Term for context based behavior?

Term for context based behavior?

We are searching data for your request:

Forums and discussions:
Manuals and reference books:
Data from registers:
Wait the end of the search in all databases.
Upon completion, a link will appear to access the found materials.

Is there term in neuroscience that describes change in neuronal circuits depending on context: low sugar/fear/… => hormones => different pathways of behavior… ?

I know that is something from Behaviorism but I search something more from "real science" like from computational biology, neuroscience etc

Human behaviour

Our editors will review what you’ve submitted and determine whether to revise the article.

Human behaviour, the potential and expressed capacity for physical, mental, and social activity during the phases of human life.

Humans, like other animal species, have a typical life course that consists of successive phases of growth, each of which is characterized by a distinct set of physical, physiological, and behavioral features. These phases are prenatal life, infancy, childhood, adolescence, and adulthood (including old age). Human development, or developmental psychology, is a field of study that attempts to describe and explain the changes in human cognitive, emotional, and behavioral capabilities and functioning over the entire life span, from the fetus to old age.

Most scientific research on human development has concentrated on the period from birth through early adolescence, owing to both the rapidity and magnitude of the psychological changes observed during those phases and to the fact that they culminate in the optimum mental functioning of early adulthood. A primary motivation of many investigators in the field has been to determine how the culminating mental abilities of adulthood were reached during the preceding phases. This essay will concentrate, therefore, on human development during the first 12 years of life.

This article discusses the development of human behaviour. For treatment of biological development, see human development. For further treatment of particular facets of behavioral development, see emotion learning theory motivation perception personality and sexual behaviour, human. Various disorders with significant behavioral manifestations are discussed in mental disorder.

Views of Biology Teacher Candidates about Context Based Approach ☆

Biology is one of the courses that explain natural events. However, in literature there are some studies claiming that the association level of students between daily life events and biology subjects is very low. Contest-based approach is described as the starting point for the development of scientific ideas in science teaching. In this approach real-life contexts are used to introduce concepts. The purpose of this study is to determine biology teacher candidates’ view about context based approach. In this study the qualitative research method was used and the participants of this study consist of 14 volunteer senior teacher candidates from the division of biology education in Hacettepe University. A questionnaire was used as a data collection tool. The collected data was analyzed by using the content analysis method. In the light of the results, it is determined that the biology teacher candidates has some lack of knowledge about context based approach and it should be improved by using the context based approach activity in courses at university level so that the teachers candidates should be aware of with guides about context based approach.


Assembly of a context-specific GeneWalk network

The first step in GeneWalk is assembly of a network that describes the relationships between genes and GO terms, starting with a list of relevant genes obtained from a specific experimental assay (Fig. 1a). These genes could be differentially expressed (DE) between some condition (such as a genetic mutation or drug treatment) and a control experiment, or the results of a high-throughput genetic screen. GeneWalk can run with any number of input genes, but the context generally becomes better defined in the presence of many (> 10) input genes (see the “Methods” section for details). A context-specific gene network (Fig. 1a, b) is then assembled using a knowledge base such as INDRA [21, 37]. Collections of INDRA statements involving at least two different differentially expressed (DE) genes or a DE gene and GO term are assembled into a gene network such that each gene is represented as a node and each statement as an edge (Fig. 1b). For comparison, we also generated a context-specific gene network using Pathway Commons [18, 43], which generally resulted in fewer gene–gene connections and no (INDRA-originating) gene–GO connections [18, 43] (Fig. 1b). This gene network, either from INDRA or PC, is then appended to a GO network [4] in which edges represent ontological relationships between GO terms as nodes (Fig. 1a). To further connect genes to GO terms in the network, we add edges between genes and their annotated GO terms (Fig. 1a), resulting in a full GeneWalk network (GWN).

Network representation learning with random walks

To determine how genes and GO terms that constitute GWN nodes relate to one another, we perform random walks in the network. A network representation learning algorithm (DeepWalk [33]) transforms the random walks into descriptions of how the nodes are embedded in the network, yielding vector representations for each node (Fig. 1c). Specifically, short random walks sample the local neighborhood of all nodes, providing a collection of neighboring node pairs, which in turn form a training set of input–output pairs for a fully connected neural network (NN) with one hidden layer (Fig. 1c). Each input and output GWN node from each sampled pair are one-hot encoded to form respectively the input and output to the NN during training. So, this NN learns which output GWN nodes have been sampled for a given input GWN node. After training, the resultant hidden layer weights form the vector representation of any (one-hot encoded) GWN input node (Fig. 1c, see the “Methods” section for further details). In this way, groups of interconnected genes and GO terms that are mechanistically or functionally linked to each other occur most frequently as sampled gene–GO term pairs, which can be scored by the cosine similarity between their NN-derived vector representations (Fig. 1c).

Gene–GO term similarity significance testing

Next, GeneWalk calculates whether the cosine similarity values between a gene and GO terms are higher than expected by chance using a significance test (Fig. 1c). A null distribution of similarity values between node vectors is generated using representation learning on networks with randomly permuted edges (Additional file 1: Supplementary Fig. S1A). Comparisons with the null distribution yield p values for all experimental gene–GO term pairs (Fig. 1c). These p values are then corrected for multiple GO annotation testing using the Benjamini-Hochberg false discovery rate (FDR), either across all gene–GO term pairs yielding a global adjusted p value (global p-adjust), or across all GO annotations per gene (gene p-adjust). To decrease variability arising from stochastic walk sampling, network representation learning and significance testing are repeated 10 times to generate the mean and 95% confidence intervals of the p-adjust estimates as the final outputs. The gene p-adjust values rank the context-specific relevance of all annotated GO terms for a pre-defined gene of interest. The global p-adjust can be used to identify relevant genes and their functions across the whole input gene list. For both global and gene p-adjust, an FDR threshold can then be set to classify all annotated GO terms that have a high cosine similarity with this gene in a statistically significant manner. We term these GO terms as “relevant” to the gene for this biological context defined by the experimental input gene set. Gene function significance arises through a high degree of interconnections with other functionally related genes in the GWN. So genes with many relevant functions are likely central to the specific biological context and thus are prime candidates for further investigation.

Identification of ground truth benchmark datasets for testing GeneWalk

To test GeneWalk and compare its predictions, we set out to identify ground truth benchmark datasets where the relevant subset of GO annotations of individual genes are known for the specific biological context. However, as far as we could determine, no such dataset exists. Existing gene function prediction benchmarks [44] were not suitable to serve as a ground truth for this learning task due to the lack of context-specificity. We considered comparing GeneWalk predictions using simulated data. However, this approach might not adequately reflect reality and would suffer from human bias, since an in silico ground truth would be constructed from chosen first principles. We recognized that GeneWalk’s task is similar to what researchers with expert knowledge do when considering a list of genes. They use their expertise to identify which GO annotations for each gene are the most relevant for the experimental context they investigate. So to test GeneWalk, we applied it to two experimental contexts in which phenotypes and molecular mechanisms are already well characterized. We unbiasedly text-mined the primary publications that first described the experimental contexts to identify the genes and their functions that were deemed relevant according to the expertise of the authors. In this manner, we generated two ground truth datasets that enable systematic and unbiased performance assessment of GeneWalk and other functional analysis approaches on the task of identifying the relevant GO terms for each gene of interest in a particular biological context.

GeneWalk application to brain myelination RNA-seq data

In the brain (Fig. 2a), neurons are myelinated in a Qki-dependent manner by oligodendrocytes [45, 46]. The Qki gene encodes an RNA binding protein involved in alternative splicing [45, 46], and conditional Qki deletion in mouse oligodendrocytes (Fig. 2a) results in severe hypomyelination and death of the animal [46]. Analysis of RNA-seq comparing animals with Qki-deficient and Qki-proficient oligodendrocytes [45] revealed 1899 DE genes (Additional file 1: Supplementary Fig. S1B).

GeneWalk identifies myelination functions from mouse brain RNA-seq. a Schematic of the experimental design in Darbelli et al. [45]. Deletion of Qki, a gene that encodes RNA-binding proteins, in oligodendrocytes results in hypomyelination in the mouse brain. RNA-seq was performed on Qki-deficient and control mice (each three biological replicates). b Schematic with statistics of the qki GeneWalk networks (GWNs) using either INDRA or Pathway Commons (PC) as a knowledge base. Also shown is a visualization of the INDRA GWN subnetwork of myelination-related genes Mal, PllP, and Plp1, all their connected genes and GO terms. Edges (gray) connecting node pairs indicate the presence of INDRA reaction statements or GO annotations between the two respective nodes. Edges between Mal and its GO connections (numbered according to rank order in c) are highlighted (bold). c GeneWalk results for Mal in the qki condition using either INDRA or Pathway Commons (PC) as a knowledge base source to assemble the GeneWalk network. All GO terms connected to Mal are rank-ordered by Benjamini-Hochberg FDR-adjusted p-value (p-adjust), indicating their functional relevance to Mal in the context of Qki deletion in oligodendrocytes. Error bars indicate 95% confidence intervals of gene p-adjust. FDR = 0.1 (dashed red line) and domains of GO annotations (square: biological process, triangle: cellular component, and circle: molecular function) are also shown. Additional file 1 shows full GeneWalk results using the INDRA or PC knowledge base. d As in c for Plp1

We initiated GeneWalk with 1861 unique Mouse Gene Database (MGD) identifiers [47] corresponding to the DE gene set (Additional File 1: Supplementary Fig. S1B), of which 94% (1750) mapped to different human orthologs using INDRA’s integrated mouse-to-human gene mappings [47, 48]. INDRA statements were retrieved for 83% of the genes, of which the vast majority (82% of the initial 1861) had at least one connected GO term (Fig. 2b). We first investigated Myelin and lymphocyte protein (Mal), Plasmolipin (Pllp), and Proteolipid protein 1 (Plp1): the three most strongly downregulated genes (Additional file 1: Supplementary Fig. S1B) that had been previously characterized as essential for myelination [49,50,51,52]. GeneWalk determined that annotated GO terms related to myelination were most relevant to these DE genes Mal, Plp1, and Pllp (Fig. 2c, d, Additional file 1: Supplementary Fig. S1C), verifying that GeneWalk can identify GO terms for each of these genes that are pertinent for the biological context.

To investigate the algorithm’s general applicability, we also performed a GeneWalk analysis using Pathway Commons (PC), which provided 5-fold fewer reaction statements (Fig. 2b, Additional file 1: Supplementary Fig. S1D) compared to the INDRA knowledge base. INDRA also provides gene–GO term connections obtained from the literature, for example Plp1 and “inflammatory response” (Fig. 2d, Additional file 2), while GeneWalk with PC utilizes GO annotations provided by the GO consortium only (Additional file 2). Nevertheless, the ordering of GO term significance for these myelination genes was similar regardless of whether PC or INDRA was used to generate the GWN (Fig. 2c, d, Additional file 1: Supplementary Fig. S1C), demonstrating that GeneWalk is robust to differences in the underlying knowledge base and the amount of available molecular information.

Performance comparison on qki ground truth between GeneWalk and alternative functional analysis methods

Most analyses of functional genomics data use gene set-based analyses to identify enriched GO terms, but they are not designed for the end-user to easily retrieve gene-specific information. To illustrate with PANTHER GO enrichment analysis, we find that Mal is absent from the gene sets corresponding to the most highly enriched biological process GO terms and only first appears as part of “ensheathment of neurons” (108 genes) and “myelination” (106 genes), the 15th and 17th term when ranked by fold enrichment (Additional file 1: Supplementary Fig. S1E), and 63rd and 70th when ranked by p-adjust from its Fisher’s exact test. Nevertheless, we systematically compared GeneWalk against eight alternative methods [1, 3, 5, 10, 13,14,15,16] (Table 1) in their ability to rank-order myelin-related GO terms above all other direct GO annotations for the three myelin genes, Mal, Plp1, and Pllp, as an initial ground truth benchmark task (Fig. 3a, see the “Methods” section for details). The ground truth rank order for these three genes was a tied rank 1 for all GO annotations that contained the string “myelin” and thus considered relevant, and a tied rank 2 for all other GO annotations that were labeled as not relevant. For fair comparison, GO annotation versions and evidence codes used by the alternative methods were matched to those of GeneWalk, as much as their publicly available software implementations allowed these specifications (see Additional file 1: Supplementary methods for details). The alternative methods yield a set of enriched GO terms with a statistical significance score that depends on the method (e.g., the p-adjust values for PANTHER). For each gene, GO annotations are sorted by their significance scores and compared to the ground truth ranking (Fig. 3a) by Kendall’s tau rank order correspondence. For example, GO term “structural constituent of myelin sheath” is relevant specifically for Mal according to GeneWalk (Fig. 2c), but it is not enriched across the whole input gene set with PANTHER (Fig. 3b). Conversely, “protein binding” is an enriched GO term with PANTHER and also a GO annotation of Mal (Fig. 3b), but it is not related to myelin and thus contributes negatively to PANTHER’s Kendall’s tau rank order score (Fig. 3c). For this initial benchmark test of top ranking myelin GO annotations for the three myelin genes, GeneWalk outperformed all alternative methods (Fig. 3c).

Systematic comparison of GeneWalk with alternative methods and model robustness analysis. a Schematic of systematic procedure to compare alternative methods with GeneWalk. The alternative methods (see Table 1 for brief descriptions and “Methods” section for details) are mostly based on a form of GO enrichment analysis, and result in a list of (globally) overrepresented GO terms with a significance value (p-adjust). For individual genes, such as Mal, we select the GO terms that are also direct annotations of that gene and form a GO annotation relevance rank order based on the method’s significance levels. Lastly for myelin-related genes Mal, Pllp, and Plp1, we compare the results of GeneWalk (gene p-adjust) and all other methods to the same ground truth ranking which is myelin terms shared 1st and all other annotations shared 2nd using Kendall’s tau to assess the rank order correspondence with the ground truth. b Example of GO annotation relevance ranking for Mal with the procedure outlined in (a) with alternative method PANTHER. c Results of systematic comparison outlined in (a), with average Kendall’s tau values (x-axis) over the three myelin genes. Error bars indicate standard error on the mean. The y-axis indicates the number of different unique GO annotations that are significant (for GeneWalk global p-adjust and for alternative methods p-adjust at FDR = 0.1) as a percentage of all unique GO annotation terms across all qki DE genes present in the GWN. d Distribution of Kendall’s tau rank order correspondences of predictions from GeneWalk and alternative methods (Table 1) to the ground truth benchmark of the qki-context where all gene GO annotations pairs mentioned by Darbelli et al. in [45] are jointly top-ranked and all other gene–GO annotations pairs are jointly bottom ranked. All methods are ordered by the median of their Kendall’s tau distribution, indicating their relative performances. Statistical differences between GeneWalk (INDRA or PC) and other methods are determined with the Wilcoxon signed-rank sum test. See Methods for details. e Bar chart of the area under receiver operating characteristic (AUROC) performance metric for GeneWalk and alternative methods (Table 1) on the benchmark described in (e) when considered as a binary classification task: identifying gene-function pairs as relevant or not. f Boxplots of the GO term levels of all significant (for GeneWalk global p-adjust and for alternative methods p-adjust at FDR = 0.1) gene–GO annotation pairs across all qki DE genes present in the GWN. A higher GO level reflects more specific concept information in the GO ontology [7]. Direct overlap comparison of GeneWalk (with INDRA) with the rankings from alternative methods is indicated with individual data points shown. For comparison of GeneWalk (with PC), see Additional file 1: Supplementary Fig. S1F. A Mann-Whitney U test indicates the statistical differences in median levels between levels significant for only GeneWalk as compared to only the alternative method, ****p < 10 −4 . g Cumulative distribution of number of connected (black) and relevant (red) GO terms per gene, alongside a simulation that uniformly randomly sampled from the number of connected terms (gray) for GWNs with INDRA. The number of relevant GO terms was smaller than with randomly sampling connections (KS test: p < 1e−16). h Hexagon density plot for all genes of interest (N = 1861) in terms of number of connected GO terms and number of relevant GO terms (at FDR = 0.1) resulting from the Qki-deficient condition GeneWalk using INDRA as a knowledge base. i Hexagon density plot of all tested gene–GO pairs (N = 28,990) as a function of GO term connectivity and similarity significance (global p-adjust, Pearson correlation r = 0.45) for the GWN described in (h)

To compare the methods further, we extended our benchmark performance analysis to a larger set of genes. We unbiasedly defined the qki-context ground truth from [45], the primary publication that describes the Qki-deletion RNA-seq experiment and the gene regulation relevant to the hypomyelination phenotype of the mouse. Through systematic, manual text mining, we tabulated all gene-biological term pairs (Additional file 3) mentioned in the same text sentences or figures from the publication [45]. Then for each gene, its GO annotations that contained the biological terms were classified as “relevant” and assigned a tied rank order 1, and the remaining annotations as “not relevant” and assigned tied rank order 2. We cannot rule out that additional genes and functions are in truth relevant, but not mentioned in publication [45]. However, our conservative methodology does capture those considered relevant enough to be mentioned by the authors, given their expert-level knowledge of the qki-context [45]. This systematic procedure resulted in 29 different listed genes (Additional file 3). Fourteen of them were DE and had at least one GO annotation that contained the corresponding biological term, which cumulated into the unbiased ground truth benchmark data set of 37 relevant gene–GO annotation pairs and 100 not-relevant pairs (Additional file 3).

On the task of ranking the relevant GO annotations higher than not-relevant annotations across all genes present in this ground truth benchmark, GeneWalk (with PC and INDRA) had the highest median rank order correspondences compared to the alternative methods (Fig. 3d). Most of the Kendall’s tau distribution differences were also statistically significant (Fig. 3d, Wilcoxon paired-rank sum test). Moreover, we compared the methods through a binary classification task (gene GO annotation pairs are relevant or not-relevant), through the metric area under receiver operating characteristic (AUROC, see the “Methods” section for details). The AUROC is determined using the quantitative significance score -log10(p-adj), but it remains a less comprehensive metric than the Kendall’s tau, since it does not consider the relative GO annotation ranking order per gene. GeneWalk (AUROC = 0.74 and 0.69 for INDRA and PC respectively) performed better than all other methods and (AUROCs < 0.67) random selection (AUROC = 0.5, Fig. 3e, Additional file 1: Supplementary Fig. S1F, see the “Methods” section for details). The GeneWalk (INDRA) network contains 3 edges (out of 186569, Fig. 2b) that originate from the ground truth publication through INDRA’s automated text mining [21, 37]. Removal of these edges from the GWN reduces its benchmark performance only marginally and all our conclusions on the comparison between GeneWalk and other methods remain unaltered (Additional File 1: Supplementary Fig. S1G).

Enrichment-based methods also provide significance values for GO terms that are transitively connected to a gene’s direct GO annotations through at least one parental relation in the GO ontology. Extending the ground truth positives to include GO terms that are parentally related to a relevant direct GO annotation does not make a difference to our results (Additional file 1: Fig. S1H), because these additional GO terms are not direct GO annotations and thus do not contribute to the ranking. When we “parentally enhanced” the methods by propagating significant p-adjust values from any such parent GO terms down to any direct GO annotation that was not called as significant, our results remained again unaffected (Additional file 1: Fig. S1I). This demonstrates that, even when considering enriched parental GO terms, enrichment-based methods do not provide the same gene-specific information as GeneWalk.

Compared to the alternative methods, GeneWalk identified more unique GO terms for all input genes (Fig. 3c). All the alternative methods, except GeneMANIA [5], seek to find a limited number of GO terms that are relevant across all members of the corresponding input gene set (Table 1). In contrast, GeneWalk’s objective is to identify GO terms relevant to individual genes by sampling its connectivity with direct GO annotations, explaining why more unique GO terms are found (Fig. 3c). Consistently, across all input genes, GeneWalk finds GO terms that are more specific in terms of concept generality compared to the other methods (Fig. 3f, Additional file 1: Supplementary Fig. S1J-L), which we quantified via each GO term’s level in the ontology [7] (Fig. 3f). We conclude that GeneWalk ranks the known molecular functions of myelin and other genes relevant to the qki-context systematically better than all tested alternative functional analysis methods and provides more detailed gene function information across the input gene set.

Systematic GeneWalk model robustness analysis

To understand the robustness of GeneWalk performances, we assessed several model assumptions. First, we found that GeneWalk is selective by focusing on the statistically relevant genes and their functions as the total number of relevant GO terms was smaller than expected by chance (KS test, p < 10 −16 for both INDRA and PC derived GWNs Fig. 3g, Additional file 1: Supplementary Fig. S2A). Fifty-four percent (1011) of the DE genes in the GWN had at least one relevant GO term (global p-adjust < 0.1, Additional file 2). Second, despite the fact that the GeneWalk algorithm contains stochastic procedures, its output predictions are reproducible between replicate runs: no statistically significant differences were observed between the global p-adjust values of a gene–GO connection pair when GeneWalk was independently run twice and compared through a two-tailed t-test with Benjamini-Hochberg multiple testing correction (with FDR = 0.01). Third, GeneWalk performance relies on the GO ontology and gene–gene interactions in the GWN (Additional file 1: Supplementary Fig. S2, see the “Methods” section): the exclusion of either of these features weakened or abolished the ability to top rank the relevance of myelin terms for Mal, Pllp, and Plp1 (Additional file 1: Supplementary Fig. S2B). Furthermore, it resulted in a much reduced correlation with the default GeneWalk model across all gene–GO annotation similarity and global p-adjust values (Additional file 1: Supplementary Fig. S2C). Fourth, GeneWalk is context-specific: the use of all expressed genes in the genome as input substantially alters predictions (Additional file 1: Supplementary Fig. S2C). Fifth, GeneWalk does not use the GO ontology transitivity property directly: performance deterioration resulted from inclusion of direct edges between transitive gene–GO relations (Additional file 1: Supplementary Fig. S2B,C). Sixth, GeneWalk performance is robust against repeating DeepWalk 3 times instead of 10 times, or the inclusion of all input DE genes, instead of only those connected through direct gene–gene edges. These modifications had little effect on all model performances (Additional file 1: Supplementary Fig. S2B,C,D), with only minor stochastic variation between replicates (Additional file 1: Supplementary Fig. S2C,D). Seventh, GeneWalk is fairly robust against variations of the network representation learning technique: the use of biased random walks through node2vec [34] or DeepWalk [33] with very long random walks did not improve and slightly reduced their respective GeneWalk performances (Additional file 1: Supplementary Fig. S2B,C). DeepWalk with infinitely long walks is mathematically equivalent to a matrix factorization approach that generates low-dimensional vector representations through spectral decomposition [53]. So GeneWalk, which employs DeepWalk with short random walks, remains preferred to these two alternative network embedding approaches. Finally, GeneWalk’s similarity null distribution randomization scheme is robust against variations: randomization of only the gene–gene and gene–GO connections instead of all GWN edges did not substantially affect the performance or resulting similarity null distribution (Additional file 1: Supplementary Fig. S2B,C). All these conclusions were reconfirmed in our rank order correspondence task applied to a second ground truth case study detailed in the next sections (Additional file 1: Supplementary Fig. S2E,F). Overall, GeneWalk utilizes the network structure of all its data sources: the gene–gene interactions, gene–GO annotations, and the GO ontology in a robust and reproducible manner with limited stochastic variation.

GeneWalk determines function relevance independent of the degree of annotation

Genes are annotated with different numbers of GO terms. To determine whether GeneWalk is biased with respect to the number of connected GO terms per gene node (the annotation degree), we compared the number of significant GO terms to node degree. The annotation degree is known to introduce a bias into enrichment analyses based on the Fisher exact test, which overestimates significance for GO terms with large annotated gene sets [13]. We found that with GeneWalk the distribution of relevant GO terms was relatively uniform for all DE genes (Fig. 3h, Additional file 1: Supplementary Fig. S3A, Likelihood Ratio test, χ 2 test p value = 1 for both INDRA and PC), showing that there was no correlation between the numbers of connected and similar GO terms. When we considered only gene–GO term connections originating from INDRA through its automated literature reading functionality, as opposed to GO annotation, we also observed a dispersed distribution (Additional file 1: Supplementary Fig. S3B), although it was not completely uniform (Likelihood Ratio test, χ 2 -test p-value < 10 −16 ). The results show that GeneWalk does not suffer from many biases in significance testing towards genes with high or low degrees of annotation.

We also asked whether a GO term with high connectivity is more likely to exhibit strong similarity to a gene simply because it is a highly connected node in the GWN. We found that this was not the case in general (Fig. 3i), although there was a weak correlation between the number of connections for a GO term and GeneWalk global p-adjust values (Pearson correlation coefficient r = 0.45). This effect could mostly be explained by a few highly connected GO terms (Additional file 1: Supplementary Fig. S3C), e.g., “cell proliferation” (1152 connections), “apoptotic process” (967 connections), or “localization” (536 connections), for which INDRA detects many genetic associations reported in the literature. However, these GO terms reflect high-level biological concepts that are rarely the specific function of an individual gene. Indeed, in the Pathway Commons-derived GWN, which only contains GO annotations, these GO terms have far fewer connections (42, 33, and 12, respectively), and the correlation between connectivity and similarity significance was lower (r = 0.26 Additional file 1: Supplementary Fig. S3D). Therefore, we conclude that GeneWalk controls for concept generality in GO term relevance ranking and does not suffer from substantial biases related to the degree of GO term connectivity.

Generation of gene-specific functions and systematic hypotheses for Plxnb3 using GeneWalk

GeneWalk helps generate gene-specific mechanistic hypotheses. Plxnb3 was one of the most strongly downregulated genes upon Qki deletion (Fig. S1B). GeneWalk revealed that more than half of its connected GO terms were relevant (gene p-adjust < 0.1), suggesting that Plxnb3 is a priority candidate with many of its annotated functions affected by the Qki deletion (Additional file 1: Supplementary Fig. S3E). Plxnb3 is expressed in oligodendrocytes specifically [54], but it is not annotated to be involved in myelination or related to Qki (Additional file 1: Supplementary Fig. S3E, Additional file 2). Furthermore, a PubMed search of Plxnb3 with the query terms “myelination” or “Qki” yielded no results. The most relevant functions of Plxnb3 were “cell–cell adhesion mediator activity,” “semaphorin receptor complex,” “regulation of GTPase activity,” “cell chemotaxis,” and “semaphorin receptor activity” (Additional file 1: Supplementary Fig. S3E), raising the possibility that Plxnb3 could contribute to the myelination process through one of these activities. This procedure illustrates how GeneWalk can be utilized in combination with differential expression strength to predict gene-specific functions and hypotheses in a systematic manner.

Nascent transcriptome response to bromodomain inhibitor JQ1 using human NET-seq

To test GeneWalk on another well-characterized model system, we reanalyzed published NET-seq data [55] describing the response of a human T-cell acute lymphoblastic leukemia (T-ALL) cell line to treatment with JQ1 (Fig. 4a), a small molecule that targets the BET bromodomain in BRD4 and other BET family members [58]. NET-seq measures RNA polymerase position genome-wide at single-nucleotide resolution [55, 59], yielding a quantitative description of the nascent transcriptome. JQ1 treatment resulted in large genome-wide transcriptional changes [55, 58]. We calculated Pol II coverage per gene and identified differentially transcribed protein-coding genes using DEseq2 [2] (Fig. 4b). INDRA statements were retrieved for 82% of DE genes (N = 2670), 79% of which had connected GO terms. GeneWalk identified relevant GO terms for 48% of DE genes (global p-adjust < 0.1, Additional file 2), similar to the statistics for the mouse brain RNA-seq data.

GeneWalk analysis of nascent transcriptome response to BRD4 inhibition in T-ALL cells. a Schematic of the experimental design in Winter et al. [55]. NET-seq was performed on JQ1-treated MOLT4 cells (1 μM for 2 h, alongside DMSO controls, two biological replicates each). JQ1 targets BRD4 and other BET bromodomain family members, causing BRD4 to dissociate from chromatin [55]. b Volcano plot showing the results of a differential expression (DE) analysis comparing RNA Polymerase II gene coverage between JQ1 and DMSO control samples. DE genes (N = 2692), indicated in red, were used as an input to GeneWalk. All other genes are depicted in black. c All enriched Biological Process GO terms (five enriched terms, Fisher exact test, FDR = 0.05) in JQ1 condition, ranked by fold enrichment, obtained by GO enrichment analysis using PANTHER [1]. Red line indicates a fold enrichment value of 1, indicating the background. d The number of different unique GO annotations (y-axis) that are significant (p-adjust < 0.1) as a percentage of all unique GO annotation terms across all JQ1 DE genes present in the GWN. Average Kendall’s tau rank order correspondences of predictions from GeneWalk and alternative methods (x-axis) over previously identified transcriptional regulators that are part of the JQ1-context (Additional file 3) [55, 56] MYC, MYB, RUNX1, RUNX2, TAL1, SATB1, ERG, ETV6, and TCF12. Error bars indicate standard error on the mean. e Distribution of Kendall’s tau rank order correspondences of predictions from GeneWalk and seven tested alternative methods (Table 1) to the ground truth benchmark of the JQ1-context where all gene GO annotations pairs mentioned in [55,56,57] are jointly top-ranked and all other gene–GO annotations pairs are jointly bottom ranked. All methods are ordered by the median of their Kendall’s tau distribution, indicating their relative performances. Statistical differences between GeneWalk (INDRA or PC) and other methods are determined with the Wilcoxon signed-rank sum test. See the “Methods” for details. f Bar chart of the area under receiver operating characteristic (AUROC) performance metric for GeneWalk and alternative methods (Table 1) on the benchmark described in (e) when considered as a binary classification task: identifying gene-function pairs as relevant or not. g Scatter plot with DE genes as data points showing the GeneWalk fraction of relevant GO terms over total number of connected GO terms (min_f, minimum value between INDRA and PC GWNs) as a function of the number of gene connections in the GWN (N gene , again minimal value between INDRA and PC). The circle size scales with the differential expression significance strength (−log10(p-adjust)) and the color hue with min_f. Twenty genes were identified with min_f > 0.5 and N gene > 30 (gray-shaded area, see Table 2 for complete list). h GeneWalk results for the transcriptional regulator RUNX1 under JQ1 treatment. Annotated biological process terms are rank-ordered by gene FDR adjusted p value. Error bars indicate 95% confidence intervals of gene p-adjust. FDR = 0.05 (dashed red line) is also shown. See Additional file 1 for full details. i As in (h) for transcriptional regulator MYB. j As in (h) for transcriptional regulator BRCA1. INDRA annotations are indicated by class: DNA damage and repair (green), chromatin, and post-translational modifications (dark blue), signaling pathways and cellular responses (light blue), transcription and gene expression (yellow), metabolism (purple), and other GO terms (gray)

Systematic comparison of GeneWalk with alternative functional analysis methods using JQ1 ground truth

PANTHER GO enrichment analysis of the JQ1 DE gene set only yielded five high-level (generic) functions such as “ncRNA metabolic process” and “chromatin organization” with low fold enrichment (range, 1.2–1.7 Fig. 4c, Fisher’s exact test, FDR = 0.05). One alternative functional analysis method, PADOG [15], was not included because it requires as input at least three replicates and the JQ1 experiment consisted of two biological replicates per treatment [55]. Thus, we benchmarked GeneWalk and the remaining seven alternatives (Table 1) to our JQ1 context. In comparison to the seven tested alternative methods (Fig. 4d), GeneWalk identified even more unique relevant GO terms than in the application to the qki study (Fig. 3c). To compare the relevance identification performance of GeneWalk against alternative methods, we generated an unbiased JQ1-context ground truth data set through the systematic text mining procedure as described for the qki-context benchmark analysis. We extracted all gene-biological term pairs mentioned in Winter et al. [55], the primary publication that described the JQ1 NET-seq experiment, as well as the abstracts from Sanda et al. [56] and Sharma et al. [57], that altogether characterized the JQ1-context in T-ALL cells: a total of 88 relevant and 196 not-relevant gene–GO annotation pairs, from 14 different DE genes (Additional file 3). The relevance rank order correspondence test for JQ1 indicated that GeneWalk with PC outperformed all the other methods when ranked by the median of the Kendall’s tau distributions (Fig. 4e), while GeneWalk with INDRA performed on par with STRING and better than the rest. With binary classification (Fig. 4f, S3F), GeneWalk (PC) performed best (AUROC = 0.80), STRING came second (AUROC = 0.73), and GeneWalk (INDRA) ranking third (AUROC = 0.67). The other methods had AUROC values around the baseline value of 0.5 (Fig. 4f, S3F), due to their lack of significant results. Removal of the 10 GeneWalk (INDRA) network edges originating from the JQ1 ground truth publications, extending the ground truth with indirect GO annotations, or “parentally enhancing” methods with enriched indirect GO annotations did not affect the above conclusions as the results remained largely unaltered (Additional File 1: Supplementary Fig. S3G-I). The performances over the combination of qki and JQ1 benchmark data (Additional file 1: Supplementary Fig. S3J-L) reconfirm the conclusion that GeneWalk overall performs better than the alternative methods on the tasks on ranking (Additional file 1: Supplementary Fig. S3J) and binary classification of relevant GO annotations (Additional file 1: Supplementary Fig. S3K-L). We conclude that these results reveal the limitations of GO enrichment analysis when many functionally unrelated genes are misregulated. GeneWalk does not suffer from this limitation, because it is based on the local regulatory network connectivity with other treatment-affected genes.

GeneWalk identifies known transcriptional regulators responding to JQ1 treatment

To test whether we could identify any previously identified transcriptional regulator genes that were affected by JQ1 treatment, we focused on genes with a high fraction of relevant GO terms over all connected terms according to GeneWalk with both INDRA and Pathway Commons knowledge bases (Fig. 4g, fraction > 0.5). We reasoned that by further selecting for genes with a large connectivity with other DE genes (Fig. 4g, gene connectivity > 30), we might identify candidate genes that mediate the observed transcriptional changes. With this procedure, we identified 21 genes (Fig. 4g, Table 2), of which 14 (Fisher Exact test, odds ratio = 13, p = 3 × 10 −8 ) had relevant transcription-related annotations (Table 2). When also including gene–GO term relations obtained through the literature with INDRA, this number rose to 17 (Fisher’s exact test, odds ratio = 11 p = 7 × 10 −7 , Table 1). Among these were RUNX1 (Fig. 4h), MYB (Fig. 4i), and TAL1, 3 out of 8 DE genes (Fisher Exact test, odds ratio = 93, p = 2 × 10 −5 ) that have previously been identified as part of a core transcriptional circuitry important to our leukemia model system [55, 56]. The other 5 DE genes with transcription-related GO annotations in this reported core circuitry are [55, 56] MYC, SATB1, ERG, ERV6, and TCF12 (Additional file 3). Additionally, RUNX2, a previously reported transcriptional regulator of T-ALL [57], was also identified by GeneWalk (Fig. 4g). All other core circuitry components previously reported in [55, 56] were either not DE and thus not part of the input gene list or did not have any transcription-related GO annotations (Additional file 3). For this test set of 9 previously identified transcriptional regulators, GeneWalk systematically ranks transcription-related GO terms as most relevant according to Kendall’s tau rank order correspondence (Fig. 4d). Lastly, GeneWalk also found newly implicated genes (Fig. 4g) such as SUPT16H (Additional file 1: Supplementary Fig. S4A), with its most relevant cellular component term being “FACT complex” (gene p-adjust = 0.01, Additional file 2), as expected, and FOXO4 (Additional file 1: Supplementary Fig. S4B) with relevant molecular functions such as “RNA polymerase II transcription factor activity, sequence-specific DNA binding” (gene p-adjust = 0.03, Additional file 2). These results demonstrate the capability of GeneWalk to systematically identify genes with relevant transcription-related functions in the context of the JQ1 response.

GeneWalk quantitatively ranks GO annotation relevance for genes with many functions

Many genes are involved in a large variety of different processes that frequently occur through the encoded-protein serving moonlighting functions in different cellular, environmental, or biological contexts [8]. These genes will have a large number of GO annotations that might not all be relevant in a particular context. GeneWalk is well suited to identify the relevant functions for genes encoding moonlighting proteins. To look at genes serving a specific role after JQ1 treatment, we identified 20 DE genes with at least 40 connected GO terms, of which no more than 50% were relevant (Additional file 1: Supplementary Fig. S4C, Additional file 4). Among them were EGFR, a gene with many established functions discussed above, and MYC, a widely studied proto-oncogene and member of the reported T-ALL core transcriptional circuitry [55]. This explains why MYC was not identified with our transcriptional regulator analysis (Fig. 4g): the majority of MYC annotations, especially those unrelated to transcription, were insignificant in the JQ1 condition (Additional file 2). BRCA1 was another downregulated gene (Fig. 4b, Additional file 1: Supplementary Fig. S4C,D) with 23% (17) of its 73 connected biological processes being relevant (Fig. 4j, FDR = 0.05, Additional file 2). GeneWalk ranked DNA damage and repair-related processes as most relevant (Fig. 4j, gene p-adjust < 0.05), followed by histone and other post-translational modification-related terms (gene p-adjust = 0.05–0.07). Transcription, metabolism, and other GO terms were the least relevant (gene p-adjust > 0.09). These results demonstrate the capability of GeneWalk to systematically prioritize context-specific functions over less plausible alternatives, which is especially useful when considering genes encoding moonlighting proteins.

GeneWalk investigation of cellular response to isoginkgetin

To investigate the context-specificity of GeneWalk model predictions, we compared the transcriptional responses induced by JQ1 to those with the biflavonoid isoginkgetin (IsoG), a plant natural product and putative anti-tumor compound whose mechanism of action remains unknown. IsoG inhibits pre-mRNA splicing in vitro and in vivo [60] and also causes widespread accumulation of PoI II at the 5′ ends of genes, indicating an additional role as a Pol II elongation inhibitor [61]. Through DE analysis of NET-seq data obtained from HeLa S3 cells treated with IsoG (Fig. 5a), we identified a total of 2940 genes as differentially transcribed, most of which exhibited upregulated Pol II occupancy (Additional file 1: Supplementary Fig. S5A, FDR = 0.001). Using INDRA and Pathway Commons as the knowledge bases, we applied GeneWalk to these DE genes and found that 18% had at least one relevant GO term (FDR = 0.1, Additional file 2).

GeneWalk determines condition-specific functions through comparison of nascent transcriptome response to IsoG and JQ1 treatment. a Schematic of the experimental design in Boswell et al. [61]. NET-seq was performed on isoginkgetin (IsoG)-treated HeLa S3 cells (30 μM IsoG for 6 h, alongside DMSO controls, two biological replicates each). The in vivo molecular targets remain incompletely characterized, as IsoG treatment causes widespread Pol II elongation inhibition. b Venn diagram detailing the overlap (Fisher’s exact test: p = 0.02, odds ratio = 1.1, 95% confidence interval [1.0, 1.3]) of DE genes between JQ1 and IsoG treatments as described in Fig. 4b and S5A. c GeneWalk results (with PC as data source) for MYC in the JQ1 (red) and IsoG (yellow) condition. Annotated biological processes are rank-ordered by FDR-adjusted p value, indicating the relative functional importance of transcription (dark blue), DNA damage and repair (green), and signaling pathways (light blue) to MYC under the IsoG condition. The top five most relevant GO terms are described in the insets. See Additional file 2 for full details. Red dashed line indicates FDR = 0.05. d As in c for SLC9A1, showing the biological process terms that are relevant in either JQ1 or IsoG condition. e Hexagon density plot for overlapping DE genes (N = 538) in terms of number of overlapping relevant GO terms (FDR = 0.1) and number of possible shared connected GO terms for the GeneWalk network using INDRA as a knowledge base

To identify candidate genes that could be involved in the IsoG-mediated transcriptional response, we searched for genes that were both strongly differentially expressed (p-adjust < 10 −25 ) and had a large fraction of functions significantly affected according to the GeneWalk analyses with both INDRA and Pathway Commons (Additional file 1: Supplementary Fig. S5B, fraction > 0.8). In this manner, we identified three genes: HES1, EGR1, and IRF1 (Additional file 1: Supplementary Fig. S5B). HES1 had “negative regulation of transcription, DNA templated” as one of the most relevant biological processes (Additional file 1: Supplementary Fig. S5C) and has been reported to inhibit transcription elongation [62]. EGR1 and IRF1 both had as most relevant term “positive regulation of transcription by RNA polymerase II” (Additional file 1: Supplementary Fig. S5D,E).

Comparison between JQ1 and IsoG analyses indicates that GeneWalk yields condition-specific gene functions

To confirm that GeneWalk’s function assignments are not constant and depend on the experimental condition, we compared GeneWalk analyses of JQ1 and IsoG treatments. Between the JQ1 and IsoG condition, 538 DE genes were shared (Fig. 5b), marginally more than expected by chance (Fisher’s exact test: p = 0.02, odds ratio = 1.1, 95% confidence interval [1.0, 1.3]). As examples, we compared the overlap of relevant GO terms of MYC and SLC9A1, which are common DE genes between JQ1 (Fig. 4b) and IsoG treatment (Additional file 1: Supplementary Fig. S5A). MYC is annotated to be involved in 29 biological processes (Fig. 5c). Between the two GeneWalk analyses, MYC showed 5 significant biological processes and 9 molecular functions for IsoG and 0 and 1 respectively for JQ1 (Fig. 5c, Additional file 1: Supplementary Fig. S6A, FDR = 0.05). “Nucleus” and “nucleoplasm” were significant cellular components in both conditions (Additional file 1: Supplementary Fig. S6B). For SLC9A1, different biological processes were significant for each condition. For example, SLC9A1 had “potassium ion transmembrane transport” and “response to acidic pH” as relevant only for the JQ1-context and “cellular sodium ion homeostasis” specific to IsoG treatment (Fig. 5d, FDR = 0.05). Thus, despite the common technical aspects such as organism under study and sequencing assay type, GeneWalk is capable of selecting which functions are specifically relevant for each experimental condition.

Overall, the numbers of shared relevant GO terms determined by GeneWalk were relatively uniformly distributed (Fig. 5e, Additional file 1: Supplementary Fig. S6C, Likelihood Ratio test, χ 2 -test p value = 1 for both INDRA and PC), indicating a lack of systematic bias in function assignment. Many genes had no shared terms between the two drug treatments (Fig. 5e), suggesting that those DE genes have different roles in each condition. We found similar results for GO terms originating from INDRA (Additional file 1: Supplementary Fig. S6D, Likelihood Ratio test, χ 2 test p value = 1). We conclude that GeneWalk is able to determine context-specific functions as a consequence of differences in the context-specific gene–gene interactions part of the GeneWalk network.

Implementation at LSU

Nationally, CUREs have included projects designed and implemented at the university but not tied to a faculty member's research (e.g., Brownell et al., 2015 Russell et al., 2015), while others have found great success implementing portions of their own research projects (e.g., Miller et al., 2013 Venesky, 2015). We have focused our CUREs solely on implementing portions of ongoing, in-house research projects, for reasons outlined in the Discussion. We recruited faculty and graduate student projects by word of mouth and by sending emails to the department asking faculty if they had research that they felt was amenable to a CURE. We then worked with the researchers to identify the elements of their research that met the following criteria:

The research performed by the undergraduates would be directly related to an ongoing research project and make a meaningful contribution to the PI's research.

The project or experiment consisted of several treatments or replicates, allowing each student group to work on a novel portion of the project

The research protocol could be broken into several subtasks, each of which could be accomplished during a three-hour weekly lab, and/or the delay between subtasks could be at least one week (e.g., setting up and running a polymerase chain reaction [PCR] one week, then running a gel and purifying the product the next week).

The students could easily be trained to complete the required techniques.

The samples used in the CUREs were not irreplaceable, and there was enough to replicate the experiment in at least two lab sections. This last criterion is crucial for comparing results across lab sections, allowing for a level of quality control to ensure that the resulting data will be useful to the PI.

The following projects have been successfully implemented as CUREs at LSU:

CCM Gateway Entry Vector Development in Chlamydomonas reinhardtii (PI: Dr. James Moroney): BIOL 1207, fall 2011, two sections, 43 students. One goal of this research group is to characterize the CO2 concentrating mechanism (CCM) present in the eukaryotic green alga C. reinhardtii. The research involved students cloning selected genes into a Gateway entry vector. Students were given a partial sequence of the gene and were instructed to match that sequence to the C. reinhardtii genome. The partial sequences were supplied by Dr. Moroney's laboratory. Once the students found the gene of interest on the C. reinhardtii genome website, they designed primers and used PCR to amplify their gene and finally put it into a Gateway entry vector. The students had a total of 12 genes of interest and were successful in cloning all 12 genes into gateway vectors. Once a gene has been placed into a Gateway entry vector, that gene can then be put into a variety of destination vectors for expression studies, protein–protein interaction, or localization studies. This provides a very useful set of tools for future work.

Effects of the BP Oil Spill on Gulf Killifish Embryos (PIs: Dr. Ben Dubansky and Dr. Fernando Galvez): BIOL 1208, spring 2011, two sections, 51 students. The goal of this research was to investigate the long-term impact of the BP Deepwater Horizon oil spill on fish populations. The students assessed the development of Gulf killifish embryos that were exposed to sediment that had received various levels of oil from the BP oil spill. Students measured hatching rate, embryonic heart rate, and mortality. The results of this research were included in Dr. Dubansky's dissertation (Dubansky, 2013) and were recently published in a peer-reviewed journal (Dubansky et al., 2013).

Controlling Ribosomal Genes in Saccharomyces cerevisiae Using the Gal1 Promoter System (PI: Dr. Raphyel Rosby): BIOL 1207 and BIOL 1208, fall 2014, eight sections, 199 students (Supplemental Materials Figure S2 and Table S1). Students generated a total of 17 mutant strains of yeast (S. cerevisiae) in which the activity of precisely one ribosomal gene was disrupted by the addition of the Gal1 promoter. This was accomplished by transforming a linear piece of DNA containing the Gal1 promoter and homology hooks to the targeted gene, isolated from a plasmid using PCR and subsequent ethanol precipitation. The targeted mutations were confirmed by colony PCR and growth assays. Each mutant strain was generated at least in duplicates by groups of three or four students in one to three sections. The generated strains are currently being used by Dr. Rosby to elucidate the signaling mechanisms involved in aberrant ribosome biogenesis that can lead to various disease conditions (ribosomopathies) in humans.

Interplay of Genetics and Reproductive Behavior in an African Cichlid, Astatotilapia burtoni (PIs: Danielle Porter and Dr. Karen Maruska): BIOL 1207, fall 2014, two sections, 47 students. Students studied and scored videos of male reproductive and territorial behaviors to quantify male dominance. They also extracted RNA and produced cDNA from tissues of female fish to investigate the expression of several genes related to feeding behaviors and reproductive state. These data were used in Ms. Porter's thesis (Porter, 2015) and provided Dr. Maruska with pilot data for future studies.

Evolutionary Tradeoffs between Thermal and Salinity Tolerance in Copepods (Tigriopus californicus) (PI: Dr. Morgan Kelly): BIOL 1209, spring 2015, two sections, 54 students. Students measured thermal and salinity tolerance of tide pool copepod (T. californicus) lines. These lines were selected for varying levels of tolerance to heat or salinity over 10 generations. Each group of two to four students in a lab section exposed one or two lines of copepods to varying salinities and temperatures and measured copepod survival. All copepod lines were measured in each section and replicated between sections. Students participated in a single-blind study (they did not know which lines they were testing until the end of the project) to minimize bias. These data are being added to a larger dataset for eventual publication. This project was continued in spring 2016.

Competitive Ability of Invasive Elephant Ear (Colocasia spp.) against Native Plants (PI: Dr. Barry Aronhime): BIOL 1503, spring 2015, two sections, 48 students (Supplemental Materials Figure S1 and Table S1). Students investigated the ability of elephant ear (Colocasia spp.), an invasive plant, to outcompete native species in greenhouse and field experiments. In the greenhouse, each group of two to four students set up experiments in which elephant ear was potted with a different native plant and measured plant traits associated with competition throughout the semester, with the experiments being replicated across two sections. They also measured elephant ear and native plant growth in a natural field setting (Bluebonnet Swamp in Baton Rouge, Louisiana) at the beginning and end of the semester. These data are being added to a larger dataset for eventual publication. This project continued in fall 2015 and spring 2016.

Biology, Biological Sciences Track, B.S.

The B.S. degree in Biology provides a comprehensive foundation for students with interests in any area of the biological sciences, including key support courses from chemistry, physics and mathematics. This degree is appropriate for students planning for careers at the bachelor's level as well as those preparing for graduate or professional study after graduation. All B.S. students complete a common core, then select upper-division courses that match their specific interests and career plans. Research experience is built into the program for all students, as is the development of skills in scientific writing and presentation.

Students can choose from two tracks (and can switch between the tracks if their interests change). The Biological Science track is appropriate for those preparing for research careers, graduate school or employment in any area of biology, while the Biomedical Science track is appropriate for students preparing for medical, dental or veterinary programs after graduation. Students preparing for secondary education, for careers that combine biology with another area, or for the allied health fields may wish to consider a B.A. program.

For additional programs and courses in this department, see

Conflict of Interest Statement

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The authors declare no competing financial interests.


This study was supported by grants from CONICYT/FONDECYT Regular (1170010), FONDAP 15150012, and the INECO Foundation.


[1] Ibanez, A., and Manes, F. 2012. Contextual social cognition and the behavioral variant of frontotemporal dementia. Neurology 78(17):1354�. doi:10.1212/WNL.0b013e3182518375

[2] Chun, M. M. 2000. Contextual cueing of visual attention. Trends Cogn. Sci. 4(5):170𠄸. doi:10.1016/S1364-6613(00)01476-5

[3] Barrett, L. F., Mesquita, B., and Gendron, M. 2011. Context in emotion perception. Curr. Direct Psychol. Sci. 20(5):286�. doi:10.1177/0963721411422522

[4] Beck, D. M., and Kastner, S. 2005. Stimulus context modulates competition in human extrastriate cortex. Nat. Neurosci. 8(8):1110𠄶. doi:10.1038/nn1501

[5] Bar, M. 2004. Visual objects in context. Nat. Rev. Neurosci. 5(8):617�. doi:10.1038/nrn1476

[6] Baez, S., and Ibanez, A. 2014. The effects of context processing on social cognition impairments in adults with Asperger’s syndrome. Front. Neurosci. 8:270. doi:10.3389/fnins.2014.00270

[7] Baez, S, Garcia, A. M., and Ibanez, A. 2016. The Social Context Network Model in psychiatric and neurological diseases. Curr. Top. Behav. Neurosci. 30:379�. doi:10.1007/7854_2016_443

[8] Masayuki, H., Takashi, I., Mitsuru, K., Tomoya, K., Hirotoshi, H., Yuko, Y., and Minoru, A. 2014. Hyperscanning MEG for understanding mother-child cerebral interactions. Front Hum Neurosci 8:118. doi:10.3389/fnhum.2014.00118

[9] Liu, N., Mok, C., Witt, E. E., Pradhan, A. H., Chen, J. E., and Reiss, A. L. 2016. NIRS-based hyperscanning reveals inter-brain neural synchronization during cooperative Jenga game with face-to-face communication. Front Hum Neurosci 10:82. doi:10.3389/fnhum.2016.00082

[10] Makeig, S., Gramann, K., Jung, T.-P., Sejnowski, T. J., and Poizner, H. 2009. Linking brain, mind and behavior: The promise of mobile brain/body imaging (MoBI). Int J Psychophys 73:985�

Term for context based behavior? - Biology

Professor Emerita
Department of Information Studies
Graduate School of Education and Information Studies
University of California, Los Angeles (UCLA)

Copyright © 2010
By CRC Press

Bates, Marcia J. (2010) Information Behavior In Encyclopedia of Library and Information Sciences, 3rd Ed.
Marcia J. Bates and Mary Niles Maack, Eds. New York: CRC Press, vol. 3, pp. 2381-2391. (Also available online at subscribing libraries.)

"Information behavior" is the currently preferred term used to describe the many ways in which human beings interact with information, in particular, the ways in which people seek and utilize information. The broad history of research on information seeking behavior over the last 50-60 years is reviewed, major landmarks are identified, and current directions in research are discussed.

Information seeking Information use Information needs Information genres Online catalogs Online searching Digital libraries Library use

"Information behavior" is the currently preferred term used to describe the many ways in which human beings interact with information, in particular, the ways in which people seek and utilize information. Information behavior is also the term of art used in library and information science to refer to a sub-discipline that engages in a wide range of types of research conducted in order to understand the human relationship to information.

Interest in this area developed out of several streams. Librarians wanted to understand library users better, government agencies wanted to understand how scientists and engineers used technical information in order to promote more rapid uptake of new research results, and social scientists generally were interested in the social uses of information in a variety of senses. In more recent years, social studies of information technology and social informatics have contributed to this area as well. Within library and information science, these various streams of research are drawn on for what they can contribute to a richer understanding of information behavior.

What, then, is information? Here, rather than review the many senses in which this term has been interpreted in the field (see ELIS entry "Information"), we will rely on a sense of the term that is an extended understanding of the concept as used in general conversation. We all recognize that people search for information on, say, the history of a small town, the population of Turkey, or how to do foreign exchange trading online. All these examples make a reasonable match with the generally understood sense of information as being factual, statistical, and/or procedural.

"Information," however, is used in a broader sense as well in the world of information behavior research. The term is generally assumed to cover all instances where people interact with their environment in any such way that leaves some impression on them—that is, adds or changes their knowledge store. These impressions can include the emotional changes that result from reading a novel, or learning that one’s friend is ill. These changes can also reflect complex interactions where information combines with pre-existing knowledge to make new understandings, or enables the individual to deduce or induce new thoughts and ideas. As the Hans Christian Andersen tale suggested, the ugly ducking did not realize that he was a swan until he came in contact with swans, saw his reflection in water, and figured out that he was himself a swan, too. (1)

These information interactions can also leave a negative impact &ndash one may ignore, deny, or reject information (2) . (See also an excellent early analysis of relations to information by Atkin. (3) ) One may also simply discover that nothing has changed &ndash the university admissions letter still hasn&rsquot come in the mail. This negative news is, of course, informative in its own way, just as a person who has ignored information has often, in some way or other, nonetheless absorbed it. In fact, probably the largest amount of all information taken in by human beings is that received passively &ndash simply through being aware that is absorbed in the context of daily living. (4)

Is this not a very broad understanding of information behavior? Indeed, does it not cover all interactions people have with their environment? Bates has argued:

In comparison to other social and behavioral science fields, we are always looking for the red thread of information in the social texture of people’s lives. When we study people we do so with the purpose of understanding information creation, seeking, and use. We do not just study people in general. … In communications research, a cousin to our field, the emphasis is on the communication process and its effects on people in information science we study that process in service of information transfer. (5)

Bates goes on to provide a specific example:

…[T]here are social scientists today who are observing people doing collaborative work through new types of networked systems in the field of computer-supported cooperative work (CSCW). The sociologist or social psychologist identifies and describes the network of relationships and the social hierarchy that develops under these circumstances. …

The information scientist, on the other hand, follows the information…. That&rsquos the red thread in the social tapestry. When we look at that social hierarchy, we are not interested in the hierarchy per se, but, rather, we ask how it impedes or promotes the transfer of information. We ask what kinds of information people prefer to communicate through this or that new channel of information technology. We always follow the information. (5)

Thus, the study of information behavior can cast a very wide net, looking into both individual interactions as well as large-scale complex group and societal interactions with information. Indeed, as we shall see, the variety of contexts in which information behavior has been studied demonstrates this breadth. But information behavior research is not communication, psychology, education, sociology, or social impacts of technology research, though all those disciplines may find the work interesting to discover. Rather, information behavior research actually studies &ndash and largely limits itself to &ndash information-related behavior.

History of Information Behavior Research

From the earliest days, librarianship in the United States had a commitment to care about and serve the users of libraries. In the founding year of American professional librarianship, 1876, Samuel Green wrote to encourage librarians to "mingle freely" with the library’s users "and help them in every way." (6) In the mid-twentieth century, the great Indian librarian, S.R. Ranganathan, promulgated his Five Laws of Librarianship, which were very much oriented to the library user:

  1. Books are for use
  2. Every reader, his book
  3. Every book its reader
  4. Save the time of the reader
  5. The library is a growing organism (7)

However, for many decades that commitment remained largely on the plane of values and had little other than anecdotal data upon which to develop library services. In the 1930&rsquos, the Graduate Library School at the University of Chicago (8) introduced the first doctoral degree in library science in the U.S. Sophisticated social science researchers, such as Douglas Waples and Bernard Berelson, brought their skills to the field. Waples (9) did research on reader preferences, and Berelson, among other things, produced a compendium of results from dozens of studies on public library use. (10) (11)

The experiences associated with the operation of "Big science" during World War II &ndash major projects such as the development of the atom bomb &ndash led government leaders to see the advantages in improving the distribution and transfer of information on new discoveries to other scientists and engineers. Major conferences on scientific information, in 1948 and 1959, led to a substantial amount of money being invested during the 1950&rsquos and 1960&rsquos in research on how scientists gathered and used information in their research work. Major example publications were the proceedings of the two science information conferences, (12) (13) the 21-report series "Project on Scientific Information Exchange in Psychology" from the American Psychological Association, (14) and the work of Garvey et al. on several disciplines. (15) (16) Other influential early works include publications by Derek de Solla Price, (17) (18) Diana Crane, (19) and A.J. Meadows. (20) (Note: In order to keep this entry&rsquos long bibliography from being even longer, referenced items are often only a sample of a person&rsquos work, and when a series of articles comes out from a project, generally only the last article in the series is referenced.)

Early on, studies on information behavior were called "use studies" (21) studies of "information seeking and gathering," or studies of "information needs and uses". (22) Gradually, the term "information seeking research" was used to include all kinds of research on people&rsquos interaction with information.

More recently, however, some researchers came to feel that "information seeking" suggested only explicit efforts to locate information, and did not include the many other ways people and information interacted. In the 1990&rsquos, the term "information behavior" came into wide use to replace "information seeking." The Old Guard objected that the phrase is a non sequitur &ndash information does not "behave" but, they lost out, and "information behavior" remains the most commonly used term today.

During the 1960&rsquos, in particular, generous funding was available in the United States for social science research, and a great deal of knowledge, based on large, well-designed studies, was developed regarding the social aspects of scientific communication and information use. (23) (24) (25) Important studies were also produced on information use and library use by the general public (26) (27) (28) (29) . Focus in the larger society during the 1960&rsquos and 1970&rsquos on identity politics of race, gender, sexual orientation, and the economically under-privileged also led to research attention being directed to information seeking of the corresponding population groups.

In the late 1960&rsquos and early 1970&rsquos, this research began to be taught in library and information educational programs in North America. (30) As scientists had been studied according to their disciplines &ndash physics, biology, etc. &ndash and many members of the general public had been studied by their social identities &ndash the poor, the elderly, etc. &ndash there was a tendency to study information-related behavior by looking at groups of these sorts. For example, an invited conference on "information service needs of the nation" was funded by the U.S. National Commission on Libraries and Information Science in 1973. Presentations were structured in a parallel format to address the needs of a number of groups, including people working in science, agriculture, business, labor, biomedicine, the arts, social services, as well as children, the geographically remote, the economically and socially deprived, the institutionalized, and the mentally and physically handicapped, among others. (31)

After the earlier attention to the natural sciences, during the 1970&rsquos research attention turned to information transfer in the social sciences. Grant funding in the U.S. receded, and pride of place went to Great Britain, where several researchers engaged in creative and revealing research on information seeking and use in the social sciences. (32) (33) (34)

Finally, in the 1980&rsquos and 1990&rsquos the under-funded humanities at last got their due, (35) (36) (37) (38) particularly with the support of large institutions such as the J. Paul Getty Trust. (39) In the 1990&rsquos interdisciplinary and area studies researchers were addressed (40) See, especially Carole Palmer, as Issue Editor, of an issue of Library Trends on interdisciplinary information seeking, (41) as well as her subsequent book. (42)

In the 2000&rsquos, Kari and Hartel made a persuasive case for studying the information behavior of people engaged in activities aimed at fulfillment and self-realization, and their own research provided examples of what could be learned along this line. (43)

Over the decades, varying amounts of information behavior research has been done in various professional contexts as well, including the health sciences, (44) law, (45) and business. (46) Among the professions, it is almost certainly the health sciences where the largest body of information behavior research has been done—probably due to abundant funding—while the education profession, despite the importance of information seeking for teachers, seems, mysteriously, to have drawn very little attention. (47) (48)

Throughout the years, a number of models have been proposed to characterize various aspects of information behavior. Paisley beautifully characterized the subjective world of the scientist, as constituting a series of contexts &ndash local work environment, research specialty, discipline, larger cultural and political world, etc. (24). In 1981 Tom Wilson described information seeking in general in a model (49) that was subsequently very widely used, and also reviewed a wide range of information behavior models in 1999. (50) Belkin et al. propounded the concept of "anomalous state of knowledge," or ASK, (51) as characterizing many information needs. That is to say, they argued that the information need is often complex and requires an extensive description to cover all the factors really at play in people&rsquos requests. Kuhlthau&rsquos Information Search Process model, based on extensive research, demonstrated how intricately the conceptualization of a paper or project was bound up with confusions and problems in searching for information. (52) Bates&rsquo "berrypicking," i.e., picking up a bit of knowledge here and a bit of knowledge there, was seen to be an appropriate description of much of human searching to meet information needs, (53) in contrast to the previous generally assumed simple query that could be answered by a single retrieval from just one database.

Though extensive research on information seeking inside and outside of library and information science had been going on since the 1950&rsquos, it was an article by Dervin and Nilan in 1986, (54) however, that seemed to provide the impetus for a great increase in interest in the subject in library and information science. The authors articulated the value of placing the user/searcher at the center of research, and paying close attention to the internal motivations and needs of the information seeker. From a minority interest of a relatively few people, information behavior research exploded in LIS after that article appeared, and doctoral students flocked to the subject area. For example, the number of articles dubbed "Use studies," the standard term used in Wilson Web&rsquos article database Library Literature and Information Full Text, doubled per year in the five years between 1985 and 1990 &ndash from 76 to 155 &ndash while in the subsequent 18 years, the annual number has gone up by less than 60 percent to 245 in 2008 (author&rsquos database search). (Of course, these results could be artifacts of the publisher&rsquos indexing practices, and a fuller exploration would be needed to verify this conclusion.)

In particular, Dervin’s conception of “sensemaking,” the effort of people to make sense of many aspects of their lives through information seeking and use, has been a dominating force in recent research on information behavior. (55)

Dervin dismissed prior studies on grounds that "the studies assumed that the information brick was being thrown into the empty bucket" &ndash i.e. into the user of information. (56) In one blow, this clever metaphor both characterized and caricatured much of the more classically empirical scientific approaches to research on information behavior, and gave qualitative research techniques and philosophies a boost. Dervin&rsquos "brick" image was unfair to the many researchers who did not take a simplistic view of information transfer, including many of the people mentioned in this article to this point. (57) However, her emphasis on the importance of sense-making in motivating information seeking legitimated the subsequent emphasis on qualitative techniques in the field, and enlarged the perspective of the whole sub-discipline of information behavior.

Indeed, over the years, increasing dissatisfaction was expressed by some researchers toward the prior orientation either to the individual seeking information, or to studying the tendencies and preferences of large social groups, such as physicists or older people. These researchers sought to expand information behavior research, drawing on several theoretical paradigms of interest in the social sciences, such as social constructivism, social constructionism, and ethnographic techniques. (58)

The surest sign of this broader interest came in the form of the "Information Seeking in Context (ISIC)" conference that came to be presented every other year, mostly in Europe, beginning in 1996. (59) Conference attendees have sought to study information behavior in a way that goes beyond traditional research designs. They argue that context must be understood in a much fuller sense they argue for rich, detailed, qualitative study of specific situations and contexts, in order to understand the very nuanced ways in which people might receive and shape information.

They draw upon many different information-related theories and models, (60) as well as on the many varieties of metatheoretical and philosophical perspectives that have become popular in the social sciences and humanities. (61) See, as examples of these newer approaches, Ellis&rsquo grounded theory approach, (62) Talja&rsquos discourse analysis of the culture of music in relation to libraries, (63) Xu&rsquos application of activity theory to interactive information retrieval, (64) Reddy and Jansen&rsquos ethnographic study of collaborative information behavior in healthcare, (65) Limberg&rsquos (66) and Ford&rsquos (67) (68) use of educational theory, and Srinivasan and Pyati&rsquos critical re-examination of information environments for diasporic groups. (69)

At the same time, research drawing on other, more classically scientific and engineering methodologies did not disappear. See Fidel and Pejtersen&rsquos use of the Cognitive Work Analysis Framework, (70) Sandstrom and Sandstrom&rsquos analysis of the methods of scientific anthropology as applied in library and information science, (71) Nicholas et al.&rsquos study of online information seeking through transaction log analysis, (72) and even Bates&rsquo use of biological and evolutionary concepts in her recent work on information (73) and browsing. (74)

Perhaps the greatest sign of maturity of the field of information behavior research came with the publication &ndash at last! &ndash of the first book comprehensively addressing information seeking, by Donald Case, in 2002, with a second edition in 2007. (75)

The popularity of the ISIC conferences demonstrates the recent efflorescence of qualitative information behavior research beyond the borders of the (sometimes self-absorbed) research culture of the United States. Scholars from the U.K. (Tom Wilson, David Ellis, Nigel Ford, Elizabeth Davenport), Ireland (Crystal Fulton), Scandinavia (Louise Limberg, Olof Sundin, Annelise Mark Pejtersen), and Finland (Pertti Vakkari, Reijo Savolainen, Sanna Talja, Jannica Heinström) have presented and published at ISIC and elsewhere. Australian (Kirsty Williamson, Theresa Anderson) and Canadian researchers (Heidi Julien, Karen Fisher, Gloria Leckie, Lynne McKechnie, Pam McKenzie, Roma Harris, Chun Wei Choo) have also been very active in recent years.

Recently, Savolainen may have marked the beginning of a new phase in information research when he urged that the qualitative research on information behavior be called instead the study of "information practice." (76) He argued that the concept of "information behavior" is primarily associated with the cognitive viewpoint, while "information practice is mainly inspired by the ideas of social constructivism." (77)

The concepts of information behavior and information practice both seem to refer to the ways in which people &lsquodeal with information.&rsquo The major difference is that within the discourse on information behavior, the &lsquodealing with information&rsquo is primarily seen to be triggered by needs and motives, while the discourse on information practice accentuates the continuity and habitualization of activities affected and shaped by social and cultural factors…. (78)

In the last several years, there has also been a very active Special Interest Group on Information Behavior, founded by Barbara Wildemuth and Karen Fisher, among others, in the American Society for Information Science and Technology (SIG-USE), which has held a number of pre-conferences and conference sessions, and offered awards for research in the area.

Information Searching vs. Information Seeking

The above discussion addressed research on how people interact with information, how and when they seek information, and what uses they make of it. But it should be understood that throughout this period of time a parallel body of research and practical application was continuing that addressed the specifics of the act of searching itself. That is, in working with paper and online resources, many problems were encountered and skills needed to succeed in the specific acts associated with locating information in a paper or online resource. Bates&rsquo articles on information searching tactics and search techniques (79) (80) promoted greater attention to the complexities of identifying sources and working one’s way through resources to locate the desired information. A long line of research followed that addressed both search success and desirable design features in information systems to promote ease of use (81) (82) (83) (84) (85) (See also ELIS entry "Information Searching and Search Models.") Even browsing, normally seen as the most unstructured method of information searching, came in for considerable attention (74). (86) (87)

Role of Technology in Information Behavior Research

In order to simplify the narrative line, the above discussion made little mention of the role of technology in information seeking and research on information seeking. But, in fact, the extraordinary changes in information technology (IT) over the last 50ᇐ years have meant that a great deal of information behavior research has also been concerned with impacts of and reactions to the kinds of interactions people experience when using new technologies for finding and communicating information.

Focus on impacts of, and roles of, IT in information behavior has been intertwined to a greater and lesser extent with the information behavior research over these decades. Early studies took a fairly stable, largely paper-based environment for granted. Indeed, Garvey&rsquos research (23) made salient, perhaps for the first time for many readers of his work, the huge, complex scientific publication cycle, from early tentative verbal presentation at talks all the way through conference presentations, summary reports, journal publication, annual reviews, and finally, incorporation of the scientific results into the established canon in textbooks.

But consciousness of the complexity of the production and publication of science was soon joined by efforts to improve, especially to speed up, the collection, storage, organization, and dissemination of that information.

Indeed, the entire discipline of information science has, in one sense, been the story of the successive absorption of a long series of IT innovations, followed, in each case, by research on the impacts of those innovations, and efforts to improve access to information through optimal design of those innovations. With the excitement generated by each new technology, the relatively stable underlying human behaviors and reactions were sometimes forgotten, and the new technology instead seen as the source of a totally new information seeking landscape. One thing we now know, however, after a lot of research on those successive waves of new technology, is that underlying human propensities with regard to information emerge again and again, as each new technology becomes familiar and its use second nature. Often, in the end, the new technologies offer speed and ease of use, while otherwise replicating previous social structures and interactions.

We know, for example, that people are willing to commit very little energy and effort to seeking information, in contrast, say, to seeking a fortune, a family, or a reputation. In fact, the truly explosive popularity of the World Wide Web as an information source may be due to the fact that the level of effort the searcher must engage in to find an answer to a question on the Web is at last so very little as to slip in under that minimal level of (least) effort that feels "natural" in information seeking. Most of the information that people eagerly seek online was once available in their local public or academic library, but the effort required to locate that information was seen as excessive in the vast majority of cases.

In the rest of this section, we will follow several IT innovations and consider their impact on information behavior research.

The first major technology in modern times to affect information seeking was the computer. Initially, its use for library information systems was limited&ndashcomputers were used to capture machine-readable versions of library catalog records ("MARC" records), which, in turn, enabled the publication of computer-produced print-on-paper book catalogs. This was followed, in short succession, by so-called "COM cats," that is, computer output microfiche catalogs, which could update book catalogs between publications of paper editions. (88)

It will be forgotten today, that in the age of card catalogs, while in one library, one could not access the catalog of any other branch of the academic or public library, or of any other library, for that matter. In academic libraries, a comprehensive copy of all the materials on campus was generally available only in the main library. Disseminating multiple copies of the full set of library records through book catalogs and COMcats in branch libraries was a significant, time-saving innovation.

These catalog innovations during the late 1960&rsquos and 1970&rsquos were followed by a true revolution in catalog accessibility&ndashthe online catalog, which was developed in the early 1980&rsquos. These constituted the first widely-available end-user information search systems, and much was learned about how untrained people did and did not succeed in this form of online searching These catalog innovations during the late 1960&rsquos and 1970&rsquos were followed by a true revolution in catalog accessibility &ndash the online catalog, which was developed in the early 1980&rsquos. These constituted the first widely-available end-user information search systems, and much was learned about how untrained people did and did not succeed in this form of online searching. (89) (90)

For a variety of reasons, the card catalog structure could not be simply translated into online form. Questions of re-design of catalog access in the new context, and the development of new and faster system designs to improve access occupied many in LIS research over the next 10-15 years. (91) (92), (93) (See also ELIS entry "Online Catalog Subject Searching.")

In the meantime, (at least) four other overlapping information-related revolutions were occupying the field as well. The first revolution occurred in the area of information retrieval, where various forms of automatic indexing and retrieval were experimented with over decades from the 1950&rsquos forward, gradually improving the speed and effectiveness of both retrieval and ranking algorithms. (94) (95) In the 1990&rsquos, search engines, such as Alta Vista and Google, drew upon these retrieval techniques to design their Web systems.

Second, in the early 1970&rsquos, online database searching was made practicable through searching against large databases on "dumb" terminals receiving and sending data over telephone wires. "Online searching" as then understood, and as then implemented by database vendors, was a complex skill that required considerable training to do well. Teaching these skills became an important part of LIS education, and drew a lot of research interest as well. That type of searching required a mix of gifts that not everyone has, and numerous studies of online searching behavior resulted (82). (96) (97)

The third revolution was the development of the Internet and World Wide Web, which enabled access to information all over the world from anywhere in the world. We are still working through the many impacts and implications of this capability for all prior information technologies and sources of information. (98) , (99) , (100)

The fourth revolution occurred with the widespread interest in creating digital libraries of all manner of textual and image material &ndash and sometimes online portals to access those resources. The Digital Libraries Initiative in the 1990&rsquos marked the moment when, at last, truly substantial amounts of research money entered the information science field. Ann Bishop and colleagues addressed at book length the socio-technical factors of digital library use. (101) The new capability of storing and easily accessing previously unimaginably large bodies of information in digital libraries led to innovative experiments in the storage and use of primary resources materials. Example studies are those of children using primary archival materials, (102) uses of texts in the field of classics in a digital library, (103) and use of a medical portal. (104)

During the 1970&rsquos to the present, many studies of information behavior involved, to a greater or lesser extent, research on people&rsquos use of and success with, these innovations in information access. On the whole, more behavioral research was done in the areas of online catalogs and online database searching than in information retrieval. For a long time, IR researchers were not particularly receptive to, or interested in, the human side of the equation, though in the 1990&rsquos, they came to realize that people needed attention, too, in the overall effort to improve retrieval. See, for example, the contrasting emphases in the two entries in this encyclopedia by Salton ("SMART System: 1961-1976") and Järvelin and Ingwersen ("User-Oriented and Cognitive Models of Information Retrieval").

With the development of frequent interaction with microcomputers in the early 1980&rsquos, the already-thriving field of human-computer interaction research, or HCI, exploded and became a still larger field. HCI paid little attention to LIS research, however, and LIS paid little attention to HCI research, probably to the detriment of both fields. However, there may have been good reasons for this mutual indifference. The specific circumstances of needing and seeking information are not well understood, for the most part, outside of LIS, and required the focused attention of LIS researchers. At the same time, HCI researchers were working to discover general principles applicable to all online and computer access, and therefore tended to ignore the distinctive features of various "application" fields, including information seeking. (See ELIS article "Human-Computer Interaction for Information Retrieval.")

In this encyclopedia, Diane Nahl&rsquos two articles on early and recent "User-centered design," as well as articles by Elaine Toms ("User-Centered Design of Information Systems") and Judith Weedman ("Design Science in the Information Sciences") address, in much greater detail, the efforts and results in this area at the intersection of information technology and the study of information behavior.

Range of Topics of Information Behavior Research

What have we learned over the years from the study of information seeking behavior? This is a hard question to answer briefly, to put it mildly, but a description of the sequence of research topics of interest over the years may give a hint of the growing understanding over time of the human relationship with information. What follows is a mere sampling.

In the 1940&rsquos and 1950&rsquos information seeking and gathering tended to be viewed implicitly as the study of the use of various forms of literature &ndash books, journals, handbooks, etc. &ndash and of various types of institutions and their services. How many books were circulated, how many reference questions were asked, how many people of what economic strata used the public library, and so on (see (10) ).

In the 1960&rsquos and beyond, studies of information seeking and use by the general public opened out the research to incorporate many sources of information, of which the library was only one. (26) (105) The first surprise was to discover how much information &ndash in both personal and professional contexts &ndash people got from friends and colleagues. In a study looking at how scientists&rsquo learned of things serendipitously, Menzel found that fellow scientists were immensely important in that process. (106) In fact, in a large number of studies, the human preference to get information from other people was soundly demonstrated.

From early on, the dominance of the "principle of least effort" in human information seeking was demonstrated over and over (25). It may not seem surprising that people try to minimize effort in finding information, but the research demonstrated that ease of access and ease of use mattered more to people than the quality of the information they found. People have a (sometimes unjustified) belief in their ability to filter the good and valid information from the faulty, hence their tendency to under-search to find the highest quality information available.

Further, information seeking is often quite unself-conscious. People are trying to solve problems in their lives, not "seek information." Activities that involve information seeking are seldom differentiated from the other actions taken to solve problems. Good research design for the study of information seeking must recognize this reality asking people what they have done lately in the way of information seeking is therefore not the way to get data with high internal validity, as a rule. Dervin has made this point repeatedly and insightfully. For a prime example of her research technique, see (107) .

Thus, as a rule, people &ndash even including Ph.D. scholars &ndash develop what search skills they have incidentally to their primary efforts at research or problem-solving, and often fail to develop a conscious repertoire of search skills and techniques to help them over difficult stages. Particularly among college students, discomfort regarding library research has been found to be severe enough to merit the term "library anxiety," and a number of studies have been done on this topic (See ELIS entry "Library Anxiety.") Along with the evidence of student difficulties with libraries, a large literature has developed on the goals and techniques of teaching "information literacy," i.e., the capability of finding and effectively evaluating desired information (See ELIS entries "Information literacy" and "Information literacy instruction.") In sum, people often vastly under-utilize available resources and are often quite inefficient in finding what they do find.

In the study of various academic disciplines, the close attention in the 1960&rsquos to the rich complexity of the culture of science enabled a subtler analysis of the information seeking in all the academic disciplines studied from the 1970&rsquos to the present time. See, for example, Patrick Wilson on the concept of "cognitive authority," (108) Julie Hurd on implications of information use patterns for library design, (109) Paisley on "information and work," (110) Robert Taylor on "information use environments," (111) Cronin on invisible colleges (i.e., informal groups of researchers with shared interests), (112) the model by Leckie et al. of information seeking in the professions, (113) and Budd (114) and Bates (115) comparing the cultures of science and humanities. In the under-studied area of archival resources, Barbara Orbach (116) and Wendy Duff and Catherine Johnson (117) have provided insightful descriptions of the use of historical archival materials.

During the 1980&rsquos and 1990&rsquos several researchers deepened the understanding of various aspects of information behavior by exploring questions and areas previously not as well understood. Elfreda Chatman looked at the information environments of janitors, women in a retirement home, and prisoners. (118) (119) (120) Cheryl Metoyer-Duran applied the concept of gatekeeping to five minority groups in Southern California, and developed sophisticated (and sometimes counter-intuitive) understandings about information flow within minority communities, and between them and the larger society. ( 121) The challenge of studying unconventional groups and domains even extended to abused women (122) and abused children. (123)

Carol Kuhlthau is another researcher who has had very wide influence in the information behavior world. She developed a model of student information seeking, which she refined over several studies that are themselves models of the art of research. Her model runs counter to many assumptions in both education and library and information science about how people tackle researching a paper or project, and how that experience can be substantially improved over past approaches. (124) Specifically, she discovered that the combined process of researching and writing a paper is complex and difficult for most people &ndash indeed, the library research is inextricably bound with the understanding and gradual formulation of the thesis of the paper. Consequently, the simple idea of "picking a topic," like picking an apple off a tree, then going to research it in the library, is not how the process reasonably can or should be expected to proceed. Yet generations of teachers and professors have left students floundering and frustrated as they moved, essentially without guidance, through this core process in paper-writing.

David Ellis&rsquo empirically-based model of common actions associated with scholarly information seeking (125) has also been influential, spurring several follow-on studies to test for similar activities in the work of people in other circumstances. And, of course, Brenda Dervin&rsquos concept of "sensemaking" as a motivation for information seeking constituted the underlying model for much information behavior research (55).

In the 1990&rsquos and 2000&rsquos, along with the growth of the ISIC community, researchers expanded their look at information behavior by incorporating the whole environment &ndash physical, social, and technological &ndash in the study of people&rsquos interactions with information. Social context and social situation were recognized as essential to the understanding of information seeking. (126) (127) Karen Fisher (nee Pettigrew) developed the concept of "information grounds" &ndash the joint creation of social environments by people in which to share conversation and information. (128) Disciplinary examples of these rich analyses include science (42) and business. (129)

Several recent studies have demonstrated particularly creative approaches to the study of children&rsquos information seeking, traditionally an under-populated area of research. Virginia Walter demonstrated that children&rsquos information needs were immense, and were in no way limited to requests made of school librarians! (130) Joanne Silverstein studied unconventional forms of information use, (131) and Ciaran Trace studied informal information creation and use by children. (132)

During recent decades, a more sophisticated understanding has also developed of information genres and the ways they are shaped by practice. In a particularly elegant study, Kling & McKim showed how pre-existing social information practices shaped the design of post-Web online information support in three scientific disciplines. (133) Peiling Wang studied at the micro level how scientists actually make use of and subsequently cite other literature in the course of their research. (134) Ann Bishop (135) and Lisa Covi (136) studied closely the interactions between people and the structure and genres of information.

Information behavior research has grown immensely from its scattered beginnings earlier in the twentieth century. We now have a much deeper and less simplistic understanding of how people interact with information. We understand information behavior better within social contexts and as integrated with cultural practices and values. The further complexity of information seeking through the use of various technologies and genres is coming to be better understood, though there is much more to be studied. In fact, even as I write, some six billion people are interacting with information worldwide, drawing on cognitive and evolutionarily shaped behaviors, on social shaping and environmental expectations, and interacting with every information technology from the book to the wireless handheld "smartphone." There is unimaginably much more to learn about information behavior.

The state of our current understandings on these topics is reviewed in over 30 articles in this encyclopedia. See, especially, the section titled "People using cultural resources" in the topical contents list of the encyclopedia.

(1) Andersen, H.C. Lucas, E. The ugly duckling. In Fairy Tales from Hans Christian Andersen, 3 rd Ed. J.M. Dent & Co.: London, 1907 379녋.

(2) Case, D.O. Andrews, J.E. Johnson, J.D Allard, S.L. Avoiding versus seeking: the relationship of information seeking to avoidance, blunting, coping, dissonance, and related concepts. Journal of the Medical Library Association 2005, 93 (3), 353넲.

(3) Atkin, C. Instrumental utilities and information seeking. In New Models for Mass Communication Research Clarke, P., Ed. Sage: Beverly Hills, Calif., 1973 Vol. 2, 205낷.

(4) Bates, M.J. Toward an integrated model of information seeking and searching. New Review of Information Behaviour Research 2002, 3, 1ᆣ. Also available at (Accessed December 2008).

(5) Bates, M.J. The invisible substrate of information science. Journal of the American Society for Information Science 1999, 50 (12), 1048.

(6) Green, S.S. Personal relations between librarians and readers. American Library Journal 1876, 1, 78.

(7) Ranganathan, S.R. The Five laws of Library Science, 2 nd Ed. Blunt and Sons: London, 1957. See also (Accessed December 2008).

(8) Richardson, J.V., Jr. The Spirit of Inquiry the Graduate Library School at Chicago, 1921 - 1951, ACRL Publications in Librarianship No. 42 American Library Association: Chicago, 1982.

(9) Waples, D. People and Print Social Aspects Of Reading In The Depression University of Chicago Press: Chicago, 1938.

(10) Berelson, B. Library&rsquos Public Columbia University Press: New York, 1949.

(11) Berelson, B. The public library, book reading, and political behavior. Library Quarterly 1945, 15 (4), 281냳.

(12) ) Royal Society of London Scientific Information Conference: 1948. Report. Royal Society: London, 1948.

(13) Proceedings of the International Conference on Scientific Information Washington, D.C., Nov. 16ᆩ, 1958.: National Academy of Sciences, National Research Council: Washington, DC, 1959. 2 vol.

(14) American Psychological Association. Project on Scientific Information Exchange in Psychology, American Psychological Association: Washington, D.C.,1963ᇘ. 21 reports.

(15) Garvey, W.D. Griffith, B.C. Scientific communication as a social system. Science 1967, 157, 1011񮇈.

(16) Garvey, W.D. Communication in the physical and social sciences. Science 1970, 11, 1166񮉥.

(17) Price, D.J.d.S. Little Science, Big Science Columbia University Press: New York, 1963.

(18) Price, D.J.d.S. Networks of scientific papers. Science 1965, 149, 510뇋.

(19) Crane, D. Invisible Colleges: Diffusion of Knowledge in Scientific Communities University of Chicago Press: Chicago, 1972.

(20) Meadows, A.J. Communication in Science Butterworth: London, 1974.

(21) Davis, R.A. Bailey, C.A. Bibliography of Use Studies Drexel Institute of Technology, Graduate School of Library Science: Philadelphia, 1964.

(22) Menzel, H. Information needs and uses in science and technology. Annual Review of Information Science and Technology 1966, 1, 41ᇙ.

(23) Garvey, W.D. Communication: The Essence of Science: Facilitating Information Exchange Among Librarians, Scientists, Engineers, and Students Pergamon Press: New York, 1979.

(24) Paisley, W.J. Information needs and uses. Annual Review of Information Science and Technology 1968, 3, 1ᆲ.

(25) Poole, H. Theories of the Middle Range Ablex: Norwood, NJ, 1985.

(26) Warner, E.S., et al. Information Needs of Urban Residents Regional Planning Council and Westat Research, Inc.: Baltimore and Rockville, Maryland, 1973 (ERIC ED 088 464).

(27) Dervin, B. Development of Strategies for Dealing With the Information Needs of Urban Residents: Phase I--Citizen Study. Final Report University of Washington, Department of Communication: Seattle, Wash., 1976 (ERIC ED 125 640).

(28) Bundy, M.L. Metropolitan public library use. Wilson Library Bulletin 1967, 41, 950뎉.

(29) Bundy, M.L. Factors influencing public library use. Wilson Library Bulletin 1967, 42, 371녆.

(30) Bates, M.J. Information science at the University of California at Berkeley in the 1960s: a memoir of student days. Library Trends 2004, 52 (4), 683늅.

(31) Cuadra, C.A. Bates, M.J., Eds. Library and Information Service Needs of the Nation: Proceedings of a Conference on the Needs of Occupational, Ethnic, and other Groups in the United States. Sponsored by the National Commission on Libraries and Information Science, University of Denver, 1973 U.S. Government Printing Office: Washington, D.C., 1974.

(32) Streatfield, D.R. Wilson, T.D. Information innovations in social services departments: a third report on Project INISS. Journal of Documentation 1982, 38, 273냡.

(33) Blake, B. Morkham, T. Skinner, A. Inside information: social welfare practitioners and their information needs. Aslib Proceedings 1979, 31, 275냣.

(34) Brittain, J.M. Information services and the structure of knowledge in the social sciences. International Social Science Journal 1979, 31 (4), 711늠.

(35) Case, D.O. Conceptual organization and retrieval of text by historians: the role of memory and metaphor. Journal of the American Society for Information Science 1991, 42 (9), 657뉤.

(36) Chu, C.M. Literary critics at work and their information needs: a research-phases model. Library & Information Science Research 1999, 21 (2), 247냙.

(37) Cobbledick, S. Information-seeking behavior of artists: exploratory interviews. Library Quarterly 1996, 66, 343넼.

(38) Wiberley, S.E., Jr. Jones, W.G. Patterns of information seeking in the humanities. College & Research Libraries 1989, 50, 638뉍.

(39) Bates, M.J. Getty end-user online searching project in the humanities: report no.6: overview and conclusions. College & Research Libraries 1996, 57, 514녯.

(40) Westbrook, L. Information needs and experiences of scholars in women&rsquos studies: problems and solutions. College and Research Libraries 2003 64 (3), 192낙.

(41) Palmer, C.L., Ed. Library Trends 1996, 45 (2), 129넶.

(42) Palmer, C.L. Work at the Boundaries of Science: Information and the Interdisciplinary Research Process Kluwer Academic Publishers: Dordrecht Boston, 2001.

(43) Kari, J. Hartel, J. Information and higher things in life: addressing the pleasurable and the profound in information science. Journal of the American Society for Information Science and Technology 2007, 58 (8), 1131񮉋.

(44) Davies, K. The information-seeking behaviour of doctors: a review of the evidence. Health Information and Libraries Journal 2007, 24 (2), 78ᇲ.

(45) Sutton, S.A. Role of attorney mental models of law in case relevance determinations: an exploratory analysis. Journal of the American Society of Information Science 1994, 45, 186낐.

(46) Auster, E. Choo, C. W. Environmental scanning by CEOs in two Canadian industries. Journal of the American Society for Information Science 1993, 44 (4), 194낓.

(47) Summers, E.G. Matheson, J. Conry, R. The effect of personal, professional, and psychological attributes, and information seeking behavior on the use of information sources by educators. Journal of the American Society for Information Science 1983, 34 (1), 75ᇩ.

(48) Williams, D., Coles, L. Evidence-based practice in teaching: An information perspective. Journal of Documentation 2007, 63(6), 812댋.

(49) Wilson, T.D. On user studies and information needs. Journal of Documentation 1981, 37, 3ᆣ.

(50) Wilson, T.D. Models in information behaviour research. Journal of Documentation 1999, 55 (3), 249냖.

(51) Belkin, N.J. Oddy, R.N. Brooks, H.M. ASK for information retrieval: part I. background and theory. Journal of Documentation 1982, 38 (2), 61ᇛ.

(52) Kuhlthau, C.C. Inside the search process: Information seeking from the user&rsquos perspective. Journal of the American Society for Information Science, 42(5), 361넻.

(53) Bates, M.J. The design of browsing and berrypicking techniques for the online search interface. Online Review 1989, 13, 407녰.

(54) Dervin, B. Nilan, M. Information needs and uses. Annual Review of Information Science and Technology 1986, 21, 3ᆵ.

(55) (accessed June 2008).

(56) Dervin, B. Information as a user construct: the relevance of perceived information needs to synthesis and interpretation. In Knowledge Structure and Use: Implications for Synthesis and Interpretation Ward, S.A., Reed, L.J. Eds. Temple University Press: Philadelphia, 1983 161.

(57) Talja, S. Hartel, J. Revisiting the user-centered turn in information science research: an intellectual history perspective. Information Research 2007, 12 (4), paper colis04. (Available atž/colis/colis04.html)

(58) Talja, S. Tuominen, K. Savolainen, R. (2005). &rsquoIsms&rsquo in information science: constructivism, collectivism and constructionism. Journal of Documentation 2005, 61 (1), 79뀭.

isic2008/?page_id=4 (accessed June 2008).

(60) Fisher, K.E. Erdelez, S. McKechnie, L.(E.F.), Eds. Theories of Information Behavior. American Society for Information Science and Technology by Information Today: Medford, N.J., c2005.

(61) Bates, M.J. An Introduction to Metatheories, Theories, and Models. In Theories of Information Behavior Fisher, K.E. Erdelez, S. McKechnie, L.(E.F.), Eds. American Society for Information Science and Technology: Medford, N.J., c2005 1ᆬ.

(62) Ellis, D. Modeling the information-seeking patterns of academic researchers: a grounded theory approach. Library Quarterly 1993 63 (4), 469놮.

(63) Talja, S. Music, Culture, and the Library: An Analysis Of Discourses Scarecrow Press: Lanham, Md., c2001.

(64) Xu, Y. Liu, C. The dynamics of interactive information retrieval, part II: an empirical study from the activity theory perspective. Journal of the American Society for Information Science and Technology 2007, 58 (7), 987뎮.

(65) Reddy, M.C. Jansen, B.J. A model for understanding collaborative information behavior in context: a study of two healthcare teams. Information Processing and Management 2008, 44 (1), 256냙.

(66) Limberg, L. Experiencing information seeking and learning: a study of the interaction between two phenomena. Information Research 1999, 5 (1). (Available at:

( 67 Ford, N. Psychological determinants of information needs: a small-scale study of higher education students. Journal of Librarianship 1986, 18 (1), 47ᇒ.

(68) Ford, N. Towards a model of learning for educational informatics. Journal of Documentation 2004, 60 (2), 183납.

(69) Srinivasan, R. Pyati, A. Diasporic information environments: reframing immigrant-focused information research. Journal of the American Society for Information Science and Technology 2007, 58 (12), 1734񮒠.

(70) Fidel, R. Pejtersen, A.M. From information behaviour research to the design of information systems: the Cognitive Work Analysis framework. Information Research 2004, 10 (1), paper 210. (Available at

(71) Sandstrom, A.R. Sandstrom, P.E. The use and misuse of anthropological methods in library and information science research. Library Quarterly 1995, 65 (2), 161낏.

(72) Nicholas, D. Huntington, P. Jamali, H.R. Watkinson, A. The information seeking behaviour of the users of digital scholarly journals. Information Processing & Management 2006, 42 (5), 1345񮌥.

(73) Bates, M.J. Information and knowledge: an evolutionary framework for information science. Information Research 2005, 10 (4), paper 239. (Available at

(74) Bates, M.J. What is browsing—really? A model drawing from behavioural science research. Information Research 2007, 12 (4), paper 330. (Available at

(75) Case, D.O. Looking for information: a survey of research on information seeking, needs, and behavior, 2nd Ed. Elsevier/Academic Press: Amsterdam, Boston, 2007.

(76) Savolainen, R. Information behavior and information practice: reviewing the "umbrella concepts" of information-seeking studies. Library Quarterly 2007, 77 (2), 109끌.

(77) Savolainen, R. Information behavior and information practice: reviewing the "umbrella concepts" of information-seeking studies. Library Quarterly 2007, 77 (2), 109.

(78) Savolainen, R. Information behavior and information practice: reviewing the "umbrella concepts" of information-seeking studies. Library Quarterly 2007, 77 (2), 126.

(79) Bates, M.J. Information search tactics. Journal of the American Society for Information Science 1979, 30, 205낞.

(80) Bates, M.J. Search techniques. Annual Review of Information Science and Technology 1981, 16, 139끱.

(81) Cochrane, P.A. Markey, K. Catalog use studies--before and after the introduction of online interactive catalogs: impact on design for subject access. Library & Information Science Research 1983, 5 (4), 337넳.

(82) Fidel, R. Online searching styles: a case-study-based model of searching behavior. Journal of the American Society for Information Science 1984, 35 (4), 211낥.

(83) Hsieh-Yee, I. Effects of search experience and subject knowledge on online search behavior: measuring the search tactics of novice and experienced searchers. Journal of the American Society for Information Science 1993, 44, 161끶.

(84) Marchionini, G. Information Seeking in Electronic Environments Cambridge University Press: Cambridge, UK, 1995.

(85) Spink, A. Wolfram, D. Jansen, B.J. Saracevic, T. Searching the web: the public and their queries. Journal of the American Society for Information Science and Technology 2001, 52 (3), 226낲.

(86) O&rsquoConnor, B. Browsing: a framework for seeking functional information. Knowledge: Creativity, Diffusion, Utilization 1993, 15 (2), 211낰.

(87) Rice, R.E. McCreadie, M. Chang, S.L. Accessing and Browsing Information and Communication MIT Press: Cambridge, Mass., 2001.

(88) Hodges T. Bloch U. Fiche or film for COM catalogs - two use tests. Library Quarterly 1982, 52 (2), 131끘.

(89) Matthews, J.R. Lawrence, G.S. Ferguson, D.K., Eds. Using Online Catalogs: A Nationwide Survey. Neal-Schuman: New York, 1983.

(90) Lynch, C.A. the next generation of public access information retrieval systems for research libraries: lessons from ten years of the MELVYL system. Information Technology and Libraries 1992, 11, 405녧.

(91) Bates, M.J. Subject access in online catalogs: a design model. Journal of the American Society for Information Science 1986, 37, 357녀.

(92) Hildreth, C.R. Intelligent Interfaces and Retrieval Methods for Subject Searching in Bibliographic Retrieval Systems, Advances in Library Information Technology #2 Library of Congress Cataloging Distribution Service: Washington, D.C., 1989.

(93) Borgman, C.L. Why are online catalogs still hard to use? Journal of the American Society for Information Science 1996, 47, 493놿.

(94) Salton, G. McGill, J. M. Introduction To Modern Information Retrieval McGraw-Hill: New York, 1983.

(95) Voorhees, E.M. Harman, D.K., Eds. TREC – Experiment and Evaluation In Information Retrieval The MIT Press: Cambridge, Mass., 2005.

(96) Saracevic, T. Kantor, P. A study of information seeking and retrieving. III. Searchers, searches, and overlap. Journal of the American Society for Information Science 1988, 39, 197날.

(97) Fidel, R. Searchers&rsquo selection of search keys: III. Searching styles. Journal of the American Society for Information Science 1991, 42 (7), 515뇗.

(98) Cronin, B. Hert, C.A. Scholarly foraging and network discovery tools. Journal of Documentation 1995, 51 (4), 388녛.

(99) Xie, H.I. Supporting ease-of-use and user control: desired features and structure of web-based online IR systems. Information Processing & Management 2003, 39 (6), 899덢.

(100) Rieh, S.Y. On the web at home: information seeking and web searching in the home environment. Journal of the American Society for Information Science and Technology 2004, 55 (8), 743늹.

(101) Bishop, A.P. Van House, N.A. Buttenfield, B.P. Digital Library Use: Social Practice in Design and Evaluation MIT Press: Cambridge, Mass., 2003.

(102) Gilliland-Swetland, A.J. Kafai, Y. Landis, W.E. integrating primary sources into the elementary school classroom: a case study of teachers&rsquo perspectives. Archivaria 1999, 48, 89뀼.

(103) Marchionini. G. Crane, G. Evaluating hypermedia and learning: methods and results from the Perseus Project, ACM Transactions on Information Systems 1994, 12 (1), 5ᆶ.

(104) Roderer, N.K. Zambrowicz, C. Zhang, D. Zhou, H. User information seeking behavior in a medical web portal environment: a preliminary study. Journal of the American Society for Information Science and Technology 2004, 55 (8), 670뉴.

(105) Chen, C. Hernon, P. (1982). Information Seeking: Assessing and Anticipating User Needs Neal Schuman Publishers: New York, 1982.

(106) Menzel, H. Planned and unplanned scientific communication. Proceedings of the International Conference on Scientific Information, Washington, D.C., Nov 16ᆩ, 1958 National Academy of Sciences, National Research Council: Washington, DC, 1959 199낻.

(107) Dervin, B. Harpring, J.E. Foreman-Wernet. L. In moments of concern: a Sense-Making study of pregnant, drug-addicted women and their information needs. The Electronic Journal of Communication 1999, 9 (2dž). Available at

(108) Wilson, P. Second-Hand Knowledge: An Inquiry Into Cognitive Authority Greenwood Press: Westport, Conn., 1983.

(109) Hurd, J.M. Interdisciplinary research in the sciences: implications for library organization. College & Research Libraries 1992, 53(4), 283냱.

(110) Paisley, W. Information and work. In Progress in Communication Sciences Dervin, B., Voigt, M., Eds. Ablex: Norwood, N.J, 1980 2, 113끭.

(111) Taylor, R.S. Information use environments. In Progress in Communication Sciences Dervin, B., Voigt, M., Eds. Ablex: Norwood, N.J, 1991 10, 217냇.

(112) Cronin, B. Invisible colleges and information transfer: a review and commentary with particular reference to the social sciences. Journal of Documentation 1982, 38 (3), 212내.

(113) Leckie, G.J. Pettigrew, K.E. Sylvain, C. Modeling the information seeking of professionals: a general model derived from research on engineers, health care professionals, and lawyers. Library Quarterly 1996, 66 (2), 161낉.

(114) Budd, J.M. Research in the two cultures: the nature of scholarship in science and the humanities. Collection Management 1989, 11, 1ᆩ.

(115) Bates, M.J. The design of databases and other information resources for humanities scholars: The Getty Online Searching Project Report No. 4. Online and CD-ROM Review 1994, 18(6), 331넜.

(116) Orbach, B.C. The view from the researcher&rsquos desk: historians&rsquo perceptions of research and repositories. American Archivist 1991, 54(1), 28ᆿ.

(117) Duff, W.M. Johnson, C.A. Where is the list with all the names? Information-seeking behavior of genealogists. American Archivist 2003, 66 (1), 79ᇳ.

(118) Chatman, E. The information world of low-skilled workers. Library and Information Science Research 1987, 9, 265냣.

(119) Chatman, E.A. The Information World Of Retired Women Greenwood Press: New York, 1992.

(120) Chatman, E.A. A theory of life in the round. Journal of the American Society for Information Science 1999, 50 (3), 207ᆥ.

(121) Metoyer-Duran, C. Gatekeepers in Ethnolinguistic Communities Ablex: Norwood, N.J., 1993.

(122) Harris, R.M. Dewdney, P. Barriers to Information : How Formal Help Systems Fail Battered Women Greenwood Press: Westport, Conn., 1994.

(123) Hersberger, J.A. Murray, A.L. Sokoloff, S.M. The information use environment of abused and neglected children. Information Research 2006, 12 (1), paper 277. (Available at

(124) Kuhlthau, C.C. Seeking Meaning : A Process Approach To Library and Information Services, 2nd ed. Libraries Unlimited: Westport, Conn., 2004.

(125) Ellis, D. Behavioural approach to information retrieval system design. Journal of Documentation 1989, 45 (3), 171난.

(126) Cool, C. The concept of situation in information science. Annual Review of Information Science and Technology 2001, 35, 5ᆾ.

(127) Courtright, C. Context in information behavior research. Annual Review of Information Science and Technology 2007, 41, 273냺.

(128) Fisher, K.E. Durrance, J.C. Hinton, M.B. Information grounds and the use of need-based services by immigrants in Queens, New York: A context-based, outcome evaluation approach. Journal of the American Society for Information Science and Technology 2004, 55 (8), 754닆.

(129) Choo, C.W. Detlor, B. Turnbull, D. Web Work: Information Seeking and Knowledge Work on the World Wide Web Kluwer Academic: Boston, 2000.

(130) Walter, V. Information needs of children. Advances in Librarianship 1994, 18, 111끉.

(131) Silverstein, J. Just curious: children&rsquos use of digital reference for unimposed queries and its importance in informal education. Library Trends 2005, 54 (2), 228낼.

(132) Trace, C.B. Information creation and the notion of membership. Journal of Documentation 2007, 63 (1), 142끬.

(133) Kling, R. McKim, G. Not just a matter of time: field differences and the shaping of electronic media in supporting scientific communication. Journal of the American Society for Information Science 2000, 51 (14), 1306񮋸.

Conceptual Shifts in Ecology

Community-based conservation has emerged at a time when the science of ecology and the various fields of applied ecology seem to be in the midst of three conceptual shifts: a shift from reductionism to a systems view of the world, a shift to include humans in the ecosystem, and a shift from an expert-based approach to participatory conservation and management ( Levin 1999 Bradshaw & Bekoff 2001 Ludwig 2001 ). I expand on each.

Systems View of the Environment

A complex adaptive system often has a number of attributes not observed in simple systems, including nonlinearity, uncertainty, emergence, scale, and self-organization ( Levin 1999 Gunderson & Holling 2002 ). These characteristics of complex systems have a number of important implications for conservation and environmental management, as can be seen from a consideration of nonlinearity and scale.

The issue of nonlinearity comes up with respect to management institutions. The older, conventional emphasis on centralized institutions and command-and-control resource management is based on linear cause-effect thinking and mechanistic views of nature. It aims to reduce natural variation in an effort to make the ecosystem more productive, predictable, and controllable. But the reduction of the range of natural variation is the very process that may lead to a loss of resilience in a system, leaving it more susceptible to crises ( Holling & Meffe 1996 ). Nonlinearity is also an issue with respect to specific relationships and processes. Nonlinear effects have been documented in conservation biology—for example, in the interactions of elephants and people. The two coexist up to a certain threshold of human population density, beyond which elephants disappear ( Hoare & du Toit 1999 ).

The issue of scale has implications for the match between institutions and ecosystems and for perspectives that may be held by different agents. Take the question of match. Can a given conservation problem be managed by a centralized agency or are there more appropriate structures of governance in which the scale of management institution is matched to the scale of the ecosystem? Often, one-size-fits-all kinds of management ignore scale issues. Such mismatches of scale may be one of the key reasons for the failure of environmental management regimes ( Folke et al. 2002 ).

One of the insights from complexity thinking is that a multiplicity of scales prevents there being one “correct” perspective in a complex system. Phenomena at each level of the scale tend to have their own emergent properties. The system must be analyzed simultaneously at different scales. In biodiversity conservation, for example, different groups of conservationists focus on different levels of biological organization: they may use different research approaches and principles at the genetic, species, and landscape levels. All these levels are the “correct” level to consider at the same time. Similarly, a number of agents or actors may hold different but equally valid perspectives on a conservation problem. Redford and Sanderson (2000:1364) allude to this phenomenon: “they [forest peoples] may speak for their version of a forest, but they do not speak for the forest we want to conserve” (emphasis added).

Including Humans in the Ecosystem: Social-Ecological Systems

There is general agreement that we can ill afford to consider humans separately from nature, especially in today's heavily human-dominated world ( Kates et al. 2001 Gunderson & Holling 2002 ). It has become increasingly important to incorporate the dynamic interactions between societies and natural systems, rather than viewing people merely as “managers” or “stressors.” There is little agreement, however, on how this can be accomplished, conceptually or methodologically.

In our work, we use the term social-ecological system to refer to the integrated concept of humans in nature ( Berkes & Folke 1998 Berkes et al. 2003 ). A number of different terms are in use to denote the idea of humans as part of ecosystems. One of them is the dwelling perspective of Ingold (2000) , which refers to the “ … practical engagement of humans with others of the dwelt-in ecosystem.” This practical engagement, building knowledge and ecological relationships, is the basis for putting humans back into the ecosystem. It involves the “skills, sensitivities and orientations that have developed through long experience of conducting one's life in a particular environment” ( Ingold 2000 :25).

The social-ecological system has many levels. The links between social and environmental systems are different at the level of the community than they are at the level of the nation state. For example, Gibson's 1999 book, Politicians and Poachers, deals with the political economy of conservation in four African countries. It shows that the forces operating at the level of the nation state (many of them related to the peculiarities of postcolonial governments) are quite different from those at the levels of region and community.

Putting humans back into the ecosystem requires using all possible sources of ecological knowledge and understanding as may be available. Using knowledge and perspectives from the community level can help build a more complete information base than may be available from scientific studies alone ( Berkes et al. 2000 ). The partnership of local communities with scientists is not an unusual phenomenon. Details of such research collaboration, and its positive outcomes for ecosystem management, have been documented, for example, by Olsson and Folke (2001) , from Sweden and by Blann et al. (2003) from Minnesota (U.S.A.).

A word of caution is appropriate here. The term community in community-based conservation is gloss for a complex phenomenon because social systems are multiscale and the term community hides a great deal of complexity. Idealized “images of coherent, long-standing, localized sources of authority tied to what are assumed to be intrinsically sustainable resource management regimes” ( Brosius et al. 1998 :165) are just that—idealized. As many conservationists know, it is often difficult to find a cohesive social group to work with in the field. Communities are elusive and constantly changing. A community is not a static, isolated group of people. Rather, it is more useful to think of communities as multidimensional, cross-scale, social-political units or networks changing through time ( Carlsson 2000 ).

Hence, it is more productive to focus not on communities but on institutions, defined as the set of rules actually used, the working rules, or rules-in-use ( Ostrom 1990 ). Institutions are humanly devised constraints that structure human interaction, made up of formal constraints (rules, laws, constitutions), informal constraints (norms of behavior, conventions, and self-imposed codes of conduct), and their enforcement characteristics. I have specifically examined those institutions that mediate between social and ecological systems ( Berkes & Folke 1998 ) and have focused on the dynamics of these institutions: their renewal and reorganization, learning and adaptation, and ability to deal with change ( Berkes et al. 2003 ).

End of Management by the Expert-Based Approach?

Many of our environmental problems, including those related to conservation, do not lend themselves to analysis by the conventional, rational approach of defining the problem, collecting data, analyzing data, and making decisions based on the results. There is too much uncertainty targets keep shifting, and the issues must often be redefined ( Kates et al. 2001 ).

These make a class of problems that Ludwig (2001) and others have called “wicked problems,” those with “no definitive formulation, no stopping rule, and no test for a solution,” problems that cannot be separated from issues of values, equity, and social justice. Ludwig (2001) argues that where there are no clearly defined objectives and where there are diverse, mutually contradictory approaches, the notion of an objective, disinterested expert no longer makes sense. Hence, a new kind of approach to science and management must be created through a process by which researchers and stakeholders interact to define important questions, objectives of study, relevant evidence, and convincing forms of argument. This kind of research, referred to by Kates et al. (2001) as sustainability science, requires place-based models because understanding the dynamic interaction between nature and society requires case studies situated in particular places.

To deal with the implications of complex systems, working partnerships can be built between managers and resource users. This is done, for example, in adaptive management, which recognizes, as a starting point, that information will never be perfect ( Holling 2001 ). The use of imperfect information for management necessitates a close cooperation and risk-sharing between the management agency and local people. Such a process requires collaboration, transparency, and accountability so that a learning environment can be created and practice can build on experience. This approach, bringing the community actively into the management process, is fundamentally different from the command-and-control style.

These three conceptual shifts in ecology—toward a systems view, inclusion of humans in the ecosystem, and management by participatory approaches—are related. They all pertain to an emerging understanding of ecosystems as complex adaptive systems in which human societies are necessarily an integral part. We can abandon Enlightenment assumptions of predictability and control we should at least be very skeptical of them. We need to recognize the limits of expertise and the advantages of participatory conservation and management. Along with the conventional biological science of conservation, there is an emerging social science of conservation that may provide a more nuanced understanding of social systems. To move toward an interdisciplinary science of conservation, we need to learn from the lessons emerging from several new interdisciplinary fields.

Why Behavioral Economics Needs Behavioral Biology

Following my first sabbatical—a two-year stint as a federal bureaucrat in Washington, DC—I resumed my teaching duties in Cornell’s department of economics in the fall of 1980. Shortly thereafter, I met Richard Thaler, who had started teaching economics in the university’s business school while I‘d been away. Over the next several years, he and I spent long hours in conversation about how our own observations of people’s behavior were often strikingly at odds with the predictions of standard economic theory.

Thaler had spent his own recent first sabbatical working with the Israeli psychologists Daniel Kahneman and Amos Tversky, both renowned for their pioneering work on systematic cognitive errors. Thaler’s 1980 article, “Toward a Positive Theory of Consumer Choice,” 1 which drew on that work, is now widely viewed as the paper that launched behavioral economics, a vibrant new field that has focused largely on the intersection of cognitive psychology and economics. In October of 2017, Thaler was awarded the Nobel Prize in economics in recognition of that work.

In 1983, I taught the first undergraduate course ever offered in behavioral economics. Because few students had ever heard the term, my first challenge was to come up with a course title that might lure some to enroll. In the end, I decided to call it “Departures from Rational Choice” (a decision I later regretted, because it triggered fruitless debates over the meaning of rationality). Naturally, there was no standard syllabus then. After much deliberation, I decided to cover material under two broad headings: “Departures from Rational Choice with Regret,” and “Departures from Rational Choice without Regret.”

Under the first heading, I listed studies that document the many systematic cognitive errors to which most of us are prone. For example, although standard rational choice models say that people will ignore sunk costs (costs that are beyond recovery at the moment of decision), such costs often influence choices in conspicuous ways. In one of Thaler’s celebrated examples, you’re about to depart for a sporting event at an arena 50 miles away when an unexpectedly heavy snowstorm begins. If your ticket is nonrefundable, your decision of whether to stay home should not be influenced by the amount you paid for it. Yet a fan who paid $100 for his ticket is significantly more likely to make the dangerous drive than an equally avid fan who happened to receive his ticket for free. The first fan is probably guilty of a cognitive error. People typically seem to regret making such errors once they become aware of them.

Under my “…Without Regret” heading, I listed studies that describe departures from the predictions of standard rational choice models that people do not seem to regret. Consider an MBA student’s decision of how much to spend on an interview suit. The standard assumption in these models is that the primary determinants of the satisfaction provided by any good are its absolute attributes, but that’s clearly not true of the utility provided by an interview suit. If you’re one of several similarly qualified applicants who all want the same investment banking job, it’s strongly in your interest to look good when you show up for your interview. But looking good is an inherently relative concept. It means looking better than other candidates. If they show up wearing $500 suits, you’ll be more likely to make a favorable first impression, and more likely to get a callback, if you show up in a $3,000 suit than if you show up in one costing only $300. Spending more is rational from the individual job seeker’s perspective, but irrational from the perspective of job seekers as a group.

Behavioral economics as it developed over the next decades did not follow the roadmap outlined in my syllabus. Instead, it has focused almost exclusively on behavior under my first heading, departures with regret. This work on cognitive errors has had an enormous impact on policy makers. Governments around the globe have been inspired by it to set up behavioral science advisory groups, popularly known as nudge units, to help citizens make better decisions. A UK study found that implementation of the British group’s recommendations had produced savings that exceeded the group’s costs by a factor of 20. 2

Work that falls under my departures-without-regret heading has been far less extensive—so much so that an instructor putting together a syllabus for a behavioral economics course today might find the presence of that heading on my early-1980s syllabus somewhat puzzling.

I continue to believe, however, that economic losses under the without-regret heading are larger by several orders of magnitude than those under the with-regret heading. Losses from without-regret departures are also substantially more stubborn—because, unlike losses resulting from cognitive errors, they cannot be remedied by unilateral individual action. Once someone learns that it is a mistake to take sunk costs into account, for example, it becomes possible to ignore them unilaterally. Collaboration with others isn’t required. But collective-action problems are a different matter. It is one thing for job candidates to recognize that all would be better off if everyone spent less on interview suits. But absent an enforceable agreement for all to cut back in tandem, each candidate’s best bet is to continue spending.

As behavioral economics continues to evolve, it would profit from adopting an even broader interdisciplinary perspective, drawing on the insights not just of economics and psychology, but also those of evolutionary biology. Traditional models of rational choice typically ignore what I view as Charles Darwin’s central insight—that life is graded on the curve. It’s not how strong, fast, or clever we are that matters, but rather how those traits compare with those of rivals. When context influences our ability to achieve important goals, as in the interview suit example, all bets regarding the efficacy of Adam Smith’s invisible hand are off. Notwithstanding the uncritically enthusiastic pronouncements of many of Smith’s modern disciples, unbridled market forces often fail to channel the behavior of self-interested individuals for the common good. On the contrary, as Darwin saw clearly, individual incentives often lead to wasteful arms races.

The losses from these arms races are often epic in scale. One of the most robust findings from the large and contentious literature on the determinants of human well-being is that beyond a certain point—one that has long since been passed in the developed world—across-the-board increases in many forms of private consumption produce no measurable gains in health or life satisfaction. 3 If all mansions were to double in size, for example, those living in them would be neither happier nor healthier than before. Existing research thus does not permit us to conclude with confidence that Americans were meaningfully better off in 2018 than in 2012, even though the inflation-adjusted total value of the nation’s goods and services was more than $2 trillion higher in 2018.

Now imagine that someone had possessed a magic wand that could have rearranged our 2012 spending patterns—say, by making the largest houses somewhat smaller, cutting expenditures on automobiles and interview suits, and reducing outlays on wedding receptions, coming-of-age parties, and the like. The resulting savings could have been spent to shorten the workweek by a few hours and provide an additional two weeks of vacation time for everyone. And more could have been spent to repair our decaying infrastructure.

Existing evidence leaves little doubt that expenditure shifts of this sort would have caused clear gains in well-being for 2012 Americans, gains that would enable us to say that Americans living in this rearranged version of 2012 would have been happier and healthier than actual Americans in 2018, even though those in the first group were $2 trillion poorer. Unless we are willing to deny the validity of that evidence, the clear implication is that our current spending patterns are wasting at least $2 trillion annually in the United States alone, a sum that almost surely dwarfs the losses caused by cognitive errors.

The things we buy are of course not the only choices that are influenced by others. As psychologists have long said, “It’s the situation, not the person.” When we see someone behave in a particular way, our impulse is to ask what sort of person would do such a thing. Psychologists say that’s the wrong way to think about it. The behavior we observe is more often driven by the social forces surrounding the actor than by traits of character and personality. Sometimes those forces influence us for ill, as when people who smoke lead their friends to take up the habit. But social influence can also be positive, as when time spent with friends who eat prudently and exercise regularly makes people more likely to adopt healthful lifestyles.

Get Evonomics in your inbox

All this is uncontroversial. Also uncontroversial is that the causal arrows run in both directions—that the social environment is itself a consequence, in the aggregate, of our own choices. The smoking rate, for example, is just the number of us who smoke divided by our total number. But because the effect of any individual choice on the social environment is minuscule, people typically ignore that second causal pathway. Few express concern, for example, that their decision to smoke might make others more likely to do so.

Are there simple policy measures that would encourage us to consider how our choices affect social environments? Cigarette taxation provides an informative case study. Because nicotine is one of the most highly addictive substances, even large increases in the price of tobacco products typically produce only small reductions in consumption for those already addicted. And the initial declines in smoking rates were indeed modest in the 1980s when American regulators began imposing significant taxes on cigarettes and restrictions on where people could smoke. Yet a small proportion of smokers did quit soon after the taxes began, and higher prices also induced a modest number of others to refrain from starting. And because smoking is highly contagious, those initial responses launched a powerful dynamic. They induced still others to quit or refrain from starting, and in every succeeding year, rates continued to fall. The cumulative effect of these responses was dramatic: The adult smoking rate in the US is now less than one-third of what it was in the mid-1960s.

Yet regulators did not invoke behavioral contagion as a rationale for cigarette taxation. Rather, they defended their restrictive measures by citing recent studies showing that exposure to second-hand smoke increased the incidence of serious illnesses among hapless bystanders. But the harm from such exposure, although real, is minuscule compared to the harm from actually being a smoker. By far the greater harm caused when someone becomes a smoker is the injury suffered by others thereby influenced to take up the habit. It’s a huge effect. One study estimated, for example, that when the proportion of smokers among a teen’s friends rose from 20 to 30 percent, she became 25 percent more likely to become or remain a smoker.

I use the term “behavioral externalities” to describe choices that affect social environments in these ways. Because social environments influence us so profoundly, both for good and ill, we have a powerful and legitimate interest in them. We would prefer to live in ones that bring out the best in us and to avoid those that harm our interests. Yet behavioral externalities have received virtually no serious attention from policy analysts, and it’s here that lie many of the most exciting opportunities for young researchers. Once you’ve been alerted to their existence, it quickly becomes apparent that behavioral externalities are ubiquitous.

Careful empirical studies have documented the importance of behavioral contagion in such diverse domains as, among many others, excessive drinking, sexual predation, cheating, bullying, obesity, greenhouse gas emissions, and compliance with public health directives. Research has tended to focus on negative peer influences, but there is also compelling evidence of positive influences. The adoption of rooftop solar panels, hybrid cars, and plant-based diets, for example, have all been shown to be highly contagious. 4

By its focus on systematic cognitive errors, behavioral economics has greatly increased our understanding of why people’s choices often fail to match those predicted by traditional rational choice models. But our failure to take full advantage of existing opportunities owes less to such errors than to our embeddedness in complex social structures. As Darwin understood clearly, our fate depends not only on our own decisions and capabilities but also on those of rivals and partners. And that, in a nutshell, is the case for a broader and more inclusive behavioral economics, one that incorporates the rich insights of behavioral biology.

Originally published at This View of Life. Read the full Advice to An Aspiring Economist series.

[1] Richard H. Thaler, “Toward A Positive Theory of Consumer Choice,” Journal of Economic Behavior and Organization, 1 (1), March 1980: 39-60.

[3] For an extensive summary of that literature, see chapters 5 and 6 of Robert H. Frank, Luxury Fever, NY: The Free Press, 1999.

[4] For a review of the relevant studies, see Robert H. Frank, Under the Influence, Princeton: Princeton University Press, 2020.

Donating = Changing Economics. And Changing the World.

Evonomics is free, it’s a labor of love, and it's an expense. We spend hundreds of hours and lots of dollars each month creating, curating, and promoting content that drives the next evolution of economics. If you're like us — if you think there’s a key leverage point here for making the world a better place — please consider donating. We’ll use your donation to deliver even more game-changing content, and to spread the word about that content to influential thinkers far and wide.

$3 / month
$7 / month
$10 / month
$25 / month

You can also become a one-time patron with a single donation in any amount.

Biology Teachers Designing Context-Based Lessons for Their Classroom Practice—The importance of rules-of-thumb

In science education in the Netherlands new, context‐based, curricula are being developed. As in any innovation, the outcome will largely depend on the teachers who design and implement lessons. Central to the study presented here is the idea that teachers, when designing lessons, use rules‐of‐thumb: notions of what a lesson should look like if certain classroom outcomes are to be reached. Our study aimed at (1) identifying the rules‐of‐thumb biology teachers use when designing context‐based lessons for their own classroom practice, and (2) assessing how these personal rules‐of‐thumb relate to formal innovative goals and lesson characteristics. Six biology teachers with varying backgrounds designed and implemented a lesson or series of lessons for their own practice, while thinking aloud. We interviewed the teachers and observed their lessons. Our results suggest that rules‐of‐thumb, which differed substantially among the teachers, indeed to a great extent guide the decisions teachers make when designing (innovative) lessons. These rules‐of‐thumb were often strongly associated with intended lesson outcomes. Also, teachers’ personal rules‐of‐thumb were more powerful in determining the lesson design than formal innovative goals and lesson characteristics. The results of this study encourage more research into how rules‐of‐thumb reflect teachers’ practical knowledge, for which suggestions are made.