Neuroscientists’ Domain Knowledge Does Not Improve Search Performance in PubMed
Bibliographic record
Abstract
A Review of: Vibert, N., Ros, C., Le Bigot, L., Ramond, M., Gatefin, J., & Rouet, J.-F. (2009). Effects of domain knowledge on reference search with the PubMed database: An experimental study. Journal of the American Society for Information Science and Technology, 60(7), 1423-1447. Objective – To determine whether neuroscientists and life scientists’ domain knowledge affects their search performance in PubMed for neuroscience topics. Design – Cross-sectional experimental study. Setting – State-funded research laboratories in the cities of Paris, Bordeaux and Poitiers, France. Subjects – There were 32 participants in the study: 16 neuroscientists and 16 life scientists with no experience in neuroscience. Both groups were similar in terms of age, gender, occupation, and online database search experience. Methods – All participants were asked to complete the same five tasks in PubMed to assess their search performance with this database. Each task consisted of finding and selecting bibliographic references on a neuroscience topic within 15 minutes. The instructions for these tasks were hidden from view during the search process. Participants performed the tasks on their office computers between May 2005 and June 2006 in the presence of one researcher who prompted them to verbally describe what they were doing and thinking as they searched. Each participant also filled out a questionnaire about their personal characteristics at the beginning of the search session and completed a second questionnaire about their knowledge of PubMed at the end. The entire experimental procedure lasted between 60 and 90 minutes and was recorded. The relevancy of the bibliographic references selected was later scored by two neuroscientists who did not participate in the study. The data were analyzed using multivariate analysis of variance (MANOVA) and qualitative analysis of verbal protocols. Main Results – The MANOVA analysis did not show any significant differences between the total scores obtained by the neuroscientists and the life scientists. Both groups were able to find relevant PubMed references for each task within the time allotted. Contrary to the researchers’ first main hypothesis, the neuroscientists’ domain knowledge did not result in a superior search performance (i.e., in less time spent searching and more relevant results) compared to that of the life scientists. However, domain knowledge did affect the method of searching, confirming the researchers’ second hypothesis. The life scientists spent more time reading the instructions for each task, included more keywords in their search queries, and opened more abstracts to select relevant references than the neuroscientists. The life scientists also used keywords that were almost exclusively taken from the instructions for each task when they searched PubMed and made significantly more mistakes than the neuroscientists. Furthermore, the participants’ knowledge of PubMed was poor as was expected, despite stating they used it very frequently. Half of the participants did not attempt to use limits even when the task called for it. The majority only used PubMed in the most basic way, that is, by typing keywords in the search box. Conclusion – Domain knowledge affects how end users search PubMed for topics in their specialty, but it does not impact their performance. Both the neuroscientists and the life scientists successfully completed the search tasks on neuroscience topics within the allotted time. Both groups had basic knowledge of PubMed, but were satisfied with their performance and results. The authors suggest that scientists would only be interested in attending a PubMed training session if they are convinced that they will learn how to search more quickly. Further experiments are needed to verify the effects of domain knowledge on search performance with topics that are more general. The search tasks used in this study were very specific, which may have positively influenced the performance of all participants. A different control group that shares less basic domain knowledge with the neuroscientists, such as mathematicians or chemists, may also be tested.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.006 | 0.078 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.003 | 0.004 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.008 | 0.002 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".