Mean prediction accuracy score across all participants
Number of study ratings submitted by all participants
Total number of different users who have submitted ratings
Users currently rating studies (active in last 60 seconds)
Completed the workshop? Enter your username below to view your personalised results page.
How well did participants predict what others would rate each study?
Top performers ranked by prediction accuracy. Scores show how well participants predicted crowd ratings. The leaderboard includes all participants who have submitted at least one rating. Click any entry to view that participant's personalised results page.
Average percentage of tokens assigned to each criterion by participants, reflecting what they considered most important when evaluating study quality. Error bars show 95% confidence intervals for the mean percentage across participants.
No data yet.
How did participants actually rate the quality of each study?
Bars show the mean quality rating that participants gave each study (1-7 scale). Error bars show the 95% confidence interval for that mean, computed from participant ratings only with the t distribution. An interval is wide when participants disagree or when few have rated the study, and none is drawn for a study with fewer than two ratings. Where an interval runs off the chart, it stops at the edge without an end cap, and hovering over the bar shows its full limits. The diamond marks the expert rating, which is not included in the bar. (The game scores predictions against a different average, which counts the expert rating as 100 participants.) All participants with at least one valid rating are included. The leaderboard shows the subset who completed the full workshop. Click on any bar to view detailed statistics and study information.
Complete breakdown of how participants rated each study's quality. Every statistic uses participant ratings only, and the expert rating is listed in its own column. 95% CI is the confidence interval for the mean. Std Dev is the sample standard deviation and measures disagreement (higher = more variable ratings). Agreement % shows consensus level. A dash means too few ratings to estimate the value. Color scales highlight relative values. Click study names to view detailed information. Click column headers to sort.
| Study | Quality | Responses | Mean Rating | 95% CI | Std Dev | Min | Max | Agreement % | Expert Rating |
|---|---|---|---|---|---|---|---|---|---|
| Loading data... | |||||||||
Author: Dr Pablo Bernabeu, Department of Education, University of Oxford
Legal disclaimer: This app and workshop were created by Dr Pablo Bernabeu in a personal capacity during spare time. The employer is not affiliated with, does not endorse, and bears no liability for this app or workshop.
Materials: github.com/pablobernabeu/Unlock_the_Lab
Licence: CC BY 4.0 â Free to use with attribution
đĄ Contributions welcome! Suggest improvements or report issues on GitHub.