Challenges and Solutions for Data Analysis in an Adult Lifespan Study of over 100,000 Online Cognitive Test Completions
Bibliographic record
Abstract
Note: This pre-print includes the accepted version of a manuscript that was published in Sage Research Methods: Doing Research Online in 2022 (https://doi.org/10.4135/9781529600). In this case study, I describe methodological insights from data analysis of an online adult lifespan dataset (over 100,000 completions, ages 15-100). The data were used to study cross-sectional age differences in cognitive performance. I cover the steps of data analysis for large-scale web-based data, namely data cleaning, analysis, and visualization techniques. In each step, I describe the unique challenges that face analysis of data collected online, and potential solutions to address them, by drawing on practical lessons and examples from this study. First, I address how to identify problematic recordings such as technical issues (incomplete data, multiple completions by the same person, etc.), unreliable self-reported demographic information (age), and cognitive task outliers (accuracy, response times). I propose rigorous data cleaning as an essential first step to ensure that analytical conclusions are reliable and unbiased. Next, I demonstrate data visualization techniques that are better suited to large online datasets than more conventional techniques (e.g., density plots or locally weighted scatterplot smoothing instead of dot-plots or linear regression). Lastly, I cover the limitations of significance testing in large online datasets, and the value of complementary approaches such as data visualization, effect size estimation, and use of parsimony criteria. I also discuss more sophisticated analysis options enabled by large online datasets, such as non-linear regression, model comparison and selection, data resampling, and addition of covariates.
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 distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.001 | 0.004 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
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 teacher head, 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".