study,estimate_as_printed,uncertainty_as_printed,sample_as_printed,definition,locator,harmonised_estimate,harmonised_se,harmonisable,poolable,transformation,reason_if_not
S1,"-0.24. (ExpCon coefficient, Model 2, LogToT; marker '.' = p<.1); text: beta_1=-0.24, 'about 21%' faster","SE (0.14); 95% CI = [-0.51, 0.03]; p=0.086","n = 96 (Table III caption); Section VI-B: N=47 (AI), N=46 (no AI)","OLS coefficient of the experimental-condition (AI) indicator on natural-log time on task, with developer-level covariates (Model 2); RCT, one complex enterprise coding task","arxiv-2410.12944v3: Table III, row ExpCon, column (2) 'Dev' (+ SE line below); Section VI-C (CI and p for Model 2)",-0.24,0.14,True,True,"log-time regression coefficient: theta = b = -0.24; SE(theta) = SE(b) = 0.14 (printed CI gives (0.03-(-0.51))/(2*1.96) = 0.1378, consistent)",
S2,+19% ('allowing AI actually increases completion time by 19%'; 'implement issues in 19% more time on average'),"none printed in the extract (95% CI shown only in Figure 1, not reproduced)","16 developers, 246 tasks; 136 issues AI-allowed, 110 issues AI-disallowed","S = E[time AI allowed]/E[time AI disallowed] - 1, estimated as exp(beta)-1 from OLS of log implementation time on AI indicator and log forecast time (Sections 2.3, D.1-D.3); issue-level RCT","arxiv-2507.09089v2: Abstract and Section 3 'Results' opening paragraph (estimator Sections 2.3, D.1, D.3)",0.1739533071,,True,False,percentage change: theta = ln(1 + 0.19) = 0.173953,"Not poolable: no SE, CI or p-value for the headline estimate is printed in the extract; its CI appears only in Figure 1 (and Figure 15), which is not reproduced, so SE(theta) cannot be obtained."
S3,"Treatment 44.29 min, Control 63.10 min, Overall 53.84 min, Improvement -29.79%***",significance stars only: *** (defined as 0<p<=0.01); no SE or CI printed,"181 subjects (randomised half to treatment, half to control)","percentage difference in mean task duration (minutes) for Copilot vs control, holding accuracy constant (control duration predicted at treatment accuracy, Appendix B); RCT","arxiv-2411.01067v2: Table 3, block 'Aggregate Results', row 'All Content', columns Treatment/Control/Improvement",-0.353679435,,True,False,percentage change: theta = ln(1 + (-0.2979)) = -0.353679 (ratio of printed durations ln(44.29/63.10) = -0.353962),"Not poolable: the only uncertainty information is significance stars (***, p<=0.01); SE(theta) cannot be obtained."
S4,"Aware 50.5 s, Control 56.9 s per email; % Time Savings 11.1",p-value 0.471,"Aware N=56, Control N=52 (Blind N=59 not used); 167 analysts in total","percentage time saving per email (full 25-email queue) of the Aware arm (agent verdicts, rationale, agent-prioritised queue) vs Control (no AI); RCT","arxiv-2511.13860v1: Section 3.3, Table 3, column 'Full Corpus', rows Aware/Control time, % Time Savings, p-value",-0.1176580435,0.163220592,True,True,percentage change c = -0.111 (11.1% saving): theta = ln(1 - 0.111) = -0.117658; APPROXIMATE SE from two-sided p: SE = |theta|/z(1-0.471/2) = 0.117658/0.720853 = 0.163221 (ratio of printed times ln(50.5/56.9) = -0.119322),
S5,"-0.278 (Treat x Post, Output, Panel A column (3))",SE (0.295) (clustered at user level); not marked significant,N 18332,"difference-in-differences coefficient of Italy x post-ban on GitHub Output (count of push, pull requests, PR comments, commit comments, create, issues events; users with affiliation), Pre 27/28.03 vs Post 03/04.04 2023","arxiv-2304.09339v1: Table 1, panel A, row Treat x Post (+ SE line), column (3) Output; row N",,,False,False,none (not harmonisable),"Not harmonisable: the effect is on output counts (GitHub events) around the withdrawal of ChatGPT access in a non-randomised difference-in-differences design, not a comparison of task-completion times with vs without AI on the same tasks."