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. (beta_1 for ExpCon, Model 2, Table III; Section VI-C: beta_1=-0.24, i.e. 21% faster)","SE (0.14); 95% CI = [-0.51, 0.03], p=0.086 (NS)","n = 96 (Table III caption); Section VI-B: N=47 AI, N=46 no-AI","OLS coefficient on experimental condition (AI exposure) with natural-log time on task (LogToT) as dependent variable, developer-level covariates (Model 2), one enterprise coding task, randomised","arxiv-2410.12944v3, Section VI-C Table III, row ExpCon, column (2) 'Dev' (second value on collapsed line L76; SE line L78); CI and p in Section VI-C text",-0.24,0.14,True,True,log-time coefficient: theta = b = -0.24; SE(theta) = SE(b) = 0.14 (printed). Check: SE from CI = (0.03-(-0.51))/(2*1.96) = 0.1378,
S2,"+19% (allowing AI increases completion time by 19%; regression-based, Section 3 / Section D.1)","none printed (95% CI shown only in Figure 1 / Figure 15, not reproduced in the extract)","16 developers, 246 tasks; 136 issues AI-allowed, 110 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; issue-level randomisation","arxiv-2507.09089v2, Abstract and Section 3 'Results' (opening paragraph); estimator Sections 2.3, D.1-D.3",0.17395330712343798,,True,False,percentage change in time: theta = ln(1 + c) = ln(1 + 0.19) = 0.173953,"No SE, CI or p-value printed in the extract; the CI exists only in a figure that is not reproduced, so SE(theta) cannot be obtained (harmonisable, not poolable)."
S3,"Overall 53.84, Treatment 44.29, Control 63.10 (minutes); Improvement -29.79%***",significance stars only: *** (paper's definition 0<p<=0.01); no SE or CI printed,"181 subjects (randomised half treatment, half control)","difference in mean task duration (minutes) between Copilot and control groups across all tasks, holding accuracy constant (control duration predicted at treatment accuracy, Appendix B), randomised","arxiv-2411.01067v2, Section 4.2 Table 3, 'Aggregate Results' block, row 'All Content', columns Treatment/Control/Improvement",-0.353679434958934,,True,False,percentage change in time: theta = ln(1 + c) = ln(1 + (-0.2979)) = -0.353679 (cf. ln(44.29/63.10) = -0.353962),"Only significance stars (p<=0.01) are printed; a p-value threshold does not identify SE(theta) (harmonisable, not poolable)."
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 (167 analysts across three arms incl. Blind N=59)","percentage time saving in mean time per email over the full 25-email queue, Aware (agent verdicts/rationale + prioritised queue) vs Control (no AI), randomised arms","arxiv-2511.13860v1, Section 3.3 Table 3 'Comparison of Email Processing Times', column 'Full Corpus', rows Aware/Control time, % Time Savings, p-value",-0.11765804346823246,0.1632205919720678,True,True,percentage change in time c = -11.1/100 = -0.111; theta = ln(1 + c) = -0.117658; approximate SE from two-sided p: SE = |theta|/z(1-0.471/2) = 0.117658/0.720853 = 0.163221 (approximate SE),
S5,"-0.278 (Treat x Post, Output, column 3, panel A)",SE (0.295) (clustered at user level); no significance star,"N = 18332 (user-day observations, users with affiliation)","difference-in-differences coefficient on aggregate GitHub Output measure (sum of Push, PR, PR comments, commit comments, Create, Issues events) for Italian vs Austrian/French users, pre 27/28.03 vs post 03/04.04.2023 around Italy's ChatGPT ban","arxiv-2304.09339v1, Section 4 Table 1, panel A, row Treat x Post, column (3) Output (Users with Affiliation)",,,False,False,none applied (not harmonisable),"Effect on output counts (GitHub events) from a DiD around withdrawal of access, not a comparison of task-completion times with vs without AI on the same tasks, and not randomised."