# Contract — T3-synthesis (P0)

**Question.** Does access to a generative-AI assistant shorten the time workers need to complete tasks? Synthesise five primary studies (extracts only).

**Target / estimand.** θ = ln(T_AI / T_noAI): log ratio of mean completion time with AI to without AI, same tasks, randomised comparison; negative = faster with AI.
Estimands: (i) the five printed values verbatim with locators; (ii) harmonised θ and SE(θ) per study when obtainable under the TASK conventions;
(iii) inverse-variance FE and DerSimonian–Laird RE pooled θ over poolable studies, with 95% CIs (1.96), Q (df=k−1), I², τ²; (iv) the synthesis conclusion.
Meaningful under: each θ estimates the same kind of quantity (log time ratio), independent studies, normal approximation; with k=2 τ² is essentially unidentified.

**Conventions.** log-time coefficient b → θ=b, SE=SE(b) (or CI width/3.92); percentage change c → θ=ln(1+c); p-value only → SE≈|θ|/z_{1−p/2} (approximate, flag);
stars/thresholds/figure-only CI → harmonisable, not poolable; effect not on completion time AI-vs-no-AI → not harmonisable.

**Outputs.** out/table1.csv (12 named columns, 5 rows); out/figure1.png (forest plot, harmonisable studies, no-SE as bare points, RE diamond);
out/results.json keys: task_id, target_definition, items{S1..S5 with estimate_as_printed, uncertainty_as_printed, sample_as_printed, definition, locator, harmonisable, harmonised_estimate, harmonised_se, poolable, transformation, reason_if_not}, k_extracted, k_harmonised, k_pooled, fe_estimate, fe_se, fe_ci_low, fe_ci_high, re_estimate, re_se, re_ci_low, re_ci_high, q_stat, i2, tau2, heterogeneity_interpretation, pooled_significant_5pct, pooled_pct_change_re, feasibility, feasibility_reason, data_issues, n_input_rows, n_analysed, random_seed;
out/report.md (Question; Data and exclusions; Method; Results; Uncertainty; Limitations and open issues); out/analysis.py (regenerates results/table/figure, prints source lines used); out/claims.json.

**Information rules.** Only the extracts are evidence; nothing remembered from the full papers (e.g. the METR CI shown in Figure 1 is NOT available). Printed numbers copied exactly.

**Invalidity / impossibility.** Fewer than 1 poolable study would make pooled keys impossible (they are non-nullable). Using non-printed numbers would invalidate. Pooling a non-time outcome (S5) would be invalid.