academic_recompute_statistics
Recompute reported statistical results—p-values, effect sizes, confidence intervals, power, and sample sizes—to verify consistency and flag mismatches in research papers.
Instructions
Recompute t/F/chi2/z p-values, two-group Cohen's d and Hedges' g, 2x2 OR/RR, CI consistency, statistical power, SD bounds, and df-implied sample sizes. Optional reported p-value comparison uses decimal round-half-up intervals; input or partial execution failures are tool errors, statistical mismatches are successful results.
Tool call template
This tool declares no required parameters. Add optional parameters from the input schema as needed.
{
"id": 1,
"jsonrpc": "2.0",
"method": "tools/call",
"params": {
"arguments": {},
"name": "academic_recompute_statistics"
}
}Input Schema
| Name | Required | Description | Default |
|---|---|---|---|
| df | No | Degrees of freedom (integer or Welch fractional) | |
| n1 | No | Group 1 size | |
| n2 | No | Group 2 size | |
| se | No | Reported standard error SE (> 0) | |
| df1 | No | Numerator degrees of freedom for F-test | |
| df2 | No | Denominator degrees of freedom for F-test | |
| sd1 | No | Group 1 SD | |
| sd2 | No | Group 2 SD | |
| beta | No | Reported regression coefficient beta | |
| alpha | No | Significance level alpha in (0, 1), default 0.05 | |
| mean1 | No | Group 1 mean | |
| mean2 | No | Group 2 mean | |
| cell_a | No | 2x2 table exposed/treated event count | |
| cell_b | No | 2x2 table exposed/treated non-event count | |
| cell_c | No | 2x2 table control event count | |
| cell_d | No | 2x2 table control non-event count | |
| f_stat | No | Reported F-statistic (right-tail ANOVA/regression) | |
| t_stat | No | Reported t-statistic | |
| chi2_df | No | Degrees of freedom for chi-square test | |
| df_kind | No | Test family for df-to-N check | |
| p_value | No | Reported p-value | |
| check_sd | No | Reported SD for bounded-scale feasibility check | |
| ci_level | No | Confidence level in (0, 1), default 0.95 | |
| ci_lower | No | Lower bound of confidence interval | |
| ci_upper | No | Upper bound of confidence interval | |
| n_params | No | Number of regression parameters (including intercept) when df_kind='regression' | |
| power_n1 | No | Group 1 sample size for achieved power | |
| power_n2 | No | Optional Group 2 sample size for achieved power | |
| sample_n | No | Optional sample size for bounded-scale SD check | |
| chi2_stat | No | Reported chi-square statistic (right-tail) | |
| ci_method | No | Optional interval construction method (e.g. wald, clopper-pearson, wilson, bootstrap) | |
| sample_df | No | Reported degrees of freedom for N-consistency check | |
| scale_max | No | Maximum possible value of the bounded scale | |
| scale_min | No | Minimum possible value of the bounded scale | |
| ci_rel_tol | No | Relative midpoint tolerance, default 0.02 | |
| reported_n | No | Reported total sample size N | |
| ci_estimate | No | Point estimate for CI consistency check | |
| ablation_tol | No | Tolerance for constant delta check, default 1e-9 | |
| ci_log_scale | No | Use geometric midpoint for ratio estimates (OR/RR/HR) | |
| sample_ratio | No | Group size ratio n2/n1 for required sample size, default 1.0 | |
| target_power | No | Target statistical power in (0, 1) for required sample size | |
| effect_size_d | No | Cohen's d for t-test power or sample-size calculation | |
| partition_tol | No | Allowed integer rounding tolerance for partition sum, default 0 | |
| ablation_matrix | No | 2D numeric matrix (rows=methods/tasks, cols=settings) for constant-offset regularity check | |
| p_value_literal | No | Optional exact reported decimal/scientific literal, preserving trailing zeros; must agree with p_value and any explicit decimals. | |
| partition_parts | No | Subset counts (e.g. [train, val, test]) | |
| partition_total | No | Declared total count for dataset split conservation check | |
| p_value_decimals | No | Optional decimal-place precision; must agree with a supplied literal. | |
| ablation_min_rows | No | Minimum row count for ablation delta check, default 3 | |
| bandwidth_rel_tol | No | Relative tolerance for storage size comparison, default 0.01 | |
| bandwidth_max_gbps | No | Hardware memory bandwidth ceiling in GB/s, default 4000.0 | |
| bandwidth_item_count | No | Total vector/item count for storage & bandwidth bound check | |
| bandwidth_vector_dim | No | Embedding dimension for storage & bandwidth bound check | |
| partition_share_unit | No | Unit for reported shares, default auto | |
| ablation_baseline_col | No | Baseline column index for ablation delta check, default 0 | |
| bandwidth_storage_unit | No | Unit for reported storage size, default auto | |
| partition_share_tol_pct | No | Allowed percentage-point tolerance for reported shares, default 0.5 | |
| partition_reported_shares | No | Optional reported subset shares or percentages | |
| bandwidth_reported_literal | No | Optional exact reported storage literal string preserving decimal precision | |
| bandwidth_reported_size_mb | No | Optional reported storage footprint in MB or MiB | |
| bandwidth_bytes_per_element | No | Bytes per element (default 4 for float32) | |
| bandwidth_reported_decimals | No | Optional explicit decimal places for reported storage size | |
| bandwidth_claims_full_system | No | Whether reported size claims to include full system text/metadata, default true | |
| bandwidth_reported_latency_us | No | Optional reported full-scan latency in microseconds | |
| bandwidth_min_metadata_overhead_bytes | No | Optional minimum provable non-vector metadata overhead in bytes |