Results Structurer

Section Structure (Target: 1000–1200 words)

Sub-section 1 — Descriptive Statistics (~150 words)

• Narrate the means, SDs, and correlations table
• Flag multicollinearity if VIF > 10 or correlation > 0.85 between predictors
• Report sample characteristics (firm size tier, sector, years of operation) in 2–3 sentences
• APA-7 format: "The mean score for environmental practices was M = 3.72 (SD = 0.81)"

Sub-section 2 — Measurement Model / Reliability (for SEM studies, ~200 words)

• Report Cronbach's Alpha (> 0.70) and Composite Reliability (> 0.70) per construct
• Report Average Variance Extracted (AVE > 0.50) for convergent validity
• Report HTMT ratio (< 0.85) or Fornell-Larcker criterion for discriminant validity
• Confirm whether measurement model fits: CFI > 0.95, RMSEA < 0.06, SRMR < 0.08

Sub-section 3 — Hypothesis Testing (~600 words)

• One paragraph per hypothesis (or hypothesis group if closely related)
• Each paragraph: state the hypothesis → report the statistic → state accepted/rejected
• APA-7 in-text format: "β = 0.34, SE = 0.07, t = 4.86, p < 0.001, 95% CI [0.21, 0.47]"
• For SEM: report path coefficient, t-value (bootstrapped), and p-value
• For regression: report β, t, p, and ΔR² where relevant
• For mediation: report indirect effect with bootstrapped CI (Hayes PROCESS or PLS bootstrapping)

Sub-section 4 — Hypothesis Summary Table (~50 words + table)

Sub-section 5 — Additional Diagnostics (~100 words, if applicable)

• CMB: report Harman's single-factor variance (< 50% threshold)
• Endogeneity: report IV results or robustness check
• Moderation: report interaction term significance and plot direction

SEM Fit Index Thresholds (baked into agent knowledge)

Flags Raised by Agent

• ⚠️ Missing: effect size (Cohen's f² or R²) not reported
• ⚠️ Missing: confidence intervals for key coefficients
• ⚠️ Non-significant result not acknowledged in text
• ⚠️ SEM fit indices below threshold — must address in limitations
• ⚠️ Results section contains interpretation — move to Discussion