Tutorial #5:
Interpret the results

As illustrated above, the summary output includes either all or some the following columns depending on the criteria used:

• Factor: the name of the factor.
• Class: the class assigned during the stratification of evidence.
• n_studies: the number of independent studies included in the factor.
• total_n: the total number of participants included in the factor.
• n_cases: the number of cases included in the factor.
• n_controls: the number of controls included in the factor.
• measure: the measured used in the calculations.
• value: the value of the pooled effect size expressed in its original metric. Note that if a factor includes only one study, its effect size is used as the pooled effect size.
• value_CI: the 95% confidence interval (CI) around the pooled effect size expressed in its original metric. Note that if a factor includes only one study, its 95% CI is used as the pooled 95% CI.
• eG: the value of the pooled effect size expressed in equivalent Hedges' g (eG).
• eG_CI: the 95% CI around the pooled effect size expressed in eG.
• eOR: the value of the pooled effect size expressed in equivalent Odds ratio (eOR).
• eOR_CI: the 95% CI around the pooled effect size expressed in eOR.
• p_value: the p-value of the pooled effect size.
• I2: the inconsistency (\eqn{I^2) value (calculated only if the number of studies in the meta-analysis is equal or larger to 2).
• PI_eG: the 95% prediction interval (PI) expressed in eG (calculated only if the number of studies in the meta-analysis is equal or larger to 3).
• PI_eOR: the 95% PI expressed in eOR (calculated only if the number of studies in the meta-analysis is equal or larger to 3).
• PI_sign: whether the 95% PI includes the null value ("notnull" vs. "null").
• egger_p: the p-value of the Egger's test for publication bias (calculated only if the number of studies in the meta-analysis is equal or larger to 3).
• egger_sign: whether the p-value of the Egger's test is < .05 ("sig." vs. "ns").
• ESB_p: the p-value of the Ioannidis' test for excess of significance.
• ESB_sign: whether the p-value of the Ioannidis' test is < .05 ("sig." vs. "ns").
• power_med: the power to detect a SMD of 0.5 at an alpha of .05 based on the number of cases and controls included in the meta-analysis (when IRR is used as effect size measure, the number of cases and controls in this calculation is assumed to be equal to half the total number of cases included in the meta-analysis).
• power: present only in the 'Personalized' classification.
  • If the user did not use the 'power' criteria to stratify the evidence, this column contains the power to detect a small effect size (SMD = 0.2), a moderate effect size (SMD = 0.5) and a large effect size (SMD = 0.8) at an alpha of .05 based on the number of cases and controls included in the meta-analysis.
  • If the user used the 'power' criteria to stratify the evidence, this column contains the power to detect the values entered by the user at an alpha of .05 based on the number of cases and controls included in the meta-analysis.
• JK_p: the largest p-value obtained in the jackknife leave-one-out meta-analysis (calculated only if the number of studies in the meta-analysis is equal or larger to 2)
• JK_sign: whether the largest p-value in the jackknife meta-analysis is < .05 ("sig." vs. "ns")
• largest_CI_eG: the 95% CI of the largest study expressed in eG
• largest_CI_eOR: the 95% CI of the largest study expressed in eOR
• largest_sign: whether the 95% CI of the largest study includes the null value ("notnull" vs. "null")
• rob: the percentage of participants included in studies at low risk of bias (calculated only if this information is indicated in the dataset)
• amstar: the AMSTAR score of the meta-analysis (calculated only if this information is indicated in the dataset)