help for confall

plot and display all possible effect estimates after estimation

confall [, eform(string) lockterms(varlist) format(%fmt) xis(string) xformat(%fmt) addaic addbic notable graph_options ]


confall plots and displays effect estimates from models with all possible combinations of potential confounders.The first independent variable is the exposure of interest and all other independent variables are potential confounders.


eform reports the estimated coefficients transformed to odds ratios, hazard ratios, relative risk i.e., exp(b) rather than b. Confidence intervals are similarly transformed.

format(%fmt) specifies the display format for presenting numbers in graph and table. format(%9.0g) is the default

xformat(%fmt) specifies the display format for presenting numbers in x axis. format(%9.0g) is the default

lockterms(varlist) specifies variables to be included in all models.

xis() specifies values for X-axis. The default is p representing p values. Alternatives include pr (rescaled X-axis using p^.23137821, but labeled as original p value) aic (Akaike Information Criterion), bic (Bayesian Information Criterion), r2 (R2 or Pseudo R2), and n (the number of confounders)}

addaic and addbic mark the effect estimate from the model with the minimum AIC and BIC, respectively.

table displays the table.

graph_options refers to options of graph twoway scatter.


. logistic diabetes BMI Age Sex IGT ACR CRP Cholesterol DiastolicBP GGT Smoking Drinking . confall, lockterm(Age Sex) eform(OR) . confall, eform(OR) xline(0.05) yline(1) . confall, eform(OR) yline(1) xis(aic)

Author Zhiqiang Wang <z.wang@uq.edu.au> The University of Queensland

Acknowledgments This work was supported by the NHMRC Australia (301024).

Also see