Skip to main content

Table 1 Summary of statistical methods based on the types of variables

From: Active clearance of chest tubes is associated with reduced postoperative complications and costs after cardiac surgery: a propensity matched analysis

Type of variable XSummary statisticsStandardized differenceStatistical test for unmatched dataStatistical test for matched data
Symmetric continuous variable\( \overline{X}\pm s \)\( \frac{{\overline{X}}_2-{\overline{X}}_0}{\sqrt{\left({s}_2^2+{s}_0^2\right)/2}} \)Independent samples t-testPaired samples t-test
Non-symmetric continuous variablem [q1 − q3]The standardized difference defined above for lnXWilcoxon’s rank sum testWilcoxon’s signed rank test
Count variablecStandardized difference for symmetric continuous variableIndependent samples t-testPaired samples t-test
Binary variable\( f\ \left(\hat{p}\%\right) \)\( \frac{\hat{p_2}-\hat{p_0}}{\sqrt{\left(\hat{p_2}\left(1-\hat{p_2}\right)+\hat{p_0}\left(1-\hat{p_0}\right)\right)/2}} \)Fisher’s exact testMcNemar’s test