Page "Virginity pledge" Paragraph 14

A fifth peer-reviewed study, also by Janet Rosenbaum published in the journal Pediatrics in 2009, found no difference in sexual behavior of pledgers and similar non-pledgers five years after pledging, but found pledgers were 10 percentage points less likely to use condoms and 6 percentage points less likely to use birth control than similar non-pledgers.

Rosenbaum's study was innovative for using Rubin causal model matching, instead of relying on regression analysis, which makes potentially untrue parametric assumptions.

According to Rosenbaum, past research findings that virginity pledgers delayed sex may have been affected by their statistical method's inability to adjust fully for pre-existing differences between pledgers and non-pledgers: pledgers are much more negative towards premarital sex prior to even taking the pledge, so would be predicted to delay sex even if they hadn't taken the pledge.

Comparing pledgers with similar non-pledgers is the only way to be certain that the effect comes from the pledge rather than the pre-existing greater beliefs of pledgers that sexuality should be restrained to the matrimonial context.