supportive message control condition or the 10 cases that compared a supportive
treatment with a no-treatment control.
In instances where tests of the moderators were not significant, an ‘‘after-the-fact’’
power analysis (O’Keefe, 2007, p. 294) was conducted following the procedures
specified by Hedges and Pigott (2004). Unlike a post hoc power test, where the
observed moderator value is used to assess power, after-the-fact power analysis is
conducted using a population value of interest for a moderator. A moderator value is
pre-selected and a power analysis is conducted to determine the amount of power
available to detect that value. The value of b.3 was used to assess power of the tests
evaluating the relationships between threat and involvement and the resistance
conferred by inoculation. The values of d.25, d.125, and d0 were selected to
assess the power associated with the test of delay as a moderator.