Skip to main content
. Author manuscript; available in PMC: 2013 Dec 1.
Published in final edited form as: Psychol Methods. 2012 Aug 20;17(4):642–664. doi: 10.1037/a0029311

Table 5.

Results of Empirical Data Analysis.

Interaction Parameters Unadjusted Reg. Adjust IPW
None β1(SE) 0.067 (0.052) 0.067 (0.052) 0.067 (0.052)
β2(SE) −0.225 (0.029) −0.177 (0.028) −0.194 (0.034)
β3(SE) −0.048 (0.045) −0.037 (0.041) −0.038 (0.049)

T & Z β1(SE) 0.043 (0.078) 0.043 (0.078) 0.043 (0.078)
(Treat & Gender) β2(SE) −0.225 (0.029) −0.177 (0.028) −0.194 (0.033)
β3(SE) −0.059 (0.067) −0.046 (0.061) −0.039 (0.071)
β6(SE) 0.042 (0.104) 0.042 (0.104) 0.042 (0.104)

M & Z β1(SE) 0.067 (0.052) 0.067 (0.052) 0.067 (0.052)
(Job-Seek & Gender) β2(SE) −0.211 (0.045) −0.171 (0.042) −0.190 (0.046)
β3(SE) −0.043 (0.045) −0.037 (0.041) −0.031 (0.049)
β3(SE) −0.024 (0.058) −0.010 (0.054) −0.007 (0.066)

T & M β1(SE) 0.067 (0.052) 0.067 (0.052) 0.067 (0.052)
(Treat & Job_Seek) β2(SE) −0.270 (0.053) −0.240 (0.050) −0.214 (0.064)
β3(SE) −0.047 (0.045) −0.035 (0.041) −0.037 (0.049)
β4(SE) 0.065 (0.063) 0.088 (0.058) 0.029 (0.076)
Note: As a reminder, the marginal structural models are given below.
  • E[M(t)] = β0M + β1t
  • E[Y(t, m)] = β0Y + β2m + β3t
  • E[M(t)] = β0M + β1t + β5z + β6zt
  • E[Y(t, m)] = β0Y + β2m + β3t + β9z + β10zm
  • E[Y(t, m)] = β0Y + β2m + β3t + β4tm