Hi everyone! I'm doing my master thesis on the effect of childhood poverty* (IV) on switching behaviors in adulthood (DV), and I am expecting an increase in the DV only in currently financially stressful situations (M) for people who grew up poor. My issue is, my moderator is categorical with only two categories (low difficulty/high difficulty) and when I run model 1 with financial difficulty as W I get strange results that the effects of the predictor (childhood poverty) are not significant at any of the two values (i.e. conditions) of the moderator even though their interaction is (p < 0.10)? See Effects1.png
I tried running it the other way around, with childhood poverty as the moderator and the results made more sense and confirmed my hypothesis. However, from what I can gather this output only gives me the results of the conditional effects of highly financially stressful situations (Cond =1) at different values of childhood poverty. (See Effects2.png) I understand that low financial stress situations would be 0, hence the effect would just be the main effect of childhood poverty, but as the simple slopes graph indicates there's a huge difference in the low condition between the participants who experienced childhood poverty and those that didn't.
Is there a way I can get the effect sizes of low financially stressful situations at different values of childhood poverty, or should I just use the main effect size of childhood poverty as justification for the blue line in the simple slopes? Many thanks!
*We defined childhood poverty as "childhood financial scarcity" hence the name of the variable ChScarc and the reverse scoring of it (positive childhood scarcity indicates growing up poor)
The crossed lines on the graph show moderation and thus interaction between the two variables. Full details can be found here:
https://www.google.com/search?client=firefox-b-1-d&ei=enbCX4nzA_-PwbkP65Wd0Ak&q=Mediation+moderation+Virginia+research+r&oq=Mediation+moderation+Virginia+research+r&gs_lcp=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&sclient=psy-ab&ved=0ahUKEwiJmMCi0KXtAhX_RzABHetKB5oQ4dUDCAw&uact=5
Best, D. Booth
Could you please show the whole PROCESS output? This would it make much easier for us.
I agree with David, the figure demonstrated the moderation effect.
For effect size, you may get some idea in the following paper.
Article Computation of Effect Size for Moderating Effects of Categor...
Rainer Duesing Sure! Model1.png is the first model with the categorical variable "condition of financial difficulty" as a moderator, and Model2.png is the second one with the continuous variable "childhood scarcity" as a moderator
In the first model it refers that the moderating effect is not significant, in the second model also refers the same the lack of significance, in this last model three levels are reported, only in the third level is significant that is to say that a higher level of the moderating variable has a significant effect as opposed to a low or medium level.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS