Hi Everyone,
I am conducting a meta-analysis using Schmidt and Hunter's (2015) approach to partially fulfill my graduate degree requirements. Although I am pretty comfortable with the analysis, I hit a snag in the moderator analysis recently. I have already consulted various resources relevant to Schmidt and Hunter's (2015) approach but could not find a satisfactory answer. I have the following questions if anyone wants to take a stab at it:
Q.1. In order to conduct a moderator analysis, I simply created two groups of studies based on the level of the moderators and compared their statistics against one another to determine their moderating effect on the relation between the primary variables. And, I did that for each moderator separately (I have five moderators; 4 categorical and 1 continuous). However, according to Schmidt and Hunter (2015), whenever there are more than 1 moderator, the authors recommend employing hierarchical sub-group analysis instead of separate sub-group analysis in order to avoid any complications regarding multicollinearity among the given moderators. If some of the moderators are correlated and I perform separate sub-group comparisons for each moderator, it may lead to misleading results. Currently I am using Schmidt and Le's (2014) software to conduct this analysis, but it doesn't offer any option for analysis where I can simply throw in all the moderators in once and analyze them simultaneously (i.e., hierarchical sub-group moderator analysis). So now I am kind of stuck here and have no clue as to what should I do?
Q.2. My second question is regarding the properties of the moderators. I have 5 moderators in which 4 of them are categorical while 1 of them is continuous. Earlier, I transformed the continuous variable into a categorical variable and performed the moderator analysis to obtain results. However, after reading a couple of articles along with Schmidt and Hunter's (2015) text, I realized that it's not ideal to dichotomize a continuous variable especially when conducting a meta-analysis. So I was wondering if there is a way to analyze a continuous moderator using a hierarchical sub-group analysis?
If someone has any insight or suggestion, please feel free to throw in here. Any kind of comment and suggestion will be highly appreciated.
Thank you in advance.
- Badges
- Science topic
- Similar topics
- Social Psychology
- Attitude