Page 1 of 1

Why we constrain the stderr to be 1 in level 1?

Posted: Fri Sep 02, 2011 3:07 am
by haizhilan
Hi, every one, I need your help, please give me your suggestion.

Currently, I want to do multi-level meta-analysis to synthesize a psychological stress scale score of the same population. I modeled the procedures of JooP Hox's method ( ... apter8.htm). But I found his Y variable had been standarzied, but my Y is not, he constrained the level 1 stderr to be 1, but I can't. Who can tell me how to set the parameters?

In my data, I colletced 10 studies, all about the psychological stress level of the same category subjects (no control group), but were conducted at different time, area. The Y variable is a normal continuous variable. Data structure showed below.

Study NO. sample size Mean of score Standard deviation of score
1 60 1.88 0.56
2 46 1.44 0.47
3 89 2.01 0.65

My objective is to get a stable estimation of score (mean and SD) of this kind of people through multi-level meta-analysis.