I have one group of people who completed measures at 3 time points and I'd like to calculate the effect size of change over time.
For that purpose, use one of the non-parametric aunts in the SPSS in the case of a single sample.
See the notes at this link. Marie Davidian is a real expert.
https://www.stat.ncsu.edu/people/davidian/courses/st790/notes.html
Best, David Booth
Hello Josephine,
A very simple way to proceed would be to compute mean (M) and standard deviation (SD) for each of the three time periods. Then, using time 1 as the baseline, define a Cohen's d ES estimate as (M_t1 - M_tx) / SD_t1 where "tx" means either time 2 or time 3, as desired. That statistic will give you the difference, in baseline SDs (and that is your ersatz common scale), from time 1 to time 2 (or time 3).
There are plenty of other schemes as well. Just enter "effect size" in YouTube or on the Amazon web sites to get a variety of explanations.
Good luck with your work.
You can simply apply three one-sample tests, one for each time point. The baseline mean will the total mean from all time-points. For each time-point, you deduct this total mean from each time-point's mean to get the mean difference and you divide this over the pooled SD. In this case, you would derive three effect size values. For the predictor of time, you can code them into two dummy variables and regress the effect size values against the two temporal regressors. With time-point 1 (or another) set as the reference time-point, you will get two beta coefficients, for instance, beta1 showing the change in effect size from time-point 1 to 2, and beta2 showing the change in effect size from time-point 1 to 3.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS