Hi statalisters ... I need some help to carry out a power test over data I obtained from one experiment. I'll explain shortly.
I performed an experiment consisting of 3 treatments that I called C, G and CG. 30 different subjects participate in each treatment. These subjects are divided into groups of 3 so that there are 10 independent groups for each treatment. In each group the 3 subjects play for 7 periods and decide how much to invest in a project.
The dataset is therefore a panel with 90 subjects observed in 7 periods. Essentially I have two variable on which I want to run the power test: the first is an individual-level variable (each individual amount contributed to the project) and the second is a group-level variable (Gini index to measure inequality in incomes of the three subjects in each group). So I have 630 obs (90 times 7 periods) for the individual-level variable and 240 obs (30 times 7 periods) for the group-level variable.
I carefully consulted stata power and sample size reference manual (looking at the power command) but I did not found any reference to the case of panel data as in my case. I would be interested in performing a retrospective analysis of power with regarding pairwise comparisons of my treatments. Any useful suggestion would be greatly appreciated. Thank you in advance.
I performed an experiment consisting of 3 treatments that I called C, G and CG. 30 different subjects participate in each treatment. These subjects are divided into groups of 3 so that there are 10 independent groups for each treatment. In each group the 3 subjects play for 7 periods and decide how much to invest in a project.
The dataset is therefore a panel with 90 subjects observed in 7 periods. Essentially I have two variable on which I want to run the power test: the first is an individual-level variable (each individual amount contributed to the project) and the second is a group-level variable (Gini index to measure inequality in incomes of the three subjects in each group). So I have 630 obs (90 times 7 periods) for the individual-level variable and 240 obs (30 times 7 periods) for the group-level variable.
I carefully consulted stata power and sample size reference manual (looking at the power command) but I did not found any reference to the case of panel data as in my case. I would be interested in performing a retrospective analysis of power with regarding pairwise comparisons of my treatments. Any useful suggestion would be greatly appreciated. Thank you in advance.
Comment