Data was collected in three different organizations. Two issues. First, the overall sample is large (N=760). However, the concern is running three separate analyses for each sample would significantly reduce the power to find effects (as there are many variables in the model). Second, permission was granted by the three organizations to collect the data, but as it was online, it is very difficult to parse out which subject was in which organization (one university, two governmental organizations). Assuming this can be parsed out properly, can one use intraclass correlational analyses using organization as a variable to account for variance explained by location? Or is there a better method to use. Your suggestions would be greatly appreciated.