I think this is likely, but easily validated -- you'd probably see a couple of clusters or some latent variables, but the raw data would still likely be useful for getting average response characteristics.