It depends on the available technique you have. If you have XRD you can try to find a splitting of non-equivalent reflections.  The attached file shows a simulation where a perovskite has been transformed into rhombohedral and then the angle has been changed from 90 to 88 degrees. Automatically several peaks split into several which where before single peak, e.g. the 211. But you have to select a good one since 100 does not split at all, and also the amount of splitting is different. Usually the splitting is much much smaller so that this is a task many people try to solve but often it is more a question of believe than of convincing results. Usually you only see a peak broadening and this can also have some other reasons.  

Thank for your kind information.