In a retrospective survey, the researchers may also like to investigate possible cause and effect relationships by observing some existing consequences and looking through the data from the past for plausible causal factors. This is different from experimental approach where data is collected in a controlled situation. For example, you wish to identify the factors characterizing persons having either high or low accident rates. You need to collect data of all the events of interest (from the past as well as present) occurring in certain subjects having high and ' low accident rates. You would then examine the data by going back, finding out relationships and their meaning.
This is also called as causal comparative approach, which is "Ex-post facto", innature i.e. data are collected by looking to past events. In social behaviour studies, it is difficult to find a cause for an effect because there are many variables or factors that influence behaviour or occurrence of events.