A longitudinal study in which variables are measured on the same units over time. For example, British Election Panel Studies have interviewed respondents after one election and again after the following election, and sometimes at various points in between. Panel surveys are particularly useful for understanding change at the individual level. Although cross-sectional studies can be used to estimate change by asking questions about past behaviour, such as vote choice at the last election, the answers can be unreliable. Furthermore, panel studies help avoid relying on aggregate data across time (see ecological association). However, they usually suffer from panel attrition—increasing levels of non-response with each successive ‘wave’ of the panel. Another potential problem is panel conditioning in which responses and/or behaviour are affected by membership of the panel.
— Stephen Fisher
