ABSTRACT
The panel design, first proposed by Lazarsfeld and Fiske (1938), circumvents the difficult problem of assessing causal direction in cross– sectional research. The cross–lagged panel design studies and compares the effects of variables on each other across time. The causal direction in panel research is not based on instantaneous relationships between simultaneously measured variables x and y. Instead, different variables are used for opposite directions: x at Time Point 1 affecting y at Time Point 2, y at Time Point 1 affecting x at Time Point 2. It is therefore supposed to be more suitable than cross–sectional research in answering, for example, whether the advertisement of a particular commodity brand causes people to buy or, conversely, its frequent consumption causes people to note its advertisement, or both effects operate reciprocally ( Zeisel, 1968). Because of its strength, the design was classified as “quasi–experimental” in the well–known evaluation of designs by Campbell and Stanley (1966).
