Abstract

Experiments are conducted to help establish cause-and-effect relationships, and they can be powerful tools for doing so. We review fundamental concepts for conducting experimental advertising research. Good experimental research involves careful consideration of independent and dependent variables, and what they are supposed to represent. To this end, we review threats to construct validity as well as offer some suggestions on how to think about external and ecological validity in ad research. We review three quasi-experimental research designs and three simple, randomized experimental designs, along with more complex factorial design experiments. Finally, we discuss ethical considerations and the crucial role researchers play in maintaining research integrity.

Experiments are designed and conducted to test whether, and to what extent, one thing causes another. Advertising researchers have used experiments to examine whether point-of-purchase advertising influences sales (e.g., Caballero and Solomon 1984; Greco and Swayne 1992), whether mood induced by a TV program affects processing of subsequent advertising (Cline and Kellaris 2007; LaTour and LaTour 2009; Shapiro and MacInnis 2002), whether repetition is an effective way to improve memory for brands (Nordhielm 2002), and whether increasing the salience of consumers’ ethnic identities affects responses to ads with patriotic appeals (Yoo and Lee 2016). Causal relationships are implicit in each of these relationships, and the best tool researchers have for determining causal relationships is experimental research.