Response to novelty as a predictor for drug effects: the pitfalls of some correlational studies

In recent years, an individual's response to novelty has been postulated to predict its response to drugs of abuse and particularly to their addictive properties (Piazza et al. 1990). The hypothesis of a relationship between the response to novelty and the effects of addictive drugs was supported by a number of animal studies that reported correlations between responses to a novel environment and various effects of drugs, such as their locomotor stimulant effects, their reinforcing action or their propensity to be self-administered (Piazza et al. 1990; Klebaur et al. 2001; Carey et al. 2003; Shimosato and Watanabe 2003). Most of these studies concluded that an animal's response to novelty predicts its subsequent response to drug administration. However, correlational studies are sometimes hampered by methodological and statistical weaknesses that preclude a proper interpretation of the results. The two most frequent weaknesses are the lack of consideration for the correlation in the control group and the calculation of spurious correlations.