In these data the magnitude of anticorrelation between withdrawal/NA and PA was only moderate, suggesting that these variables may not be at the opposite ends of a single affect dimension. In other words, these pilot data were not consistent with the circumplex model of emotion often utilized in EMA smoking studies which posits that NA and PA represent opposing ends of the same affective dimension (Russell, 1980; Russell & Carroll, 1999; Shiffman, 2013) and provide further evidence that these two variables should be examined as separate constructs (Watson & Tellegen, 1985).
