When Do Warmth and Competence Sell Best? The “Golden Quadrant” Shifts as a Function of Congruity With the Product Type, Targets’ Individual Differences, and Advertising Appeal Type

Abstract

Three experiments reported here uniquely test the effectiveness of warm versus competent advertising strategy as a function of congruence with other elements of the advertising context. These are product involvement (Experiment 1, n = 96), consumers’ smartphone anxiety (Experiment 2, n = 60), or self-versus other-profitability (Experiment 3, n = 100). As expected, the “golden quadrant” (optimum warmth and competence for advertising effectiveness) does shift: Competence is more important for high-involving products, but warmth wins for highly anxious participants or when the highly involving service is accompanied by people-focused appeals. An expansion of the stereotype content model is discussed in the context of the congruity principle.

ACKNOWLEDGMENTS

We thank Harriet Small and Jennifer Shardlow for their help in data collection and initial analyses. Additional thanks go to Dr. Charlie Frowd and Dr. Sarah Bayless, The University of Winchester, UK, for their comments on the early drafts of this paper.

Notes

¹All four portrayals were rated as more than average in terms of attractiveness without being overly attractive (scores ranged between .72 and 1.71 on a −3 to 3 attractiveness Osgood differential scale). Almost all of the four male pictures were evaluated similarly in terms of attractiveness with two exceptions: Picture 3 scored higher (M = 1.39, SD = 1.54) than Picture 2 (M = .72, SD = 1.07), and this effect was of medium strength (d = 0.51); Picture 2 also scored lower than nontraditional Picture 5 (M = 1.71, SD = 1.21), and this effect was strong (d = 0.87). Therefore, an additional 4 (ad version) × 2 (product involvement) analysis of variance (ANOVA) was run, but this showed that the two versions of the two ad types did not influence ad effectiveness differently. The same held for Experiment 2 where these stimuli were used. Therefore, the scores were collapsed across the two versions of each ad type, and further analyses were run using ad type on two levels only: warm or competent.

²Two additional analyses were also conducted. First, the sample source was entered as a third variable (students vs. other). Because any effects of this variable on ad effectiveness were likely due to chance, sample source was excluded from further analyses. Second, the data from the whole sample were also analyzed using moderated regression where the anxiety level was treated as a continuous variable—a procedure that avoids median split and the associated reduction of sample size and loss of power. This analysis also detected an Ad Type × Anxiety interaction effect (β = .45). This effect was of the same shape as one returned by an ANOVA and the analysis reported above. As this analysis is more powerful than ANOVA, it also detected a main effect of anxiety (β = −.38), which indicated that the lower smartphone anxiety, the greater the reported effectiveness of both ad types. This effect is unsurprising and speaks to the validity of the adapted Smartphone Anxiety Scale. Taken together, these results qualify findings from Experiment 1: They indicate that indeed competence may not always be more important than warmth for high-involving products. Similarly to Experiment 1, competent ads were slightly more effective than warm ones, especially for the lowly smartphone-anxious people. However, for highly smartphone-anxious individuals, warmth was more effective than competence. Thus, again the “golden quadrant” shifted as a function of the relevance of warmth and competence to specific consumers.