Preferring familiar emotions: As you want (and like) it?

Abstract

Do people want to feel emotions that are familiar to them? In two studies, participants rated how much they typically felt various emotions (i.e., familiarity of the emotion) and how much they generally wanted to experience these emotions. We found that, in general, people wanted to feel pleasant emotions more than unpleasant emotions. However, for both pleasant and unpleasant emotions, people more (vs. less) familiar with an emotion also wanted to experience it more. Links between the familiarity of an emotion and wanting to experience that emotion were not explained by the concurrent experience of familiar emotions. Also, we show that although familiar emotions were also liked more, liking did not fully account for wanting familiar emotions. Finally, the familiarity of emotions mediated the links between trait affect and the emotions people wanted to feel. We propose that people are motivated to feel familiar emotions, in part, because of their instrumental value.

Keywords:

• Emotion
• Emotion regulation
• Familiarity
• Wanting
• Liking

Notes

¹ We also tested whether there is a curvilinear relationship between familiarity and wanting. For each emotion, we used hierarchical regressions to test whether the quadratic term of familiarity accounted for any additional variance beyond the linear term of familiarity by examining the r-squared change. Results indicated that the quadratic familiarity of cheerfulness did not account for any additional variance in predicting wanting cheerfulness, ΔR ²=.002, F(1, 54) < 1, p=.72, nor did the quadratic familiarity of anger account for any additional variance in predicting wanting anger, ΔR ²=.009, F(1, 54) < 1, p=.44. This shows that in this dataset there were no curvilinear relationships between the familiarity of an emotion and wanting to experience it.

² Although happiness is sometimes examined as a broader construct incorporating well-being and psychological health (Diener, Suh, Lucas, & Smith, 1999), consistent with prior work, we use this term to refer to the discrete positive state (e.g., Russell, 2003).

³ As we did in the pilot study, we tested whether there was a curvilinear relationship between familiarity and wanting of emotions. The quadratic familiarity terms did not account for additional variance in predicting wanting in the case of happiness, ΔR ²=.007, F(1, 136) = 1.08, p=.30, anger, ΔR ²=.00, F(1, 136) < 1, p=.91, or fear, ΔR ²=.00, F(1, 135) < 1, p=.91.

⁴ To control for error and the intercorrelation of variables, we replicated these analyses using structural equation modelling. The first model examined happiness and we entered familiarity, concurrent experience, and liking as intercorrelated predictors of wanting. Familiarity remained a significant predictor (β = 0.15, p=.054), concurrent experience was a significant predictor (β = 0.17, p=.038), and liking was a significant predictor (β = 0.38, p<.001). The second model examined anger and we entered familiarity, concurrent experience, and liking as intercorrelated predictors of wanting. Familiarity remained a significant predictor (β = 0.17, p=.035), concurrent experience was a significant predictor (β = 0.17, p=.033), and liking was a significant predictor (β = 0.25, p=.003). The third model examined fear and we entered familiarity, concurrent experience, and liking as intercorrelated predictors of wanting. Familiarity remained a significant predictor (β = 0.22, p=.004), concurrent experience was a significant predictor (β = 0.20, p=.007), and liking was a significant predictor (β = 0.37, p<.001).

⁵ We extended these analyses using structural equation modelling. In the first model, we entered extraversion, concurrent experience of happiness, and liking happiness as intercorrelated predictors of wanting happiness. Extraversion was not a significant predictor (β = 0.06, p=.42), the concurrent experience of happiness was a significant predictor (β = 0.22, p=.004), and liking happiness was a significant predictor (β = 0.39, p<.001). In the second model, we entered trait anger, the concurrent experience of anger, and liking anger as intercorrelated predictors of wanting anger. Trait anger was not a significant predictor (β = 0.12, p=.16), the concurrent experience of anger was a significant predictor (β = 0.20, p=.015), and liking anger was a significant predictor (β = 0.25, p=.004). In the third model, we entered neuroticism, the concurrent experience of fear, and liking fear as intercorrelated predictors of wanting fear. Neuroticism was a significant predictor (β = 0.20, p=.008), the concurrent experience of fear was a significant predictor (β = 0.22, p=.004), and liking fear was a significant predictor (β = 0.38, p<.001). Given that neuroticism was a significant predictor of wanting fear, we tested a final model in which both neuroticism and the familiarity of fear were entered as intercorrelated predictors of wanting fear to test whether familiarity remained a significant predictor when neuroticism was in the model. In this case, neuroticism was no longer a significant predictor (β = 0.16, p=.08) whereas familiarity of fear remained significant (β = 0.24, p=.01).

Additional information

Funding

Funding: This work was supported by a National Science Foundation grant (SES 0920918) to Maya Tamir.