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.

People vary in the extent to which they experience particular emotions. Some people experience happiness more than others while some experience anger or fear more than others. This investigation tested whether and how the emotions people typically feel (i.e., familiar emotions) are linked to the emotions people want to feel. To the extent that the emotions people want to experience can shape emotion regulation and experience (e.g., Mauss, Tamir, Anderson, & Savino, 2011; Tamir & Ford, 2012; Tsai, Miao, Seppala, Fung, & Yeung, 2007), examining the interplay between the emotions people typically feel and the emotions they want to feel may contribute to our understanding of affective individual differences and the potential processes that underlie them.

Possible links between familiar emotions and desirable emotions

The degree to which people want to experience a particular emotion may be associated with how familiar that emotional experience is to them. At least five different predictions can be made about the patterns of associations between familiarity of emotions and how motivated people are to experience them. These are described below.

First, the emotions people want to feel may depend on aspects that are unrelated to the familiarity of those emotional experiences (e.g., they depend solely on whether the emotion is pleasant or unpleasant to experience). In this case, the familiarity of emotions and the degree to which people want to experience them should not be significantly associated.

Second, familiarity might make people more sensitive to the hedonic implications of an emotion. In particular, people who are more familiar with pleasant emotions may be more motivated to experience them, whereas people who are more familiar with unpleasant emotions may be even less motivated to experience them. In this case, the link between familiarity and wanting would depend on the valence of the emotions at hand. In the case of pleasant emotions, familiarity of an emotion should be positively associated with the extent to which people want to experience it. In the case of unpleasant emotions, familiarity of an emotion should be negatively associated with the extent to which people want to experience it.

Third, familiarity might lead to habituation. Indeed, evidence suggests people can habituate to many kinds of experiences (e.g., Bolger, DeLongis, Kessler, & Schilling, 1989). If people get used to familiar emotions, the link between familiarity and wanting should still depend on the valence of the emotions at hand. In the case of pleasant emotions, familiarity of an emotion should be negatively associated with the extent to which people want to experience it. In the case of unpleasant emotions, familiarity of an emotion should be positively associated with the extent to which people want to experience it.

Fourth, if some quality of familiar emotions makes them less desirable, the familiarity of emotions would be negatively associated with wanting to experience these emotions, regardless of whether emotions are pleasant or unpleasant.

Finally, if some quality of familiar emotions makes them more desirable, the familiarity of emotions would be positively associated with wanting to experience these emotions, regardless of whether emotions are pleasant or unpleasant.

The first goal of this investigation was to identify whether and what kind of links exist between the familiarity of emotions and how much people want to experience these emotions. We begin by reviewing the gaps in the available literature and suggesting possible ways to address them.

Review of existing literature

Several studies to date have examined links between familiar emotional experiences and what people want to feel. These studies typically assessed broad affective traits that are often linked to emotional experiences (e.g., extraversion and neuroticism) and preferences for affective states that vary on core affective dimensions (i.e., valence and arousal). People higher (vs. lower) in extraversion tend to be more familiar with positive affect, whereas people higher (vs. lower) in neuroticism tend to be more familiar with negative affect (Watson & Clark, 1992). Thus, assessing individual differences in affective traits like extraversion and/or neuroticism can also index familiarity of positive and negative emotions, respectively. Studies that assessed individual differences in affective traits and preferences for affective states have found that affective traits are indeed associated with preferences for affective states, ruling out the possibility that familiarity and preferences for emotions are unrelated to each other. The nature of these links, however, has not been consistent across studies.

Some studies have found that people higher (vs. lower) in extraversion tend to show stronger preferences for positive affect (Augustine, Hemenover, Larsen, & Shulman, 2010; Kampfe & Mitte, 2009; Rusting & Larsen, 1995; Tsai, Knutson, & Fung, 2006). Such findings show that the more familiar someone is with positive affect, the more they want to experience it. This pattern is inconsistent with the possibility that familiar emotions are less desirable. This pattern is also inconsistent with the possibility that people become habituated to familiar emotions. Instead, this pattern is consistent with two alternative accounts. First, it is possible that people become sensitised to the hedonic implications of familiar emotions (i.e., people want pleasant emotions more and unpleasant emotions less when those emotions are familiar). Second, it is possible that familiar emotions are more desirable overall.

To test these two alternative accounts, it is necessary to examine the links between unpleasant familiar emotions and preferences for such emotions. If people are sensitised to the hedonic implications of familiar emotions, people who are more familiar with unpleasant emotions (e.g., those higher in neuroticism) should be less motivated than others to experience unpleasant emotions. However, if familiar emotions become more desirable overall, people who are more familiar with unpleasant emotions (e.g., those higher in neuroticism) should be more motivated than others to experience unpleasant emotions.

Unfortunately, research on neuroticism and emotional preferences has been relatively inconsistent. Some studies have found that people higher (vs. lower) in neuroticism show stronger preferences for negative affect (Kampfe & Mitte, 2009; Tamir, 2005), other studies have found that people higher (vs. lower) in neuroticism show stronger preferences for positive affect and do not differ in preferences for negative affect (Augustine et al., 2010; Rusting & Larsen, 1995), and still other studies have found that people higher (vs. lower) in neuroticism show weaker preferences for positive affect (Tsai et al., 2006).

One possible reason for this inconsistency may be that affective traits (e.g., neuroticism) are broad constructs that include not only affective characteristics, but also motivational and behavioural ones. Thus, the associations between affective traits and emotional preferences may be influenced by various factors, other than the familiarity of emotions. To test whether familiarity of emotions is related to how much people want to experience them, it is necessary to assess individual differences in familiarity, per se.

In this investigation, therefore, we examined whether the familiarity of an emotion was linked to the extent to which people want to experience that emotion. By examining the pattern of associations between familiarity and wanting of both pleasant and unpleasant emotions, this investigation sought to test whether only pleasant familiar emotions or all familiar emotions are more desirable. A second aim of this investigation was to examine why this might be the case.

Exploring familiarity and wanting

If people who are more familiar with unpleasant emotions want to experience unpleasant emotions less, this would suggest that the more people experience an emotion, the more sensitised they become to its hedonic value. An alternative, however, is that people who are more familiar with unpleasant emotions would want to experience them more. Below we propose several mechanisms that may account for such a pattern.

People are often motivated to maintain their concurrent affective state (Mayer & Stevens, 1994) and familiar emotions are more likely to be experienced at any given time (Watson & Clark, 1984). Thus, it is possible that links between familiarity and wanting are, in fact, driven by concurrent emotions. If so, it may be that the associations between familiarity and wanting are spurious, and depend on what emotions people happen to be feeling when asked about what they want to feel. We tested this possibility in the present investigation by assessing familiarity of emotions, wanting emotions, as well as concurrent emotional experiences.

It is also possible that there are meaningful links between familiarity of emotions and how much people want to experience them. For instance, it is possible that the more familiar an emotion is, the more people like that emotion. To the extent that familiarity breeds liking (Zajonc, 1980), it is possible that people like pleasant emotions more, and dislike unpleasant emotions less, the more familiar those emotions are. Liking (i.e., finding something pleasant to experience) and wanting (i.e., being motivated to attain something) tend to be related to one another (Berridge, 1996). If people like what is familiar and want what they like, then the more people like familiar emotions, the more motivated they would be to experience them. Although they are often related, however, liking and wanting involve different mechanisms that correspond to neurally and conceptually separable pathways (e.g., Berridge & Robinson, 2003). Whereas liking is hedonic in nature and involves experiencing pleasure in response to particular stimuli, wanting is motivational in nature and involves goal-related drives to seek particular stimuli. It is possible, therefore, for people to want something but not necessarily because they like it. According to the instrumental approach to emotion regulation (e.g., Tamir, 2009), people may want to experience emotions for various reasons that are unrelated to their hedonic quality. For instance, people pursuing a confrontational goal were more motivated to experience anger because they believed anger would be useful to them (Tamir & Ford, 2012). To the extent that people want to experience familiar emotions more, this may be driven by hedonic reasons (i.e., liking) but may also be driven by non-hedonic reasons (e.g., instrumentality).

We tested these possibilities in the current investigation. To do so, we assessed wanting and liking of emotions separately. We examined whether familiarity of emotions is associated with liking them and, if so, whether such links account for the associations between familiarity of an emotion and wanting to experience it.

Current investigation

We conducted a Pilot Study and Study 1 to test our theoretical questions. In both studies, we assessed the extent to which people were familiar with the experience of various emotions and the extent to which people wanted to experience these emotions. In both studies, we assessed concurrent emotions, to examine whether they account for any links between familiarity and wanting. We also measured liking of emotions, separately from wanting, to test whether familiarity and wanting are related to liking, and if so, whether liking accounts for links between familiarity and wanting of emotions. To avoid carryover effects, in both studies, core constructs were assessed in separate sessions. Specifically, familiarity, wanting, and liking were assessed in three separate sessions in the Pilot Study, and familiarity was assessed in a separate session from wanting and liking in Study 1. Finally, in Study 1, we also measured affective traits to examine whether the emotions people want to feel are more strongly associated with the familiarity of emotions than with broad affective dispositions.

We predicted that people would want to experience emotions that are more familiar to them, regardless of whether those emotions are pleasant or unpleasant. We predicted that such associations would remain even when controlling for concurrent emotions. We also expected participants to like familiar emotions more. However, we did not expect liking to fully account for links between familiarity of emotions and wanting to experience them. Finally, we expected that familiarity of emotions would drive the associations between broad affective traits and the emotions people want to experience.

PILOT STUDY

We examined whether people who are more familiar with cheerfulness and anger are more likely to want to experience them. These emotions were selected to represent two widely studied discrete emotions that differ in hedonic quality (i.e., pleasant vs. unpleasant, respectively). To minimise the risk of multicollinearity and response biases (e.g., desire for consistency), we measured familiarity, wanting, and liking in three separate sessions, several days apart.

METHODS

Participants

Participants (N=60, 52% females, M_age=23.5 years) were recruited to complete this study as part of a larger research project. All participants were undergraduate students at a large Israeli university.

Materials and procedure

Participants completed the study in three separate sessions, each 3–5 days apart. During the first session, we assessed liking emotions by asking participants how pleasant it was for them to feel cheerful and angry. During the second session, we assessed concurrent emotions by asking to what extent they felt cheerful and angry right now, and assessed wanting emotions by asking participants to what extent they generally wanted to feel cheerful and angry. During the third session, we assessed familiarity of emotions by asking to what extent they generally felt cheerful and angry. All responses were rated on a scale of 1 (Not at all) to 5 (Extremely). Within each session, scales were administered in the order listed and were interspersed among other unrelated study materials (e.g., various reaction time and perceptual tasks). The average time between the first and the second session was 3.5 days and the average time between the second and the third session was 3.8 days.

Results

Mean level patterns

We first conducted a series of t-tests to examine, on average, what emotions people were familiar with, were currently experiencing, wanted to feel, and liked to feel, regardless of individual differences. As might be expected, on average, people were more familiar with cheerfulness than with anger, t(58) =9.81, p<.001, were currently experiencing more cheerfulness than anger, t (58) = 6.02,p <.001, wanted to feel more cheerfulness than anger, t(57) =23.90, p<.001, and liked cheerfulness more than anger, t (59) = 24.27, p <.001 (see Table 1 for descriptive statistics).

Table 1.

Descriptive statistics and simple correlations between study variables (Pilot Study)

	Familiarity		Concurrent		Wanting		Liking	Descriptives
	1	2	3	4	5	6	7	Mean (SD)
Familiarity
1. Cheerfulness	—							3.29 (0.98)
2. Anger	.04	—						1.71 (0.79)
Concurrent
3. Cheerfulness	.52*	.03	—					2.39 (1.16)
4. Anger	.15	.25†	−.20	—				1.24 (0.70)
Wanting
5. Cheerfulness	.37*	.16	.32*	.10	—			4.40 (0.88)
6. Anger	.03	.47*	−.08	−.13	−.02	—		1.24 (0.47)
Liking
7. Cheerfulness	.51*	.12	.35*	.12	.73*	−.08	—	4.50 (0.75)
8. Anger	−.15	.41*	−.19	.00	.01	.50*	— .17	1.28 (0.58)

Note:

^*p<.05;

^†p<.06.

Associations between familiarity and wanting

To examine whether individual differences in the familiarity of emotions are associated with wanting to experience these emotions, we ran a series of simple correlations. As shown in Table 1, we found that people more (vs. less) familiar with cheerfulness wanted to experience cheerfulness more. Furthermore, we found that people more (vs. less) familiar with anger wanted to experience anger more.¹

Familiarity, wanting, and concurrent emotions

First, in the case of cheerfulness, as expected and shown in Table 1, people more (vs. less) familiar with cheerfulness were more likely to experience cheerfulness when also rating how much they wanted it. The concurrent experience of cheerfulness, in turn, was also associated with wanting to experience cheerfulness. When familiarity of cheerfulness and concurrent cheerfulness were entered as simultaneous predictors of wanting cheerfulness in a regression analysis, familiarity of cheerfulness was a marginally significant predictor, β = 0.27, t(55) = 1.87, p = .067, and concurrent cheerfulness was not significant, β = 0.18, t(55) = 1.26,p = .21.

Second, in the case of anger, there was a tendency for people more (vs. less) familiar with anger to experience more anger when also rating how much they wanted it. The concurrent experience of anger, however, was not positively associated with wanting to experience anger. When familiarity of anger and concurrent anger were entered as simultaneous predictors of wanting anger in a regression analysis, the familiarity of anger remained a significant predictor, β = 0.53, t(55) = 4.47,p <.001, and concurrent anger was a significant negative predictor, β = −0.26, t(55) =2.16, p = .035.

Familiarity, wanting, and liking emotions

As expected and shown in Table 1, the more familiar people were with cheerfulness, the more they liked it. Similarly, the more familiar people were with anger, the more they liked it. To test whether liking accounted for the link between familiarity and wanting, we ran the simple regressions described above, adding liking as a simultaneous predictor.

First, in the case of cheerfulness, we found that the familiarity of cheerfulness was no longer a significant predictor, β = −0.05, t(54) < 1, p = .69, concurrent cheerfulness was not a significant predictor, β = 0.10, t(54) < 1, p = .36, but liking cheerfulness remained a significant predictor, β = 0.72, t(54) =6.72, p<.001. Because familiarity of cheerfulness was rendered non-significant when liking cheerfulness was included in the model, we tested whether liking cheerfulness mediated the link between the familiarity of cheerfulness and wanting cheerfulness. A bias-corrected bootstrap (Preacher & Hayes, 2008) with 5,000 samples confirmed that liking cheerfulness significantly mediated the link between familiarity of cheerfulness and wanting cheerfulness, CI95 [.13, .57].

Second, in the case of anger, when the familiarity of anger, concurrent anger, and liking anger were entered as simultaneous predictors in a regression analysis predicting wanting anger, familiarity of anger remained a significant predictor, β = 0.38, t(54) = 3.13, p = .003, liking anger was a significant predictor, β = 0.34, t(54) = 2.90, p = .005, and concurrent anger was a marginally significant predictor, β = −0.22, t(54) = 1.95,p = .057.

Discussion

These results lead to several conclusions. First, people do indeed want to experience more familiar emotions. Furthermore, this pattern applies to both pleasant (i.e., cheerfulness) and unpleasant (i.e., anger) emotions. These findings are inconsistent with the possibility that people become sensitised to the hedonic implications of familiar emotions or that they habituate to them. Instead, these findings suggest that familiar emotions are more desirable, whether pleasant or unpleasant.

Second, concurrent cheerfulness was not a significant predictor of wanting cheerfulness when familiarity of cheerfulness was included in the model, but neither was familiarity of cheerfulness, suggesting that concurrent cheerfulness may contribute to the association between familiarity and wanting cheerfulness. In the case of anger, however, the pattern was clear. Concurrent anger did not account for the association between the familiarity of anger and wanting anger because these associations remained significant when controlling for concurrent anger.

Third, people tended to like familiar emotions, and this was the case for both pleasant and unpleasant emotions. Interestingly, liking mediated the link between familiarity and wanting in the case of cheerfulness, but liking did not account for the link between familiarity and wanting in the case of anger. This suggests that whereas people more familiar with cheerfulness want it more because they like it more, people more familiar with anger want it more, but not necessarily because they like it more.

This pilot study offered preliminary support for our hypotheses but had several limitations. First, only two emotions were assessed (cheerfulness and anger). Second, a single item was used to measure each emotion. Finally, trait affect (e.g., extraver-sion) was not assessed. Study 1 was designed to replicate the findings of the pilot study and address each of these limitations.

STUDY 1

First, to establish generalisability, we assessed three different emotions in Study 1: happiness,² anger, and fear. Second, we used a larger sample and more reliable measures composed of multiple items to assess each emotion. Third, to test whether the emotions people want to feel are linked to their familiarity per se, or to trait affect more broadly, we also examined associations between familiarity of happiness, anger, and fear and related indices of trait affect (i.e., extraversion, trait anger, and neuroticism, respectively). Finally, similar to the Pilot Study, to minimise the risk of response biases, we measured familiarity in a separate session, approximately five months before the measurement of liking and wanting.

Methods

Participants

Participants (N = 141, 56% females, M_age =20.25 years) were recruited to complete this study as part of a larger research project. All participants were undergraduate students at an American college. Two participants were omitted due to extreme responses ( >2.5 SDs from mean), although the main results remained unchanged when these participants were included in the analyses.

Materials

Familiarity of emotions

Participants were asked to what extent they generally felt cheerful, happy, and joyful (happiness; α = .79), angry, irritated, and annoyed (anger; α =.85) and feaful, worried, and nervous (fear; α = .76) on a scale of 0 (Not at all) to 6 (Extremely).

Wanting emotions

Participants were asked to what extent they generally wanted to feel cheerful, happy, and joyful (happiness; α = .89), angry, irritated, and annoyed (anger; α = .91) and fearful, worried, and nervous (fear; α = .85) on a scale of 0 (Not at all) to 6 (Extremely).

Liking emotions

Participants were asked how pleasant it was when they felt cheerful, happy, and joyful (happiness; α = .85), angry, irritated, and annoyed (anger; α = .84) and feaful, worried, and nervous (fear; α = .85) on a scale of 0 (Not at all) to 6 (Extremely).

Concurrent emotions

Participants were asked to what extent they felt happiness, anger, and fear right now on a scale of 0 (Not at all) to 6 (Extremely).

Trait affect

Participants completed the Big Five extraversion subscale (Goldberg, 1999), the Trait Anger Scale (Spielberger, Jacobs, Russel, & Crane, 1983), and the Big Five neuroticism subscale (Goldberg, 1999). The extraversion sub-scale includes 10 items (e.g., I am the life of the party; α = .85) that participants rated on a scale of 1 (Very inaccurate) to 5 (Very accurate). The trait anger scale includes 10 items (e.g., I have a fiery temper; α = .85) that participants rated on a scale from 1 (Almost never) to 4 (Almost always). The neuroticism subscale includes 10 items (e.g., I get stressed out easily; α = .84) that participants rated on a scale of 1 (Very inaccurate) to 5 (Very accurate).

Procedure

Participants completed the study in two sessions. During the first session we assessed familiarity of emotions and trait affect. During the second session we assessed concurrent emotions, wanting emotions, and liking emotions, in this order, with multiple unrelated surveys interspersed between scales (average time between first and second sessions = 19.0 weeks; range = 13.6–27.6 weeks).

Results

Mean level patterns of emotions

We first conducted a series of repeated-measures analyses of variance (ANOVAs) to examine what emotions people were familiar with, were currently experiencing, wanted to feel, and liked to feel, on average, regardless of individual differences. When examining familiarity of emotions, there was a main effect for Emotion, F(2, 272) = 134.73, p<.001, such that, on average, people were more familiar with happiness than anger and fear, ts(137) > 12.51,ps <.001. When examining concurrent emotions, there was a main effect for Emotion, F(2, 280) = 203.23, p <.001, such that on average, people were currently feeling more happiness than anger and fear, ts(140) > 14.21, ps<.001, and were currently feeling more fear than anger, t(140) = 2.84, p = .005. When examining wanting, there was a main effect for Emotion, F(2, 278) = 410.42, p < .001, such that on average, people wanted to feel more happiness than anger and fear, ts(139) > 20.80, ps<.001, and wanted to feel more fear than anger, t(139) = 2.92, p = .004. Finally, when examining liking, there was a main effect for Emotion, F(2, 270) = 425.18,p <.001, such that on average, people liked happiness more than anger and fear, ts(135) > 20.98, ps<.001. See Table 2 for all descriptive statistics.

Table 2.

Descriptive statistics and simple correlations between study variables (Study 1)

	Familiarity			Concurrent			Wanting			Liking			Trait affect		Descriptives
	1	2	3	4	5	6	7	8	9	10	11	12	13	14	Mean (SD)
Familiarity
1. Happiness	—														3.53 (1.04)
2. Anger	−.17*	—													1.63 (1.17
3. Fear	−.14	.58*	—												1.75 (1.19)
Concurrent
4. Happiness	.39*	−.13	−.17*	—											3.41 (1.30)
5. Anger	−.07	.26*	.20*	−.31*	—										0.68 (1.04)
6. Fear	.01	.18*	.30*	−.24*	.41*	—									0.98 (1.28)
Wanting
7. Happiness	.29*	.05	.02	.31*	−.14	−.08	—								4.47 (1.34)
8. Anger	−.04	.28*	.17*	−.24*	.27*	.29*	−.43*	—							0.74 (1.18)
9. Fear	−.06	.23*	.32*	−.19*	.33*	.32*	−.39*	.78*	—						0.92 (1.05)
Liking
10. Happiness	.18*	−.28*	−.20*	.22*	−.28*	−.14	.45*	−.24*	−.30*	—					4.97 (1.09)
11. Anger	−.16	.26*	.16	−.11	.23*	.14	−.21*	.33*	.40*	−.54*	—				1.18 (1.23)
12. Fear	−.05	.15	.10	−.09	.19*	.14	−22*	.29*	.43*	−.56*	.81*	—			1.20 (1.28)
Trait affect
13. Extraversion	.39*	−.23*	−.33*	.14	−.09	−.17*	.17*	−.12	−.13	.20*	−.08	−.02	—		3.39 (0.59)
14. Trait anger	−.05	.40*	.18*	−.11	.20*	.13	−.05	.23*	.15	−.20*	.32*	.17	.08	—	1.88 (0.54)
15. Neuroticism	−.20*	.46*	.50*	−.20*	.18*	.26*	−.02	.27*	.28*	−.07	.14	.07	−.23*	.40*	2.97 (0.61)

Note:

^*p<.05.

Associations between familiarity and wanting

To examine whether familiarity of emotions was associated with wanting to experience these emotions, we ran a series of simple correlations. As shown in Table 2, we found that people more (vs. less) familiar with happiness wanted to experience happiness more, people more (vs. less) familiar with anger wanted to experience anger more, and people more (vs. less) familiar with fear wanted to experience fear more.³

Familiarity, wanting, and concurrent emotions

First, in the case of happiness, as expected and shown in Table 2, people more (vs. less) familiar with happiness were more likely to experience happiness when also rating how much they wanted it. The concurrent experience of happiness, in turn, was also associated with wanting to experience happiness. When familiarity of happiness and concurrent happiness were entered as simultaneous predictors of wanting happiness in a regression analysis, familiarity of happiness was a significant predictor, β =0.19, t(136) =2.21, p = .029, and concurrent happiness was also significant, β = 0.25, t(136) = 2.87, p = .005.

Second, in the case of anger, people more (vs. less) familiar with anger were more likely to experience anger when also rating how much they wanted it. The concurrent experience of anger, in turn, was also associated with wanting to experience anger. When familiarity of anger and concurrent anger were entered as simultaneous predictors of wanting anger in a regression analysis, familiarity of anger was a significant predictor, β = 0.23, t(136) = 2.71, p = .008, and concurrent anger was also significant, β = 0.22, t(136) =2.59, p = .011.

Third, in the case of fear, people more (vs. less) familiar with fear were more likely to experience fear when also rating how much they wanted it. The concurrent experience of fear, in turn, was also associated with wanting to experience fear. When familiarity of fear and concurrent fear were entered as simultaneous predictors of wanting fear in a regression analysis, familiarity of fear was a significant predictor, β = 0.25, t(135) = 2.95, p = .004, and concurrent fear was also significant, β = 0.24, t(135) = 2.93, p = .004.

Familiarity, wanting, and liking emotions

As expected and shown in Table 2, the more familiar people were with happiness, the more they liked it. Similarly, the more familiar people were with anger, the more they liked it. The correlation between familiarity of fear and liking of fear, however, was not significant. To test whether liking accounted for the link between familiarity and wanting, we ran the simple regressions described above, adding liking as a simultaneous predictor.

First, in the case of happiness, we found that the familiarity of happiness was a marginal predictor, β =0.16, t(131) = 1.90, p = .059, concurrent happiness was a significant predictor, β = 0.17, t(131) = 2.08, p = .04, and liking happiness was a significant predictor, β = 0.38, t(131) = 4.89, p < .001. Because familiarity of happiness was rendered non-significant when liking happiness was included in the model, we tested whether liking happiness mediated the link between the familiarity of happiness and wanting happiness. A bias-corrected bootstrap with 5,000 samples confirmed that liking happiness significantly mediated the link between familiarity of happiness and wanting happiness, CI₉₅ [.02, .19].

Second, in the case of anger, we found that the familiarity of anger was a significant predictor, β =0.17, t(131) = 2.06, p = .042, concurrent anger was a marginal predictor, β = 0.16, t(131) = 1.97, p = .052, and liking anger was a significant predictor, β =0.25, t(131) = 2.96, p = .004.

Third, in the case of fear, we found that the familiarity of fear was a significant predictor, β = 0.21, t(130) = 2.69, p = .008, concurrent fear was a significant predictor, β = 0.20, t(130) = 2.58, p = .011, and liking fear was a significant predictor, β =0.38, t(130) = 5.04, p <.001.⁴

Familiarity of emotions, affective traits, and wanting emotions

We predicted that the familiarity of an emotion would mediate the link between trait affect and wanting. Indeed, in the case of happiness, the familiarity of happiness was positively correlated with extraversion, and extraversion was positively correlated with wanting happiness (see Table 2). When both the familiarity of happiness and extraversion were entered in a regression as simultaneous predictors of wanting happiness, familiarity of happiness remained a significant predictor, β = 0.26, t(136) = 2.95, p = .004, while extraversion did not, β = 0.07, t(136) < 1, p = .46. A bias-corrected bootstrap with 5,000 samples confirmed that familiarity of happiness significantly mediated the link between extraversion and wanting happiness, CI₉₅ [.08, .45].

In the case of anger, the familiarity of anger was positively correlated with trait anger, and trait anger was positively correlated with wanting anger (see Table 2). When both the familiarity of anger and trait anger were entered in a regression as simultaneous predictors of wanting anger, familiarity of anger remained a significant predictor, β = 0.22, t(136) = 2.52, p = .013, while trait anger did not, β =0.14, t(136) = 1.57, p=.12. A bias-corrected bootstrap with 5,000 samples confirmed that familiarity of anger significantly mediated the link between trait anger and wanting anger, CI₉₅ [.05, .42].

Finally, in the case of fear, the familiarity of fear was positively correlated with neuroticism, and neuroticism was positively correlated with wanting fear (see Table 2). When both the familiarity of fear and neuroticism were entered in a regression as simultaneous predictors of wanting fear, familiarity of fear remained a significant predictor, β = 0.24, t(135) = 2.56, p = .011, while neuroticism did not, β = 0.16, t(135) = 1.72, p = .09. A bias-corrected bootstrap with 5,000 samples confirmed that familiarity of fear significantly mediated the link between neuroticism and wanting fear, CI₉₅ [.03, .43].⁵

Discussion

The findings of Study 1 demonstrate that the more familiar people are with an emotion— whether pleasant or unpleasant—the more they want to experience it. This pattern was obtained for happiness, anger, and fear, suggesting that familiar emotions, both pleasant and unpleasant, are more desirable. Furthermore, in all cases, links between familiarity of an emotion and wanting to experience it were not explained by the concurrent experience of that emotion.

Our findings show that people want to experience familiar emotions and that—at least in the case of happiness and anger—people like to experience familiar emotions. Liking, however, did not fully account for the link between familiarity and wanting. Familiarity of anger and fear still predicted wanting anger and fear, respectively, when liking and concurrent emotions were also included as predictors. Familiarity of happiness, however, no longer significantly predicted wanting happiness when liking and concurrent happiness were included as predictors. Indeed, liking happiness mediated the link between familiarity and wanting. This raises the possibility that different factors may contribute to the links between familiarity and wanting pleasant versus unpleasant emotions.

Finally, we found that familiarity mediated the links between trait affect and wanting emotions. This supports our claim that links between trait affect and the emotions people want to experience, as reported in prior studies, are driven specifically by the familiarity of emotions and demonstrates the utility of measuring familiarity per se.

GENERAL DISCUSSION

This investigation sought to identify whether and what kind of links exist between the familiarity of emotions and the extent to which people want to experience these emotions. We found that, on average, participants wanted to experience happiness and did not want to experience anger and fear. However, the degree to which they wanted to experience these emotions differed significantly as a function of how familiar these emotions were to the individual. People more (vs. less) familiar with an emotion wanted to experience this emotion more. This was the case regardless of whether the emotion was pleasant or unpleasant to experience. Such associations were demonstrated across two samples in two cultures and could not be fully explained by concurrent emotional experiences. We also found that familiarity was a more powerful predictor of the emotions people want to feel, compared to trait emotions. This pattern of findings helps to disambiguate the previously conflicting literature regarding the links between trait emotions and familiarity.

Although there are many reasons to expect familiarity of emotions to be linked to wanting emotions, the current pattern of findings is inconsistent with some possible accounts and is more consistent with others. In particular, our findings suggest that links between familiarity and wanting emotions may not result from a process of habituation: the more people tend to experience pleasant emotions, the more they want to experience them. Our findings are also inconsistent with the possibility of hedonic sensitisation: although one might expect people who often feel bad to be particularly motivated to avoid bad feelings, our findings show that the more familiar people are with an emotional experience, whether pleasant or unpleasant, the more desirable they find it to be.

Why are familiar emotions more desirable?

People typically want things that they like (Berridge, 1996). Indeed, our findings show that in most cases, people for whom an emotion was relatively more (vs. less) familiar tended to like that emotion more. For instance, people who tended to experience more anger found this emotion more, rather than less, pleasant to experience. Liking, however, did not always account for the links between familiarity and wanting. In the case of cheerfulness (Pilot Study) and happiness (Study 1), liking did mediate the link between familiarity and wanting. However, in the case of anger (Pilot Study and Study 1) and fear (Study 1), people wanted to experience more familiar unpleasant emotions regardless of how much they liked them. Although further evidence is needed, this suggests that the mechanism linking familiarity to wanting may differ by valence. Specifically, liking may play some role in the link between familiarity and wanting in the case of pleasant emotions, but not in the case of unpleasant emotions.

The fact that liking does not fully account for the link between familiarity and wanting emotions is consistent with the idea that liking and wanting involve separate neural mechanisms (Berridge & Robinson, 2003). It is therefore possible to want something whether one likes it or not. Still, if liking does not explain the link between familiarity of emotions and wanting to experience them, particularly with regard to unpleasant emotions, then what does? Building on an instrumental approach to emotion regulation (e.g., Tamir, 2009), we suggest that people may be motivated to experience familiar emotions because such emotions offer certain benefits that are not necessarily hedonic in nature.

First, familiar emotions may offer cognitive benefits. For example, they may be easier to experience. Due to processing fluency, that which is often experienced tends to be more easily experienced in the future (Reber, Schwarz, & Winkielman, 2004). As such, perhaps experiencing familiar emotions is less distracting and effortful. Second, familiar emotions may offer motivational benefits. For example, they may facilitate goal pursuit. If familiar emotions promote goals that are important to the individual, experiencing such emotions may help people achieve their goals. Indeed, we have previously found that people high in neuroticism, who often prioritise avoidance goals, show stronger preferences for worry and fear in potentially threatening situations (Tamir, 2005). Finally, familiar emotions may offer epistemic benefits. According to self-verification theory, people strive to verify their self-perceptions (e.g., Swann, Chang-Schneider, & Angulo, 2008; Wood, Heimpel, Manwell, & Whittington, 2009). Such patterns may also apply to self-relevant emotional experiences such that experiencing familiar emotions might help people confirm their emotional self-concept (e.g., Tamir & Robinson, 2004; Tamir, Robinson, & Clore, 2002; Wood et al., 2009). Although our findings do not speak directly to these possibilities, they suggest that these and other possibilities should be directly tested.

Implications and future directions

The present findings have several important implications. Because emotional preferences shape emotion regulation and subsequent emotion experience (e.g., Tamir & Ford, 2012), people for whom an emotional experience is more familiar may be more likely to actively maintain that experience. This implies, perhaps ironically, that people may actively contribute to their own pleasant and unpleasant emotional experiences (see also Borkovec & Roemer, 1995). Although potentially provocative, this idea is consistent with our data and testing it in future research could have pragmatic implications.

Because our design was correlational, our data did not address the causal direction of the link between familiarity and wanting emotions. Another possibility, therefore, is that the more people want to experience an emotion, the more they try to maintain or increase that emotional experience, and thus the more they experience it (i.e., the more familiar with it they become) as a result. It is also possible that both directions operate simultaneously (namely, familiarity breeds wanting, and wanting breeds familiarity). To test these possibilities, it would be useful to employ longitudinal designs that could examine changes in familiarity and wanting over time.

Future research should also address some of the limitations of the present studies. First, our measures of wanting relied on self-report. In the future, it may also be useful to include behavioural indices of emotional preferences (e.g., Tamir & Ford, 2009, 2012). Second, we focused on happiness (or cheerfulness), anger, and fear. In the future, it would be useful to extend our findings by examining other emotions as well (e.g., guilt or pride). Finally, our studies focused on healthy populations. However, the familiarity of emotional experiences is an important defining feature of several clinical disorders (e.g., mood disorders). An important question, therefore, is whether the current patterns also apply to clinical populations. For example, are people who are more anxious more motivated to maintain their anxiety compared to those who are less anxious (e.g., Borkovec & Roemer, 1995)? Future research could explore such possibilities and uncover the relationships between the emotions people typically feel and the emotions they want to feel.