Wanting Less Closeness in Romantic Relationships

Abstract

Although the experience of closeness has received considerable attention in the close relationships literature, wanting less closeness with a romantic partner is less understood. Study 1 identified the lay understanding of what it means to want less closeness in a romantic relationship. In Studies 2 and 3, participants primed with descriptors of wanting less closeness, compared to those primed with control words, reported wanting significantly less closeness in their current relationship. As would be expected from models of adult attachment, those high on attachment-related avoidance responded more strongly to the experimental manipulation than those low on this dimension.

ACKNOWLEDGMENTS

Many thanks to Alison Bloom-Feshbach, Setareh Deljo-Zargarani, Samara Thomas, and Will Gardner-Mims for their assistance with collecting and coding data for this report.

Notes

Note. N = 75.

¹The full list of 99 features is available from the first author.

Note. N = 78–80 (some participants failed to rate every word); words rated on a scale from 1 (extremely atypical feature of wanting less closeness) to 7 (extremely typical feature of wanting less closeness).

²To explore the possibility of an underlying prototype structure, we originally included a puzzle populated by words that had received the lowest centrality ratings in Study 1b. Exploratory analyses suggested the concept “wanting less closeness” is not cognitively organized as a prototype. That is, peripheral features also significantly affected desired closeness; if this concept were organized as a prototype, one would expect the central features, but not the peripheral features, to affect desired closeness. To streamline the presentation of results relevant to the questions at the heart of the current article, we include data only from the Too Close and Neutral Control conditions.

³We realized too late that the meaning of “space” and “freedom” are antithetical to the meaning of “need space” and “want freedom.” If anything, their inclusion in the Study 2 and Study 3 priming tasks should weaken the effectiveness of the manipulation and could thus be expected to work against our hypotheses.

Note. For the Neutral Control condition, sample sizes for correlations ranged from 26 to 27; sample size for the Too Close condition was 26. None of the premanipulation variables differed significantly between the two experimental conditions (ps ranged from .17 to .93). Although actual closeness did not differ significantly between conditions (p = .12), desired closeness did, t(50) = 2.34, p = .02.

†p < .10. *p < .05. **p < .01.

⁴An alternate analytic approach predicts the discrepancy between desired IOS and current IOS. In both Study 2 and Study 3, the same key results emerge when using discrepancy scores as when using regressed change. In particular, the main effect for condition and the interaction between condition and avoidance remain significant and in the predicted direction.

Note. Variables included in interactions were centered prior to analysis. Step 1 statistics are from Model 1, Step 2 statistics are from Model 2, and so on. In Study 2, the Neutral Control condition received dummy code 0 and the Too Close condition received dummy code 1. In Study 3, the Negative Control condition received dummy code 0 and the Too Close condition received dummy code 1.

†p < .10. *p < .05. **p = .01.

⁵Although we didn't have any specific predictions regarding attachment-related anxiety, the significant interaction between it and condition is notable (β = .42, p < .05). The pattern of the effect suggests that, for participants in the Too Close condition, attachment-related anxiety was positively associated with a desire for relatively more closeness (simple slope t = 2.83, p < .05); that is, the manipulation—populated with words an anxious person would find threatening to the relationship—created a yearning for connection. In the Neutral Control condition, attachment-related anxiety was not much associated with desired closeness (simple slope t = −.80, ns). This effect did not replicate in Study 3.

⁶To evaluate whether the manipulation influenced actual closeness, we predicted actual closeness using the same set of predictors and covariates as those used when predicting desired closeness. Although condition did not significantly predict actual closeness, condition and attachment-related anxiety interacted to significantly predict actual closeness (p < .01). For individuals in the Neutral Control condition, attachment-related anxiety was positively associated with actual closeness (simple slope t = 2.29, p < .05), but for individuals in the Too Close condition, anxiety was negatively associated with actual closeness (simple slope t = −2.34, p < .05).

Note. For the Negative Control condition, sample sizes for correlations ranged from 42 to 43; for the Too Close condition, samples sizes ranged from 47 to 48. None of the premanipulation variables differed significantly between the two experimental conditions (ps ranged from .44 to .95). Neither actual closeness nor desired closeness differed significantly between the two experimental conditions (ps = .64 and .22, respectively).

†p < .10. *p < .05. **p < .01. ***p < .001.

⁷Although attachment-related anxiety did not interact significantly with condition, as occurred in Study 2, it did interact significantly with attachment-related avoidance to predict desired closeness (β = −.19, p < .05). Among people with high attachment-related anxiety, also being high in attachment-related avoidance was negatively associated with desired closeness (simple slope t = −2.33, p < .05); this association was less pronounced among people low in attachment-related anxiety (simple slope t = −1.79, p < .10). This effect was not present in Study 2.

⁸To evaluate whether the manipulation influenced actual closeness, we predicted actual closeness using the same set of predictors and covariates as those used when predicting desired closeness. Neither condition nor its interactions with avoidance and anxiety significantly predicted actual closeness.