ABSTRACT
Gratitude reinforces social bonds. This relationship-regulating function depends on whether and how it is expressed. People can express gratitude in different ways: Beneficiaries may emphasise how they profited from a benefit (self-benefiting) or focus on the benefactor's actions and characteristics related to it (other-praising). What underlies these expressive styles remains unclear. Based on findings that other-praising gratitude expressions have unique positive effects on interpersonal relationships, four studies (N = 1,188) investigated a novel antecedent of these expressions: unexpectedness of the benefit. In Study 1, we content-coded participants’ thank-you notes for an actual Christmas present. Path modelling revealed that unexpectedness of the benefit predicted other-praising, whereas happiness with the present predicted self-benefiting. These results were robust to relevant covariates and mirrored by participants’ self-reported self-benefiting and other-praising intentions. Studies 2–4 (preregistered) investigated samples from two different populations and experimentally manipulated (un)expectedness of recalled or imagined benefits. Given mixed experimental results, we conducted an internal meta-analysis. Across experimental studies, unexpected benefits increased other-praising, albeit weakly so, but not self-benefiting. In addition, the effect of unexpectedness on other-praising was significantly different from that on self-benefiting. We discuss potential processes and moderators of the effect of unexpected benefits on gratitude expressions.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Notes
1 The coder was instructed that expressions of surprise about the present did not form part of self-benefiting. However, if at all, surprise should enhance self-benefiting codings through the self-focus involved, thus working against our hypothesis.
2 Confirmatory factor analyses are reported in the SOM.
3 Preregistered as secondary analyses, we report mixed-measures ANOVAs in the SOM (pp. 7, 12, 18). In none of the experimental studies did the condition by expression interaction term reach statistical significance, ps ≥ .185.
4 Even though a preregistration was submitted for this study on Aspredicted.org, the first author accidentally failed to approve it following the other authors’ approvals. A screenshot of the (intended) preregistration, as approved by the other authors, is available in the SOM.
5 A fixed-effects analysis and separate univariate meta-analyses of the effects of (un)expectedness on other-praising and self-benefiting yielded highly similar results.