Withdraw or affiliate? The role of humiliation during initiation rituals

Abstract

Initiation rituals can take different forms and empirical evidence is inconsistent as to whether these rituals promote affiliation among novices. We argue that experienced humiliation during initiations leads to less affiliation among novices, in particular when one is initiated as sole group member rather than as part of the group. We examined this hypothesis in three studies, using different paradigms. In Study 1 (N = 123), perceived severity of an initiation in the past was associated with lower affiliation with other novices; this relationship was mediated by experienced humiliation. Study 2 (N = 64) showed that public derogation in the lab led to more humiliation when participants were the only victim than when they were derogated as a group. Study 3 (N = 248), a vignette study, showed that a similar effect of social context was mediated by expected support from other novices. We conclude that severe initiations may, due to experienced humiliation, result in less rather than more affiliation with fellow novices.

Keywords:

• Humiliation
• Initiation rituals
• Affiliation
• Social support
• Hazing

Acknowledgements

We thank Leyi Auyeung, Carly van Drie, Loes Halfschepel, Jolien Kater, Anita Nyari, Greet Westerbeek and Nathalie de Zwart for their help with data collection (Study 1 and Study 2) and the content analysis (Study 1), and Prof. Dr Hilde Huizenga for her help with the analyses (Study 2).

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ Sixteen of these respondents stopped filling out the questionnaire when they were asked for a description of their most humiliating experience during the initiation. We did use the first part of their data. There were also 30 respondents who opened the questionnaire, but only filled out the first demographic questions and then quit. Data of these participants were not used.

² In addition, we measured several other variables in all three studies, namely state self-esteem, group identification, different emotion labels and behavioural intentions and general questions about (attitudes of) membership in fraternities/sororities and involvement in hazing practices. We do not report the results for these variables; however, data of these measures may be obtained from the first author upon request.

³ We asked the same questions with regard to members of the fraternity/sorority in general, also during the initiation and at this moment (Cronbach's α = .84 and .79, respectively). Results may be obtained from the first author upon request.

⁴ In all three studies we also measured anger and tendencies to aggress against the initiator (Study 1 and 3) or derogator (Study 2). Correlations between these measures and humiliation are all positive and significant, ranging from r = .24 to r = .75.

⁵ These results are not reported in the present study because in this condition psychological phenomena other than the one we are focusing on in the current research may be at work. However, data can be requested from the first author. Importantly, all analyses show the same pattern if this condition is included.

⁶ We did this for ethical reasons, but also to make the episode more credible, as we feared that a very derogative comment would have probably raised suspicion in this setting.

⁷ We also included a behavioural measure of affiliation (distance of participant chair to confederate chair). However, we faced some problems with the measurement of this variable and the data very likely became inaccurate. When included in the analyses, we found no effect on this measure and no relation with other variables.

⁸ Although we are aware of differences between embarrassment and shame, we decided to use both emotion labels in one scale. We had two reasons for this. First, in Dutch, these emotion labels have more similar meanings than in English. This was also confirmed by a strong positive relation between these items (r = .48, p < .001). Second, the current research was not aimed at testing possible differences between shame and embarrassment in the context of initiations.

⁹ If participants answered these questions incorrectly, we suspected that they did not pay enough attention to our manipulation (of which one runs a higher risk in the case of Mechanical Turk participants). On the basis of this check we discarded data of 10 participants (see “Participants and Procedure”).

¹⁰ We could not use the exact same scale as in Study 2 as the situations were very different and some items used in the lab study were not relevant for the vignette-study. Furthermore, the questionnaire had to be short(er), as participant fatigue can pose a threat in M-Turk studies. However, five of the seven items used in Study 3 were the same as in Study 2. When we analysed the data in Study 1 and Study 2 using this 5-item humiliation scale instead of the longer item scales, we found the same results, except for the mediation of the effect of Social Context on withdrawal tendencies by humiliation in Study 2. This mediation is not significant using the 5-item scale. Thus, although we replicate this mediation in Study 3, we should interpret this finding with some caution.

Additional information

Funding

This research was partly funded by a grant from the European Union awarded to the University of Amsterdam (amongst others) for their work on the European FP-7 project SAFIRE (Project Reference: 241744).