Parent and adolescent conversations about hurt: How interaction patterns predict empathic accuracy and perceived understanding

ABSTRACT

Although hurt feelings might be inevitable in parent–adolescent relationships, follow-up conversations provide an opportunity for reconciliation and learning. This research considers how interaction patterns affect empathic accuracy and perceived understanding by examining the conversations of 98 parent–adolescent dyads regarding an event when adolescents felt hurt by something their parent said or did. Analysis of the interactions revealed three distinct interaction patterns: parental probing/adolescent withdrawal, mutual confrontation, and supportive listening. Although empathic accuracy remained low for all groups, supportive listening was associated with greater empathic accuracy for children and greater perceived understanding for both parents and adolescents. Analysis of specific thoughts reported during video review of the interactions revealed that parents often over-attributed negative thoughts to children, especially parents in the probing/withdrawal group.

KEYWORDS:

• Empathic accuracy
• perceived understanding
• parent–adolescent relationships
• hurt feelings

Acknowledgements

The authors would like to thank several research assistants who helped in the collection and coding of these data: Melea Andrys, Alexandra Baltimore, Ashley Bente, Nicole Cihla, Brooke Duval, Clairissa Holmes, Kelly Hogan, Alyssa Kaczor, Ariel Kleinschmidt, Kristen Kubichan, Lauren McGrath, Jamie Patrick, Joshua Pederson, Allison Reiter, Lindsay Schmidt, Audrey Shelton, Allison Ulstad, and Bennett Wiltfang.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes on contributors

Rachel M. McLaren is an associate professor in the Department of Communication Studies at the University of Iowa (Ph.D., Pennsylvania State University).

Alan Sillars (Ph.D., University of Wisconsin) is a professor emeritus in the Department of Communication Studies at the University of Montana.

Notes

1 Sillars et al. (2010) had parent–adolescent triads discuss three conflict topics (one each chosen by child, father, and mother) over a 15-minute period. All but a small number of families completed three topics. Thus, we judged 5 minutes per topic to be sufficient.

2 In retrospect, the second item (“ … did not see things the same way”) could assess perceived agreement as well as perceived understanding. Given the relatively low internal consistency of the two-item scale for adolescents (α = .55), we repeated all analyses using the more straightforward item (“ … understood each other”) as a single-item measure. As there were only minor differences in results, we retained the two-item scale for perceived understanding.

3 Please contact the first author for detailed training manuals for account interaction coding.

4 Ratings for confirmation, warmth, and involvement were based on description of these behaviors by Koenig Kellas and Trees (2005). Although these authors distinguish confirmation and attentiveness as two aspects of perspective-taking, raters had difficulty separating these concepts, so we combined elements of attentiveness and confirmation into a single rating for confirmation.

5 In a separate analysis of these data (McLaren & Sillars, 2014), observers rated role consensus in the conversations (i.e., mutual affirmation of victim-perpetrator roles). Role consensus was substantially lower in the mutual confrontation cluster (M = 2.5) than in supportive listening (M = 4.1) and probing/withdrawal (M = 3.8) clusters, F(2, 93) = 48.33, p < .001.

6 O'Keefe (2003) argues that alpha adjustments based only on the number of tests reduces statistical power, leading to Type II error, and creates absurd inconsistencies. For example, two researchers analyzing the same data set, but who perform a different number of tests, can reach opposite conclusions when rejecting the null hypothesis. Similarly, a researcher who performs 10 tests at different times (or who analyzes these tests simultaneously but reports them in different articles) retains greater statistical power than a researcher who performs and reports all tests at once. Whereas O’Keefe argues that family-wise alpha adjustments based on the number of tests are never appropriate, Tutzauer (2003) argues that alpha should only be adjusted in specific circumstances, mainly when making a strong claim that requires all hypotheses within a theory to be reported supported.

7 In this study, adolescent information-giving correlated positively with parental confirmation of adolescents (r = .36, p < .001) and adolescent confirmation of parents (r = .46, p < .001).

Additional information

Funding

This study was funded by an internal grant from the University of Iowa.