Abstract
This article presents three studies furthering validity evidence for a self-report measure of active-empathic listening (AEL). Study 1 investigates the temporal stability of the AEL scale, revealing a statistically sound model with no decline in general fit over time, supporting the scale's measurement of an individual trait-like difference. Studies 2 and 3 investigate the contribution of trait-level AEL and various characteristics of situations to the utilization of AEL. A general discussion focuses on areas for future research with respect to how AEL might help (or hinder) the development and maintenance of close, personal relationships.
Acknowledgments
The authors would like to thank Terrance Brown, Sierra Crump, Theresa MacDonald, and Kenissa McKay for their assistance with data collection.
Notes
Note. Means with different subscripts are significantly different (p < .001) within columns.
Note. Estimated marginal means with different subscripts are significantly different at p < .01.
aParticipants could check multiple boxes for these questions. Percentage represents the percentage of participants that did check a given box.
We use the term temporal stability instead of test-retest reliability because we are primarily interested in the consistency of scores across time. As stated by DeVellis (Citation2003), “[Referring] to invariance in scores overtime as temporal stability is preferable because it does not suggest, as does test-retest reliability, that measurement error is the source of any instability we observe” (p. 44, emphasis in original).
Profile analysis can also test for flatness of profiles which, in our study, ascertains whether situational AEL scores differ across situations. The profile analysis returned results statistically and substantively identical to those obtained by the SEM analysis.
The multivariate effect combines the linear, quadratic, and cubic effects similar to a standard ANOVA. Thus, inspection of the linear trend irrespective of an overall multivariate effect is analogous to testing specific contrasts of interest as outlined in several notable publications (Rosenthal, Rosnow, & Rubin, Citation2000). This approach is preferable for a variety of reasons, not the least of which is that it is more powerful than the omnibus approach. Given the low sample size for this analysis, we feel confident in interpreting a p-value below .10, though we are in no way engaging in “wishful thinking” (O'Keefe, Citation2007).