Abstract
The role that the involvement of parents may play in the treatment outcome of their children with anxiety disorders is still under debate. Some studies dealing with other disorders have examined the role that the expressed emotion (EE) construct (parental overinvolvement, criticism and hostility) may play in treatment outcome and relapse. Given that some of these aspects have been associated with social anxiety for a long time, it was hypothesized that EE may be associated with lower treatment outcome. The sample was composed of 16 adolescents who benefited from a school-based, cognitive-behavioural intervention aimed at overcoming social anxiety. Then, parents were classified with high or low EE. The results revealed that the adolescents whose parents had low EE showed a statistically significant reduction of their social anxiety scores at posttest, as opposed to adolescents of parents with high expressed emotion. These findings suggest that parental psychopathology (parents with high EE) should be taken into consideration to prevent poor adolescent treatment outcome.
Results
Non-parametric tests were used given: a) the limited number of youth–parent dyad in each condition (6 parents with high EE, and 10 parents with low EE); b) results from the Shapiro–Wilk test indicating that data are not from a normally distributed population; and c) the inequality of variance, as revealed by Mauchly's sphericity and Levene's tests.
First, between group differences were tested using the Mann–Whitney U test. Results revealed the absence of significant differences between the two groups on change scores (p = .19). Within group differences were also analyzed. Wilcoxon signed-rank test was used. Means and standard deviations (in parenthesis) for each group and measures are as follows: on the SAS-A/Total score, the group composed of adolescents with parents with low EE showed means of 66.6 (9.42), 47.7 (10.57) and 47 (12.29) at pretest, posttest and 6-month follow-up, respectively. On the SPAI-B, M = 40.66 (8.22), 27.96 (15.78) and 24.98 (11.96) at pretest, posttest and 6-month follow-up, respectively. For adolescents with parents with high EE, the scores at pretest, posttest and 6-month follow-up on the SAS-A were as follows: 64.16 (8.79), 50.2 (13.02) and 50 (12.67). Similarly, on the SPAI-B, the means and standard deviations were 38.61 (5.11), 26.64 (13.07) and 25.02 (12.01) at pretest, posttest and 6-month follow-up.
Within group differences analyses results revealed that the adolescents whose parents had low expressed emotion showed a statistically significant reduction of their social anxiety scores at posttest, as opposed to adolescents with high EE parents. Particularly, adolescents with parents with low EE reported significant differences on the SAS-A/Total score, Z = −2.8, p < .05, and the SPAI-B, Z = −2.91, p < .05. Furthermore, the effect size for these differences was high. We have adopted the criteria proposed by Cohen (1988), in which .2 means a small effect size, .5 means a medium effect size and .8 means a large effect size. A large effect size allows statistical significance with no hazard for the sensitivity of the research. Effects sizes were high: 1.81 and .97 for the SAS-A/Total score and the SPAI-B, respectively. Contrary to adolescents with parents with high EE, those with parents' low EE also scored significantly lower at the 6-month follow-up as opposed to pretest on the SAS-A/Total score, Z = −2.7, p < .01, and the SPAI-B, Z = −2.31, p < .05. No significant differences at posttest as opposed to follow-up were observed, but reduction in scores was maintained at 6-month follow-up.
