Longitudinal Associations Between Teasing and Health-related Quality of Life Among Treatment-seeking Overweight and Obese Youth

Chad D. Jensen; PhD ; Ric G. Steele; PhD; ABPP

Disclosures

J Pediatr Psychol. 2012;37(4):438-447. 

In This Article

Results

Demographic and anthropometric statistics are displayed in Table I. Means and standard deviations of primary study variables are presented in Table II. A preliminary analysis of variance indicated that no significant mean differences in teasing or HRQOL existed across ethnic groups at any of three measurement occasions.

Measurement Model

Preliminary confirmatory factor analyses were conducted to determine whether the two latent constructs of interest were measured comparably across measurement periods (i.e., pretreatment, posttreatment, 1-year follow-up; configural invariance). The initial, freely estimated model constructed using child-report Teasing and HRQOL data, demonstrated acceptable fit [χ2 (102, n = 93) = 151.20, p = .001, RMSEA = .062, NNFI = .95, CFI = .97], suggesting that the pattern of free and fixed parameters was equivalent across measurement occasions. Next, the loadings of the indicators on the latent constructs were equated across measurements to test for weak factorial invariance, a model that also demonstrated acceptable fit [χ2 (110, n = 93) = 156.78, p = .002, RMSEA = .06, NNFI = .95, CFI = .97]. No significant changes in model fit were observed based on the RMSEA model test (i.e., does the RMSEA value of the nested model fall within the 90% RMSEA confidence interval of the comparison model?; Little, 1997). Finally, tests of strong factorial invariance were conducted by equating the intercepts. Results indicated that the model fit was acceptable [χ2 (118, n = 93) = 168.69, p = .001, RMSEA = .57, NNFI = .95, CFI = .97]. Again, no significant changes in model fit were observed based on the RMSEA model test. Taken together, these tests indicate that the constructs included in the model (WRT and CT) were invariant when measured across time periods.

Measurement invariance was established for parent-report measures of quality of life using an identical procedure. The initial CFA model including parent reports of HRQOL and child reports of Teasing demonstrated excellent model fit, χ2 (102, n = 93) = 115.22, p = .175, RMSEA = .02, NNFI = .99, CFI = .99. The subsequent test of weak factorial invariance also indicated close model fit, χ2 (114, n = 93) = 147.12, p = .020, RMSEA = .04, NNFI = .97, CFI = .98. No significant changes in fit were observed based on the RMSEA Model Test. Finally, the strong factorial (i.e., intercept) invariance model demonstrated acceptable fit, χ2 (122, n = 93) = 163.91, p = .006, RMSEA = .05, NNFI = .97, CFI = .97. Similarly, no significant difference in model fit was observed between weak and strong invariance models based on the RMSEA Model test.

Structural Models

After establishing measurement equivalence over time, associations between latent constructs (i.e., Teasing and HRQOL) over time were examined.

Child Report The first structural model, constructed using child-report data, demonstrated close model fit and no significant difference in model fit was detected relative to the weak invariance model [χ2 (114, n = 93) = 167.70, p = <.001, Δχ2 (4, n = 93) = 4.98, p > .25, RMSEA = .059, NNFI = .99, CFI = .97], indicating that the specified directional associations are tenable. Consistent with the first study hypothesis, child-reported Teasing and HRQOL remained significantly negatively correlated at each of the three assessments.

Next, the study hypothesis postulating a bidirectional association between Teasing and HRQOL (i.e., teasing would predict HRQOL and HRQOL would predict teasing prospectively) was examined. To test this hypothesis, longitudinal structural paths were specified between Time 1 Teasing and Time 2 HRQOL and between Time 2 Teasing and Time 3 HRQOL. Teasing at Time 1 did not significantly predict HRQOL at Time 2 nor did Teasing at Time 2 predict HRQOL at Time 3 (β = .07, p > .05; β = .039, p > .05, respectively). However, findings indicated that HRQOL at Time 1 significantly predicted Teasing at Time 2 (β = .33, p < .01) and HRQOL at Time 2 predicted Teasing at Time 3 (β = .40, p < .01), with higher HRQOL predicting lower levels of teasing. In summary, the hypothesized predictive relationship with Teasing predicting HRQOL was not supported. However, the hypothesis that HRQOL would predict subsequent teasing was confirmed.

Finally, the hypothesis predicting that both latent constructs would remain temporally stable was evaluated by examining the autoregressive pathways within constructs at each measurement occasion. In these analyses, Teasing at Time 1 significantly predicted teasing at Time 2 (β = .35, p < .01) and Time 2 teasing also predicted Teasing at Time 3 (β = .35, p < .01). Similarly, Time 1 HRQOL significantly predicted HRQOL at time 2 (β = .50, p < .01) and Time 2 HRQOL predicted HRQOL at Time 3 (β = .41, p < .01).

Since hypothesized latent regressions with teasing predicting QOL were not significant, these pathways were removed one at a time, resulting in a close-fitting final structural model [χ2 (116, n = 93) = 162.23, p < .005, RMSEA = .058, NNFI = .96, CFI = .97] presented in Figure 1. Again, no significant difference in model fit was detected relative to the weak invariance model [Δχ2 (6, n = 93) = 5.14, p > .25].

Parent Report Next, an identical structural model was specified for parent report of HRQOL and child report of teasing. This model demonstrated close model fit (χ2 (114, n = 93) = 133.90, p = .09, RMSEA = .027, NNFI = .99, CFI = .99), however, none of the specified cross-lagged regressive paths were significant. These paths were subsequently removed, yielding a close-fitting structural model [χ2 (118, n = 93) = 138.41, p = .09, RMSEA = .028, NNFI = .99, CFI = .99]. No significant difference in model fit was detected relative to the weak invariance model [Δχ2 (8, n = 93) = 8.71, p > .25].

Comments

3090D553-9492-4563-8681-AD288FA52ACE
Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as:

processing....