Psychometric Characteristics of Strengths Knowledge Scale and Strengths Use Scale Among Adolescents

Introduction

Character strengths are relatively stable and positive patterns of cognition, emotion, and behaviors were manifested by individuals (Peterson & Seligman, 2004). Recently, character strengths were introduced into positive education (Seligman, Ernst, Gillham, Reivich, & Linkins, 2009), for the positive correlation with academic performance and class-related behavior among students (International Positive Education Network, 2016).

Studies demonstrated that characters are positively associated with grade point average (GPA; Lounsbury, Fisher, Levy, & Welsh, 2009). Interventions based on character strengths have been constructed to facilitate the emotional and cognitive development of students. Quinlan, Swain, and Vella-Brodrick (2012) proposed that strengths knowledge and strengths use may be the underlying mechanisms of how strengths-based interventions work and promote individual development. Self-determination theory suggests that both strengths knowledge and strengths use can satisfy the fundamental psychological needs of individuals and hence raise self-motivation and well-being (Ryan & Deci, 2000). Strengths knowledge refers to the awareness and recognition of one’s strengths, and strengths use means using one’s strengths intentionally (Govindji & Linley, 2007). Accordingly, Govindji and Linley (2007) developed the Strengths Knowledge Scale (SKS) and Strengths Use Scale (SUS) to measure the extent of recognizing and using strengths. Both scales show good psychometric characteristics in Western countries (Wood, Linley, Maltby, Kashdan, & Hurling, 2011). However, the situation among adolescents has not been reported in existing literature.

The present study evaluated the psychometric properties of the abovementioned scales by the exploratory structural equation modeling (ESEM). ESEM combines the framework and benefits of exploratory factor analysis, confirmatory factor analysis, and structural equation model (Marsh, Liem, Martin, Morin, & Nagengast, 2011). Compared with confirmatory factor analysis, ESEM has better model–data fit, and more accurate interfactor correlations in the framework (Marsh et al., 2011; Marsh et al., 2013). As positive education and corresponding interventions can promote cognitive and emotional development of students, we hypothesized that strengths knowledge/strength use positively relates to GPA, flourishing, and positive affect (PA), but negatively relates to negative affect (NA) and psychological symptoms.

Method

Results and Discussion

The ESEM analyses were conducted by Mplus 7.4 with the robust maximum likelihood (MLR) estimation. An initial analysis revealed that the goodness-of-fit indices of the 22-item and two-factor related structure achieved adequate fit (χ² = 615.342, df = 188, comparative fit index [CFI] = .911, Tucker–Lewis index [TLI] = .891, standardized root mean square residual [SRMR] = .037, root mean square error of approximation [RMSEA] = .072, 90% confidence interval [CI] = [.066, .078]). However, Item 2 showed insignificant loading on the strengths knowledge factor (loading = −0.007), and the modification indices showed a strong residual covariance between Item 10 and Item 11 (Modification Index = 92.98, Expected Parameter Change = 0.278). The reason for the insignificant loading of Item 2 was wording effects which may confound the factor structure of the scale (Ye, 2009). Negatively worded items in the questionnaire will generate wording effects which are affected by age and reading ability (Marsh, 1996). Thus, a revised model that excluded Item 2 and included the residual covariance between Item 10 and Item 11 was constructed. The goodness-of-fit indices of the revised 21-item structure were significantly improved (χ² = 484.887, df = 168, CFI = .931, TLI = .914, SRMR = .035, RMSEA = .065, 90% CI = [.059, .072]). The standardized factor loadings on SKS ranged from .51 to .93, and that of SUS ranged from .52 to .91. The average variance extracted (AVE) of SKS and SUS are .60 and .61, respectively. The final model also demonstrated good internal consistency reliability for the SKS (ω = .925) and SUS (ω = .959).

The item-rest correlations ranged from .65 to .84. Descriptive statistics and correlations of the strengths knowledge, strengths use, and variables are shown in Table 1. Strengths knowledge and strengths use are positively related to flourishing, PA, and GPA (r = .15 to .66, p < .001), whereas negatively related to NA, depression, anxiety, and stress (r = −.10 to −.32, p < .001). Specifically, strengths knowledge demonstrates stronger correlations to other variables than strengths use, except PA. Multiple regression analysis was conducted to identify the significant contributors to academic achievement.

OPEN IN VIEWER

Table 1.

Descriptive Statistics and Correlations of the SKSU and Other Criteria-Related Variables.

	1. SK	2. SU	3. Flourishing	4. PA	5. NA	6. Depression	7. Anxiety	8. Stress	9. GPA
1	—	.73***	.66***	.43***	−.22***	−.32***	−.27***	−.26***	.22***
2		—	.61***	.50***	−.10*	−.23***	−.16***	−.19***	.15**
3			—	.43***	−.28***	−.44***	−.32***	−.32***	.10*
4				—	.20***	−.07	.02	.01	.12*
5					—	.60***	.62***	.59***	<−.01
6						—	.84***	.83***	−.05
7							—	.88***	−.02
8								—	−.02
M	4.88	4.73	5.31	2.88	1.95	0.59	0.73	0.82	77.45
SD	0.80	0.97	1.01	0.77	0.76	0.65	0.62	0.62	11.31

Note. SK = strengths knowledge; SU = strengths use; PA = positive affect; NA = negative affect; GPA = grade point average.

*

p < .05. **p < .01. ***p < .001.

The results in Table 2 were straightforward. Strengths knowledge was the only significant factor (B = 3.84; t = 3.59, p < .001) in explaining GPA, and the whole model contributed 5.50% (F = 3.11, p < .01) of the variances. These results highlighted the significance of strengths knowledge. However, other studies implied that strengths use was more important than strengths knowledge for well-being. Among our results, the closer association between GPA and strengths knowledge might result from that both of them belong to cognitive aspect (Govindji & Linley, 2007; Richardson, Abraham, & Bond, 2012), while others results might be due to the lack of direct and detailed guidelines on how to use strengths. Future research should be cautious about the guidelines. Also, further studies should be conducted to examine the temporal and causal relationships among strengths knowledge, strengths use, and GPA. Another finding that required further discussion was the positive correlation (r = .20, p < .001) between PA and NA. Watson et al. (1988) assumed that PA and NA were separate and independent constructs in the original mode. However, further studies in different cultures found negative (Jovanović & Gavrilov-Jerković, 2016) or insignificant correlation (Zhang, Diao, & Constance, 2004) between them. These inconsistent results implied that culture might moderate the relationship between PA and NA (Schimmack, 2009). In this study, NA positively associated to depression, anxiety, and stress, whereas PA showed insignificant correlations with them. It might suggest that the Chinese are not good at expressing positive emotions (Diener, Suh, Smith, & Shao, 1995). Accordingly, future studies should incorporate cultural factors in studying the relationship between PA and NA. Finally, the participants were recruited from 10 different schools, but the analyses did not take the sample dependency into consideration. Further studies may adopt multilevel modeling to handle the nesting of students within the classes/schools.

OPEN IN VIEWER

Table 2.

Regression of Psychological Variables on GPA.

Independent variables	Dependent variable: GPA
	β	t	VIF
Strengths Knowledge	3.842	3.593***	2.602
Strengths Use	−0.253	−0.295	2.479
Flourishing	−0.970	−1.223	2.257
Positive Affect	0.526	0.597	1.639
Negative Affect	0.473	0.487	1.927
Depression	−1.724	−1.007	4.424
Anxiety	1.022	0.502	5.702
Stress	0.523	0.272	5.066
R2 (F)	0.055 (3.106**)

Note. O’Brien (2007) suggested that the VIF bigger than 10 indicated the serious multi-collinearity issue of multiple regression model. GPA = grade point average; VIF = variance inflation factor.

**

p < .01. ***p < .001.