Social Network Site Use Motives Scale (SUMS) was developed under the assumption that it consists of six factors, but only four factors were extracted as a result of Exploratory Factor Analysis. The goal of the present study was to investigate whether SUMS consists of four or six factors using Confirmatory Factor Analysis (CFA) and Exploratory Structural Equation Modeling (ESEM) approach.
A Korean college student sample (n=600; mean age, 21 years; 58% female) filled out the SUMS and the Social Network Site Addiction Proneness Scale. CFA and ESEM were used to assess the factor structure of the SUMS.
Results indicated that a four-factor solution to the SUMS had inadequate fit in the sample examined using both CFA and ESEM and a six-factor solution to the SUMS had insufficient fit using CFA, whereas fit was optimal using ESEM for the six-factor model. In addition, the scale showed adequate convergent validity and reliability.
These findings support the six-factor model of SNS use motives and suggest that ESEM is a more appropriate method than CFA for examining the factor structure of the SUMS. The results displayed the usefulness of the ESEM framework in the investigation of use motives.
A social networking site (SNS) (Facebook, Twitter, Instagram, etc.) is an online vehicle for creating self-descriptive profiles, communicating with friends, and meeting other people who share an interest [
In recent years, SNS addiction has become prevalent in many countries. Cheng et al. [
It has been reported that SNS addiction may cause academic problems such as low academic achievement [
Previous studies indicated that use motives are important to understand and intervene SNS addiction [
The SNS Use Motives Scale (SUMS) [
Despite its appeal as a potential measure of SNS use motives, Shin and Lim’s factor analytic findings [
The factor structure of the SUMS was assessed with exploratory factor analysis (EFA) in the previous study [
However, the CFA where items are hypothesized to load on their respective factors without allowing cross-loadings onto any of the other factors could lead to low model fit indices and inflated inter-factor correlations, limiting the discriminant validity of the scale [
The final objective of this study was to examine the convergent validity, internal consistency, and test-retest reliability of the SUMS. For convergent validity, the relations between the SUMS factors and SNS addiction will be tested. According to previous studies of SNS use motives [
Using the convenient sampling method, six hundred college students were recruited from introductory psychology courses at a university located in Gyengbuk, South Korea. The ages of the participants ranged from 18 to 28, with a mean age of 21.17 (SD=2.06). The participants were 41.7% male and 58.3% female. This study was approved by the Daegu University Institutional Review Board (IRB # 2020-022-08).
The SUMS is a measure of SNS use motives [
The SAPS is a measure of SNS addiction [
Data collection was performed offline in 2018–19. Written consent was sought from each participant before data collection and the respondents completed a battery of self-administered questionnaires. Data collection was performed in classrooms in groups of about 40 people using pen or pencil, and a researcher was available to answer any queries arising from the questionnaires. Participants were assured of anonymity and confidentiality and were free not to take part in the study. The average time to complete the battery of questionnaires was about 15 to 20 minutes. Debriefing was conducted after completion of the questionnaires. As a reward for participating in the study, 1,000 won worth of school supplies were provided to the participants.
Data analyses were carried out using SPSS 21 (IBM Corp., Armonk, NY, USA) and Mplus 7.0 (Muthén & Muthén, Los Angeles, CA, USA). The factor structure was examined using CFA and ESEM with maximum likelihood estimation which assumes that the variables are continuous and follow a multivariate normal distribution. In the present data, skewness of all variables were found to be between -2 to +2 and kurtosis of all variables were found to be between -7 to +7, thus it can be considered that these data follow a multivariate normal distribution.
Four alternative models were tested and compared: 1) a six-factor model with CFA; 2) a six-factor model with ESEM (
A minimum cut-off of 0.95 for Comparative Fit Index (CFI), minimum cut-off of 0.95 for Tucker-Lewis Index (TLI), a maximum cut-off of 0.06 for Root Mean Square Error of Approximation (RMSEA), and a maximum cut-off of 0.08 for Standardized Root Mean Square Residual (SRMR) were considered as indication of acceptable fit [
As can be seen in
Correlations between SNS use motives and SNS addiction are shown in
The test of internal consistency produced a Cronbach’s alpha of 0.895 for the enhancement factor, of 0.931 for the conformity factor, of 0.908 for the pastime factor, of 0.877 for the social factor, of 0.868 for the information factor, and of 0.914 for the coping factor. All Cronbach’s alphas were higher than the recommended threshold level of 0.70 (
The aim of this study was to examine whether SUMS consists of four or six factors using CFA and ESEM approach. The results of the present study showed that the six-dimensional model yielded considerably better fit than the four-factor model. These findings support Shin and Lim’s [
The current study showed that items of SUMS had small but significant, secondary loadings on factors other than the intended factor. Because cross-loadings were not constrained to zero in ESEM, the ESEM analyses were consistently found to provide better fit than CFA. These results could convince SNS addiction researchers to consider using ESEM when studying SNS use motives.
Given the presence of cross-loadings in the measurement model of use motives, simply relying on CFA and not using ESEM when investigating the factor structure of SNS use motives can lead to wrong conclusion. For example, poor-fitting CFA solutions may lead researchers to conclude that the sixfactor model cannot be adequate for the SNS use motives. Therefore, simply relying on CFA and not using ESEM can lead to a premature abandonment of otherwise promising theories in the field of SNS use motives and SNS addiction [
Another consequence of the exclusion of cross-loadings is inflated factor correlations. Previous studies with the addiction scales have shown that ESEM consistently results in smaller factor correlations than CFA [
These results are not without limitations. First, this study is self-report study, which could be inflated by common method variance. Second, because the cross-sectional nature of the current study precludes any causal interpretation, studies using longitudinal designs are needed to identify a cause-and-effect relationship between SNS use motives and SNS addiction. Third, because the participants of this study were limited to college students and it was reported that the SNS usage rate in Korean 20s (91.9%) was the highest compared to other age groups [
Despite these limitations, the results demonstrate Shin and Lim’s [
The datasets generated or analyzed during the study are available from the corresponding author on reasonable request.
The author has no potential conflicts of interest to disclose.
This research was supported by Daegu University Research Grant, 2018.
Exploratory structural equation modeling (six factor model). Factor 1, enhancement motive; Factor 2, conformity motive; Factor 3, pastime motive; Factor 4, social motive; Factor 5, information motive; Factor 6, coping motive.
Confirmatory factor analysis (six factor model). All factor loadings are standardized and are statistically significant, p<0.001. Factor 1, enhancement motive; Factor 2, conformity motive; Factor 3, pastime motive; Factor 4, social motive; Factor 5, information motive; Factor 6, coping motive.
Goodness-of-fit from factor analyses
Model | χ2 | df | CFI | TLI | SRMR | RMSEA | 90% CI | AIC | SSABIC |
---|---|---|---|---|---|---|---|---|---|
Four factor | |||||||||
ESEM | 1,681.044 | 321 | 0.899 | 0.863 | 0.035 | 0.084 | 0.080–0.088 | 41,395.397 | 41,608.061 |
CFA | 2,487.365 | 399 | 0.844 | 0.830 | 0.080 | 0.093 | 0.090–0.097 | 42,045.719 | 42,163.050 |
Six factor | |||||||||
ESEM | 703.834 | 270 | 0.968 | 0.948 | 0.017 | 0.052 | 0.047–0.056 | 40,520.188 | 40,795.183 |
CFA | 1,332.670 | 390 | 0.930 | 0.922 | 0.054 | 0.063 | 0.060–0.067 | 40,909.024 | 41,037.355 |
χ2, chi-square goodness of fit test; df, degree of freedom; CFI, comparative fit index; TLI, Tucker-Lewis index; SRMR, standardized root mean square residual; RMSEA, root mean square error of approximation; AIC, Akaike information criteria; SABIC, sample size-adjusted Bayesian information criteria
Parameter estimates for the CFA and ESEM solutions of the SUMS
Item | CFA |
ESEM |
|||||
---|---|---|---|---|---|---|---|
Factor I | Factor II | Factor III | Factor IV | Factor V | Factor VI | ||
1 | 0.748 |
0.157 | -0.041 | -0.001 | -0.010 | 0.665 |
0.014 |
2 | 0.766 |
0.645 |
-0.049 | 0.156 | -0.042 | 0.118 | -0.024 |
3 | 0.640 |
0.028 | 0.023 | 0.109 | 0.665 |
0.019 | -0.155 |
4 | 0.688 |
0.316 |
0.035 | 0.004 | 0.017 | 0.016 | 0.566 |
5 | 0.762 |
0.070 | -0.030 | 0.672 |
0.160 | -0.073 | 0.076 |
6 | 0.786 |
0.038 | 0.893 |
-0.020 | -0.113 | -0.010 | -0.031 |
7 | 0.669 |
0.053 | 0.146 | -0.145 | -0.007 | 0.619 |
0.104 |
8 | 0.772 |
0.570 |
-0.014 | 0.105 | 0.079 | 0.194 | -0.051 |
9 | 0.834 |
-0.043 | -0.024 | 0.019 | 0.854 |
0.073 | -0.001 |
10 | 0.558 |
0.107 | 0.235 | -0.047 | -0.015 | -0.006 | 0.626 |
11 | 0.837 |
0.051 | 0.044 | 0.818 |
0.005 | 0.002 | -0.036 |
12 | 0.835 |
0.015 | 0.760 |
0.053 | 0.109 | -0.011 | -0.015 |
13 | 0.785 |
0.036 | -0.010 | 0.146 | 0.109 | 0.674 |
-0.065 |
14 | 0.855 |
0.806 |
0.030 | -0.017 | 0.016 | 0.021 | 0.093 |
15 | 0.711 |
0.095 | 0.047 | -0.032 | 0.658 |
-0.014 | -0.007 |
16 | 0.875 |
0.003 | -0.025 | 0.059 | -0.016 | 0.068 | 0.862 |
17 | 0.840 |
-0.010 | 0.004 | 0.799 |
-0.044 | 0.090 | 0.088 |
18 | 0.856 |
0.002 | 0.814 |
-0.015 | -0.013 | 0.050 | 0.077 |
19 | 0.857 |
-0.090 | 0.026 | 0.075 | -0.010 | 0.900 |
0.003 |
20 | 0.864 |
0.766 |
0.001 | 0.124 | -0.007 | -0.003 | 0.032 |
21 | 0.902 |
-0.015 | 0.046 | -0.007 | 0.808 |
0.039 | 0.112 |
22 | 0.860 |
-0.069 | -0.030 | 0.055 | 0.046 | 0.049 | 0.872 |
23 | 0.827 |
0.198 | -0.010 | 0.667 |
0.016 | 0.057 | -0.030 |
24 | 0.900 |
-0.069 | 0.844 |
0.041 | 0.083 | 0.056 | -0.016 |
25 | 0.726 |
0.040 | 0.016 | -0.072 | 0.039 | 0.659 |
0.116 |
26 | 0.744 |
0.632 |
0.056 | 0.007 | 0.073 | -0.017 | 0.146 |
27 | 0.751 |
0.074 | 0.329 |
-0.060 | 0.455 |
-0.047 | 0.107 |
28 | 0.827 |
0.051 | 0.093 | 0.038 | 0.028 | -0.049 | 0.742 |
29 | 0.813 |
0.001 | 0.060 | 0.764 |
-0.009 | -0.016 | 0.132 |
30 | 0.893 |
-0.035 | 0.829 |
0.030 | 0.067 | 0.001 | 0.034 |
loadings are above 0.30;
each item loaded on its corresponding factor, while all cross-loadings were constrained to be zero.
Factor I, enhancement motive; Factor II, conformity motive; Factor III, pastime motive; Factor IV, social motive; Factor V, information motive; Factor VI, coping motive; CFA, confirmatory factor analysis; ESEM, exploratory structural equation modeling
Correlations, descriptive statistics, and reliability
1 | 2 | 3 | 4 | 5 | 6 | SA ESEM | SA CFA | |
---|---|---|---|---|---|---|---|---|
1-enhancement | - | 0.25 | 0.65 | 0.44 | 0.54 | 0.45 | 0.49 | 0.67 |
2-conformity | 0.32 | - | 0.11 | 0.69 | 0.26 | 0.53 | 0.48 | 0.58 |
3-pastime | 0.76 | 0.22 | - | 0.24 | 0.39 | 0.31 | 0.29 | 0.53 |
4-social | 0.51 | 0.76 | 0.34 | - | 0.42 | 0.37 | 0.41 | 0.60 |
5-information | 0.64 | 0.35 | 0.50 | 0.51 | - | 0.34 | 0.25 | 0.45 |
6-coping | 0.59 | 0.59 | 0.47 | 0.52 | 0.46 | - | 0.57 | 0.78 |
Mean | 16.58 | 9.07 | 17.43 | 12.06 | 14.89 | 11.16 | ||
SD | 4.10 | 4.27 | 4.66 | 4.66 | 4.52 | 4.82 | ||
Cronbach’s alpha | 0.895 | 0.931 | 0.908 | 0.877 | 0.868 | 0.914 |
Correlations above the diagonal are obtained from a ESEM solution. Correlations below the diagonal are obtained from an CFA solution. All correlations are significant. Correlations between use motives and SNS addiction are obtained from a seven-factor ESEM or CFA solution.
SA, SNS addiction; enhancement, enhancement motive; conformity, conformity motive; pastime, pastime motive; social, social motive; information, information motive; coping, coping motive. SA, SNS addiction; ESEM, exploratory structural equation modeling; CFA, confirmatory factor analysis; SD, standard deviation