منبع کنترل و استفاده از تلفن همراه: پیامدها برای کیفیت خواب، عملکرد تحصیلی و بهزیستن ذهنی

Abstract This study centers on the following hypothesis: that individuals with an external locus of control, in comparison to individuals with an internal locus of control, have less control over their cell phone use (i.e., more likely to use at bedtime; more likely to use in class and while studying) and are consequently more vulnerable to the negative outcomes associated with excessive cell phone use (i.e., poor sleep quality, reduced academic performance, and reduced subjective well-being). Methods: Undergraduate college students (N = 516) participated in the study by completing validated surveys assessing their cell phone use, locus of control, sleep quality, academic performance, and subjective well-being. A path model was used to examine how locus of control relates to students’ cell phone use and the key outcome variables. Results: The model exhibited reasonable model fit with all paths being statistically significant and in the hypothesized direction. Conclusion: By enabling an individual to better control cell phone use at inopportune times, a greater internal locus of control may mitigate some of the negative outcomes associated with high frequency cell phone use; conversely, an individual with a greater external locus of control may have difficulty controlling use at inopportune times and the negative effects associated with high frequency use may be exacerbated.

1. Introduction The cellular telephone (i.e., mobile phone, smartphone and hereafter cell phone) is central to young people’s lives. The device is a primary hub for two-way communication (calling, texting, email, etc.), online social networking, leisure time entertainment, information gathering, and data management pertinent to daily life. For this reason, many young people use the cell phone throughout the day and increasingly describe it as something they cannot live without (Pew Research, 2014). Yet, research has emerged identifying multiple negative outcomes associated with excessive daily cell phone use. Negative outcomes include reduced academic performance ( Dietz and Henrich, 2014, Jacobsen and Forste, 2011, Junco and Cotton, 2012, Lepp et al., 2014, Lepp et al., 2015, Rosen et al., 2013, Wei et al., 2012 and Wood et al., 2012), poor sleep quality ( Fossum et al., 2014, Lanaj et al., 2014, Lemola et al., 2015, Munezawa et al., 2011, Murdock, 2013 and Thomée et al., 2011), decreased mental health ( Beranuy et al., 2009, Harwood et al., 2014, Jenaro et al., 2007, Lepp et al., 2014 and Rosen et al., 2014), increased sedentary behavior, decreased cardiorespiratory fitness and decreased intensity of planned exercise ( Barkley et al., 2015, Lepp et al., 2013 and Rebold et al., 2015), and decreased life satisfaction ( Lepp et al., 2014). As these relationships become more clear, it is important to explore the psychology behind them. Of particular interest to this study is an individual’s locus of control ( Johnson et al., 2015 and Rotter, 1966). Briefly, locus of control describes an individual’s beliefs about their ability to control the environment as well as the outcomes of their behavior. Individuals with a greater internal locus of control tend to believe that control is centered within themselves; conversely, individuals with a greater external locus of control tend to believe that control is centered outside of themselves and are therefore more likely to attribute behavioral outcomes to environmental influences. Given that the cell phone is a prominent feature of the modern environment, it is worth examining how locus of control may be associated with cell phone use and associated outcomes. Considering the current research, we hypothesize that individuals with an external locus of control, in comparison to individuals with an internal locus of control, have less control over their cell phone use and are consequently more vulnerable to the aforementioned negative outcomes associated with excessive cell phone use. To put this in less formal terms, we are suggesting that individuals with an external locus of control have surrendered some control over their environment and behavioral outcomes to their cell phone. Testing this idea enabled us to develop a model describing the relationship between young people’s cell phone use and important behavioral outcomes (academic performance, sleep quality, and subjective well-being) and the role locus of control plays in the relationship.

4. Results 4.1. Descriptive statistics, examination of outliers, and assumption checking The age range of the data set (n = 516) was 18–29 with a mean of 20 (SD = 1.48). Females comprised 80% of the data set (n = 412), which is higher than the percentage of females (59%) in the overall undergraduate student body of the university. As reported below, the potential influence of gender on the model was thoroughly examined. Next, an examination of outliers for the open ended data was conducted (i.e., total daily cell phone use, average cell phone use during class time and study time). Following the method of Rosen et al. (2013), values that were more than three standard deviations from the mean were truncated to exactly three standard deviations from the mean. This procedure was applied to the measure of total daily cell phone use for four participants. Average cell phone use for the sample was 347 min·day−1 (SD = 249.13). This procedure was also applied to the measure assessing the frequency with which students check their cell phone during class time and study time for six participants. On average, students checked their cell phone 12 times·hour−1 (SD = 10.91) during class and study time. Descriptive statistics for all variables in the model are presented in Table 1. Table 1. Descriptive statistics for the major variables in the path model (N = 492). M SD Min Max Total daily cell phone usea 346.75 249.13 0.00 1130.00 Cell phone use at night scale 25.28 6.67 8.00 40.00 Cell phone use during class & study timeb 12.31 10.91 0.50 53.43 Locus of Control scale (LOC) 0.00 0.66 −2.22 1.21 College Grade Point Average (GPA)c 3.31 0.54 0.00 4.00 Pittsburg Sleep Quality Index (PSQI) 5.91 3.54 0.00 19.00 Satisfaction With Life scale (SWL) 24.69 7.27 5.00 35.00 a Min·day−1. b Times checked·hour−1. C Calculated on a 4.0 scale. Table options Before conducting the analyses, we also examined the completeness of the data matrix and the normality of the variables used. For the current data, there were 24 students who had at least one missing value on the variables used in the model. Due to the relatively small proportion of missing data (i.e., 4.7%), we allowed the default missing data treatment method in the computer package, listwise deletion, to drop these cases from the analysis (Rubin, 1987). In LISREL 9.1, the default estimation method is maximum likelihood (ML) that assumes a multivariate normal data distribution. Therefore, we performed the multivariate normality test for the variables in the path model. The test results indicated that the assumption of a multivariate normal data distribution may not hold in the sample data set (χ2 = 861.05, p < .001). As a solution, we adopted an alternative estimation method, Weighted Least Squares (WLS), in conjunction with the asymptotic covariance matrix to correct the bias resulted from the non-normal data ( Jöreskog and Sörbom, 1999 and Jöreskog et al., 2001). Lastly, the effect of gender on the model’s dependent variables was assessed using multiple regression analyses. The gender effect was not significant for cell phone use during class time and study time (t = −0.68, p = .50) and PSQI (t = 0.48, p = .63); marginally significant for SWL (t = 2.49, p = .013); and significant for cell phone use at night (t = 4.23, p < .001) and college GPA (t = 3.50, p < .001). In a more conservative approach, gender was added as a covariate for not only cell phone use at night and college GPA, but also SWL in the path analysis. 4.2. Path analysis The zero-order Pearson correlation matrix for the variables in the path model is displayed in Table 2. The correlation analysis identified 16 statistically significant relationships in the matrix with a p-value ⩽ .05. Total cell phone use was positively related to both cell phone use at night (r = .33, p < .01) and cell phone use during class and study time (r = .33, p < .01) and cell phone use at night was also positively related to cell phone use during class and study time (r = .24, p < .01). Locus of control was significantly correlated with all the dependent variables in the model where higher scores indicate a more internal locus of control. As such, locus of control was positively related to college GPA (r = .20, p < .01) and life satisfaction (r = .44, p < .01), and negatively related to cell phone use at night (r = −.15, p < .01), cell phone use during class and study time (r = −.14, p < .01), and sleep quality as measured by PSQI where higher scores indicate poorer sleep quality (r = −.35, p < .01). Notably, locus of control was not related to total daily cell phone use (r = −.02, p = .73). Lastly, the inter-correlations among college GPA, PSQI, and life satisfaction were also statistically significant with college GPA negatively related to PSQI (r = −.11, p < .05) but positively related to life satisfaction (r = .24, p < .01) and PSQI negatively related to life satisfaction (r = −.35, p < .01). Thus, a higher GPA was associated with better sleep quality and increased satisfaction with life; likewise, better sleep quality was associated with increased satisfaction with life. Table 2. Pearson correlations between variables in the path model (N = 492). Measure 1 2 3 4 5 6 1. Total daily cell phone usea – 2. Cell phone use at night .332⁎⁎ – 3. Cell phone use during class & study time .327⁎⁎ .243⁎⁎ – 4. Locus of Control (LOC) −.016 −.147⁎⁎ −.138⁎⁎ – 5. College Grade Point Average (GPA) −.173⁎⁎ −.021 −.155⁎⁎ .204⁎⁎ – 6. Pittsburg Sleep Quality Index (PSQI) .082 .299⁎⁎ .122⁎⁎ −.351⁎⁎ −.111⁎ – 7. Satisfaction With Life (SWL) −.025 −.005 −.114⁎ .440⁎⁎ .235⁎⁎ −.347⁎⁎ a min·day−1. ⁎⁎ p ⩽ .01 (2-tailed). ⁎ p ⩽ .05 (2-tailed). Table options Given these significant correlations, we proceeded to test the hypothesized path model which also included gender as a covariate for cell phone use at night, college GPA, and satisfaction with life (Fig. 2). Judging by the five goodness-of-fit indices, the model fits the data well (i.e., χ2 = 44.16, df = 10, p < .001; χ2/df = 4.4; RMSEA = .058; GFI = .987; SRMR = .033; CFI = .942). Specifically, all the model goodness-of-fit indices except the χ2 fell within the desired range for a good or reasonable model fit. Although the χ2 statistic was statistically significant, the χ2/df ratio had a value of 4.4 being slightly greater than 3, indicating the model fits the data reasonably well after the adjustment of the inflation caused by the large sample size ( Kline, 2004). RMSEA (.058) was only slightly greater than .05, showing that our model was very close to a good fit with reasonable error of approximation. The error terms of cell phone use at night and cell phone use during class and study time were significantly correlated (r(e) = .13, p < .01), suggesting that the cell phone use at inopportune times share some common behavior patterns beyond their relationship with total cell phone use and locus of control. Gender was a statistically significant covariate for cell phone use at night (β = .26, p < .001), college GPA (β = .21, p < .001), and satisfaction with life (β = .14, p < .01), indicating that there were gender differences in these variables. These results showed that female students on average tended to use more cell phone before going to bed, had better academic performance, and were happier with their life, when compared to their male counterparts. The effects of gender were controlled, which led to more accurate interpretation of the relationships between the independent and the dependent variables in the model. Path model with standardized path coefficients.Note1:∗∗p<.01; ∗∗∗p<.001; Gender ... Fig. 2. Path model with standardized path coefficients. Note1:∗∗p < .01; ∗∗∗p < .001; Gender was a covariate added for cell phone use at night, GPA, & SWL; error terms of cell phone use at night and cell phone use during class and study time were allowed to correlate. Note2: Total_CP = Total daily cell phone use; CP_Night = Cell phone use at night; CP_Class = Cell phone use during class/study time; LOC = Locus of Control; GPA = College Grade Point Average; PSQI = Pittsburg Sleep Quality Index (higher scores indicate poorer sleep quality); SWL = Satisfaction with Life scale. Figure options Fig. 2 presents the standardized coefficients for all the paths specified in the model. To be specific, the standardized weighted least squares estimates for the path coefficients and the equation error variance of the dependent variables are tabulated in Table 3. All paths were statistically significant and in the hypothesized direction. The standardized coefficients for the paths initiated from locus of control and total daily cell phone use to cell phone use at night were −.18 (p < .001) and .29 (p < .001), respectively. Thus, as locus of control became more internal, inopportune cell phone use at night decreased. Conversely, as total daily cell phone use increased, inopportune cell phone use at night increased. A total of 20% of the variance in cell phone use at night was explained by locus of control, total daily cell phone use, and the covariate of gender. The standardized coefficients for the paths from locus of control and cell phone use at night to sleep quality (PSQI) were −.31 (p < .001) and .25 (p < .001), respectively. Thus, as locus of control became more internal, sleep quality improved. Conversely, as inopportune cell phone use at night increased, sleep quality worsened. A total of 19% of variability in sleep quality was accounted for by locus of control and cell phone use at night. The standardized coefficients for the paths initiated from locus of control and total daily cell phone use to cell phone use during class and study time were −.13 (p < .01) and .32 (p < .001), respectively. Thus, as locus of control became more internal, inopportune cell phone use in academic settings decreased. Conversely, as total daily cell phone use increased, inopportune cell phone use in academic settings increased. A total of 12% of the variability in cell phone use during class and study time was explained by locus of control and total daily cell phone use. The standardized coefficients for the paths initiated from locus of control and cell phone use during class and study time to college GPA were .16 (p < .001) and −.13 (p < .01), respectively. Thus, as locus of control became more internal, GPA increased. Conversely, as inopportune cell phone use in academic settings increased, GPA decreased. A total of 10% of the variability in college GPA was accounted for by locus of control, cell phone use during class and study time, and the covariate of gender. Finally, the standardized coefficients for the three paths initiated from locus of control, sleep quality (PSQI), and college GPA to life satisfaction (SWL) were .32 (p < .001), −.23 (p < .001), and .11 (p < .01), respectively. Thus, as locus of control became more internal, sleep quality improved, and GPA increased then life satisfaction increased. A total of 27% of the variability in life satisfaction was explained by locus of control, sleep quality, college GPA, and the covariate of gender. Table 3. Weighted least squares estimates for hypothesized paths in the model (N = 492). Standardized estimate t Path coefficients Total_CP → CP_Night (H1) 0.29 6.38∗∗∗ Total_CP → CP_Class (H2) 0.32 6.01∗∗∗ LOC → CP_Night (H3) −0.18 −4.23∗∗∗ LOC → CP_Class (H4) −0.13 −2.85∗∗ CP_Night → PSQI (H5) 0.25 6.00∗∗∗ LOC → PSQI (H6) −0.31 −6.64∗∗∗ CP_Class → GPA (H7) −0.13 −2.73∗∗ LOC → GPA (H8) 0.16 3.49∗∗∗ PSQI → SWL (H9) −0.23 −5.33∗∗∗ GPA → SWL (H10) 0.11 2.90∗∗ LOC → SWL (H11) 0.32 7.52∗∗∗ Equation error variances CP_Night 0.80 18.15∗∗∗ CP_Class 0.88 10.63∗∗∗ PSQI 0.81 14.48∗∗∗ GPA 0.90 10.03∗∗∗ SWL 0.73 11.34∗∗∗ Note1:∗∗p < .01; ∗∗∗p < .001; gender was a covariate added for cell phone use at night, GPA, & SWL; error terms of cell phone use at night and cell phone use during class and study time were allowed to correlate. Note2: Total_CP = Total daily cell phone use; CP_Night = Cell phone use at night; CP_Class = Cell phone use during class/study time; LOC = Locus of Control; GPA = College Grade Point Average; PSQI = Pittsburg Sleep Quality Index (higher scores indicate poorer sleep quality); SWL = Satisfaction with Life scale.