Self Esteem and Social Media usage between people of different gender and ages.
# Results
## Comparison of Reading ,Self Esteem and Social Media usage between people of different gender and ages.
A regression analysis was created to evaluate if the Reading ,Self Esteem and Social Media varied across the gender and age between the people of different gender and ages. All the investigated variables such as depression, anxiety and stress, depicting the mental health of Facebook and non-Facebook users were statistically tested to conduct the one-way ANOVA. The variables did not contain any extreme outliers or multicollinearity (correlation was less than 0.9). The homogeneity of variances was checked using Leven’s test and come out be non-significant and hence was acceptable to conduct one-way ANOVA. However, Shaprio-Wilk test comes out to be significant for all variables and was not acceptable. (see Table 1).
“`{r, out.width = “90%”, out.height = “100%”, fig.align = “center”, fig.pos=’h’, out.extra = ”, fig.cap = “\\label{fig:fig2}”, echo=FALSE}
knitr::include_graphics(“C:/Users/Faheem/Desktop/Ishani/MANCOVA/images/Table 1.png”)
“`
To remedy this, the Kruskal-Wallis rank test was adopted, which is deemed fit to deal with non-parametric and as an alternative to the one-way ANOVA test. Table 2. Table 3 and Table 4 reports the comparison of the results of the Kruskal-Wallis rank test of depression, anxiety and stress respectively between the Facebook users and non-Facebook users.
Depression level did not significantly vary across male and female among Facebook users. However, it varied significantly by the age groups: Up to 30 and 30 plus. A Kruskal-Wallis rank test with post hoc Dunn’s comparison revealed that the Facebook users having middle-age (30 plus) possessed a higher level of depression than the corresponding youngest (Up to 30) respondents. Whereas, the depression level did not vary significantly among the non-Facebook users (see Table 2).
“`{r, out.width = “90%”, out.height = “100%”, fig.align = “center”, fig.pos=’h’, out.extra = ”, fig.cap = “\\label{fig:fig2}”, echo=FALSE}
knitr::include_graphics(“C:/Users/Faheem/Desktop/Ishani/MANCOVA/images/Table 2.png”)
“`
Anxiety level among Facebook users significantly varied by both gender and age. A Kruskal-Wallis rank test with post hoc Dunn’s comparison indicated that male were more exposed to anxiety levels than the female. Similarly, a Kruskal-Wallis rank test with post hoc Dunn’s comparison suggested that the Facebook users having middle-age (30 plus) possessed the higher level of anxiety than the corresponding youngest (Up to 30) respondents. Anxiety level did not significantly vary by the gender and age among the non-Facebook users (see Table 3).
“`{r, out.width = “90%”, out.height = “100%”, fig.align = “center”, fig.pos=’h’, out.extra = ”, fig.cap = “\\label{fig:fig2}”, echo=FALSE}
knitr::include_graphics(“C:/Users/Faheem/Desktop/Ishani/MANCOVA/images/Table 3.png”)
“`
Stress level among Facebook users significantly varied by both gender and age. A Kruskal-Wallis rank test with post hoc Dunn’s comparison indicated that male were possessing a higher level of stress than the corresponding female respondents. Also, a Kruskal-Wallis rank test with post hoc Dunn’s comparison highlighted that the Facebook users having middle-age (30 plus) possessed the higher level of stress than the corresponding youngest (Up to 30) respondents. Anxiety level did not significantly vary by the gender and age among the non-Facebook users (see Table 4).
“`{r, out.width = “90%”, out.height = “100%”, fig.align = “center”, fig.pos=’h’, out.extra = ”, fig.cap = “\\label{fig:fig2}”, echo=FALSE}
knitr::include_graphics(“C:/Users/Faheem/Desktop/Ishani/MANCOVA/images/Table 4.png”)
“`
## Hierarchal Regression Analysis
# Depression
The first block-wise regression analysis of depression indicated that sociodemographic variable such as gender was not a significant predictor of the depression in both Facebook and non-Facebook users. While other sociodemographic variables such as age (reference) for both Facebook and non-Facebook users and specifically, age (Up to 30) found to be a significant predictor of the depression in non-Facebook users. The second block-wise regression analysis of depression indicated that explanatory variable such as social media use was a significant predictor and internet use was not significant predicator the depression in both Facebook and non-Facebook users. The depression was better predicted with the sociodemographic variables F(1, 788) = 2.9968, p < 0.08382. , accounting additional variability of 0.06% than the corresponding social media use in Facebook users. Conversely in non-Facebook users, the depression was better predicted by the social media use F(1, 153) = 8.7117, p < 0.00366 **, explaining additional 2.23% od variability than the corresponding sociodemographic variables. Furthermore, an analysis of the interaction terms indicated that age (30 plus) moderated the relationship between social media use and depression in Facebook users. On the other hand, none of the interaction terms predicted the depression level in non-Facebook users (see Table 5).
# Anxiety
The first block-wise regression analysis of anxiety indicated that both sociodemographic variable such as gender and age: age(reference), age (Up to 30) and age (30 plus) were the significant predictors of the anxiety in Facebook users. The gender was not a significant predictor and age (Up to 30) found to be a significant predictor of the anxiety in non-Facebook users. The second block-wise regression analysis of anxiety indicated that explanatory variable such as social media use was not a significant predictor for anxiety in Facebook users, whereas it was significant for anxiety in non-Facebook users. On the contrary, internet use was a significant predictor of anxiety in Facebook users, whereas it was not a significant predictor of anxiety in non-Facebook users. In the non-Facebook users, anxiety was better predicted by the age (Up to 30) F(1, 153) = 4.2383, p < 0.04123*, accounting additional variability of 3.85% than the corresponding social media use. Furthermore, an analysis of the interaction terms indicated that gender moderated the relationship between social media use and anxiety in Facebook users. On the other hand, none of the interaction terms predicted the anxiety level in non-Facebook users (see Table 5).
# Stress
The first block-wise regression analysis of stress indicated that both sociodemographic variables such as gender and age(reference) were the significant predictors of the stress in Facebook users. The gender was not a significant predictor and age: age(reference) and age (Up to 30) found to be a significant predictor of the stress in non-Facebook users. The second block-wise regression analysis of depression indicated that explanatory variable such as social media use was a significant predictor and internet use was not significant predicator the depression in both Facebook and non-Facebook users. The stress was better predicted with the sociodemographic variables F(1, 787) = 18.7934, p < 1.646e-05***, accounting additional variability of 2.68% than the corresponding social media use in Facebook users. Conversely, in the non-Facebook users, the stress was better predicted by the social media use F (1, 153) = 6.9325, p < 0.009333,explaining additional 2.76% variability than the corresponding sociodemographic variables. None of the interaction terms predicted the stress level in both Facebook and non-Facebook users (see Table 5).
“`{r, out.width = “90%”, out.height = “100%”, fig.align = “center”, fig.pos=’h’, out.extra = ”, fig.cap = “\\label{fig:fig2}”, echo=FALSE}
knitr::include_graphics(“C:/Users/Faheem/Desktop/Ishani/MANCOVA/images/Table 5.png”)
“`
# Discussion