Introduction
The issue of smartphone addiction is one that most have heard of by now. As a generation that has grown up with the use of these gadgets are now a staple in most people’s day-to-day routine. A study by Deng et al. (2018) shows that on average, “participants spent 2 hours 39 minutes on their smartphone and made 101 app switches per day.”
Despite the widespread propagation of the problems this addiction may cause, these issues have mostly been brushed aside or overlooked. With addiction comes an influence on other parts of our lives, most often being negative or destructive in nature. The topic of interest comes in relating this issue with possible correlated factors. Delving more extensively in this matter allows for the addition of other factors such as the purpose of this usage as well as the demography of the population of interest. Social networking is revealed to be the category of most frequent use at an average of 44 minutes per day (Deng et al., 2018).
Relating to this issue is the prevalent complication of stress. While some applications such as games may help in reducing the amount of stress in certain demographics, the use of such categories of apps may correlate to a higher level of perceived stress in the population. While other factors such as age may be more deeply correlated to these levels of stress, the study on the strength of this correlation may provide further insight into the effects of smartphone addiction.
In this study, the Analysis of Variance is used to test the hypothesis that there is no significant difference between the various age groups and their corresponding smartphone application-based addiction scale and perceived stress levels.
The general intent of the paper is to determine the existence of a significant difference among the studied age groups in terms of two variables — 1) perceived stress levels and 2) smartphone addiction, individually. The acquired data for the study was extracted from a larger dataset and will primarily focus on samples recorded from participants aged 18 to 45 years old.
Methodology
The Analysis of Variance, an analysis tool, was employed to identify if there is a statistically significant difference between the means of the independent groups identified at a significance level of:
Description of the Data
The dataset utilized in this paper was acquired from an online journal article from Data in Brief titled “The connection between risk of smartphone addiction, type of smartphone use, life satisfaction, and perceived stress dataset.” In their paper, Vujić and Szabo focused on analyzing smartphone addiction, specifically through the four variables (Smartphone addiction, Satisfaction with life, Perceived Stress, and Hedonic smartphone use) between males and females.
A copy of the raw dataset can be accessed online through Mendeley Data. The following tables provide a brief overview of the extracted dataset from the study.
Table 1. Overview of Data for Smartphone application-based addiction scale (SABAS)
Table 2. Overview of Data for Perceived Stress Levels (PSS)
On the Collection of Data
As declared by Vujić and Szabo, the responses were collected anonymously through Qualtrics, an online research platform. Participation in the data collection was voluntary. Considering the respondents’ voluntary participation, the adopted instruments (SABAS and PSS) were specifically chosen for their validity, simplicity, and brevity.
ANOVA
Note that the same process was implemented in the ANOVA testing for the SABAS and PSS.
Parameter of Interest
The parameters of interest are the smartphone application-based addiction scale (SABAS) and Perceived Stress Levels (PSS).
Null Hypothesis
\[H_{0}:\tau_{1}=\tau_{2}=\tau_{3}=\tau_{4}= 0 \]
Alternative Hypothesis
\[H_{1}:\tau_{i} \neq 0 \text{ for any }i\]
Rejection Criteria
The null hypothesis must be rejected if the calculated F ratio is greater than the value \[f_{\alpha,a-1,a(n-1)}\] — the critical F value or in this case, \[f_{0.05,3,357 }=2.62991651\] Otherwise, there is no sufficient evidence to reject the null hypothesis, hence one fails to reject it.
Test Statistic
\[ F_{0}=\frac{SS_{treatments}/(a-1))}{SS_{E}/[a(n-1))]]}=\frac{MS_{treatments}}{MS_{E}} \]
Results and Discussion
The figures in this section show the residual plots against perceived stress and smartphone addiction for the different age groups. From the plots, one can observe that the spread of data points are almost similar among age groups.
While the ANOVA concludes that there is a difference in perceived stress and smartphone addiction among the age groups, the same cannot be easily concluded by mere inspection of the plots. Hence, this analysis suggests future analyses or researches to conduct a post-hoc analysis to ascertain which age group(s) truly stands out from the rest.
ANOVA Tables
The F values for both tests were calculated using the equations listed in the methodology portion. All of the values necessary to carry out these tests were shown in the ANOVA table from the spreadsheet. In both tests, the F value was shown to be greater than the F-critical value, which is a value based on the number of treatments and observations in the dataset. Given these calculations, the null hypothesis which states that there is no significant change in the treatment effect between observed groups is rejected in favor of the alternative hypothesis.
The implication of these test results is that the different treatment groups - or in our case, age groups - were shown to have a significant and observable difference in their treatment effects. Along with this, it is implied, and was backed by the data, that the mean value of each treatment group differs significantly from the mean of the entire sample (“grand mean”). Hence, the amount of time different age groups spends on their phones varies significantly.
Although it was shown that there is a relationship between the age of a participant and their likelihood of experiencing stress and that of exhibiting a risk of smartphone addiction, it is still not certain if there are more variables at play, or if age itself is even the strongest factor in a person’s likelihood of having these experiences.
The relationship between smartphone addiction and stress was not tested in this study. Further analyses must be conducted to see if one has an effect on the other. However, since these variables were tested separately from each other in this study, the results were not affected by any potential correlation between these two.
Conclusion
Given the calculations that result in: \[F > F_{critical} \] the null hypothesis which states that there is no significant change in the treatment effect between observed groups is rejected in favor of the alternative hypothesis. This infers that significant change in the treatment effect among the groups has been observed. Certain limitations of the scope of this paper may include age being a major influence in the purpose for which certain populations use smartphones.
In a world where all the answers to our questions may be found with a handheld device, certain measures should be implemented to safeguard the well-being of smartphone users. Given the findings of this study, it can be recommended to collect more data on the degree of smartphone addiction in particularly vulnerable age groups (such as ages 18 to 23) and construct protective measures to counter the risk of smartphone addiction.
References
Yeh Y., Chang C., Ting Y. & Chen S. (October 2020). Effects of Mindful Learning Using a Smartphone Lens in Everyday Life and Beliefs toward Mobile based Learning on Creativity Enhancement. Educational Technology & Society. 23 (4), 45-58. https://www.jstor.org/stable/26981743
