Abstract

1. LITERATURE


1.1. RESEARCH SUBJECT
Internet usage of the individuals is mostly related to social media (Hootsuite & We Are Social, 2020). People use the social media usage to get news, communicate, access to information, and for free-time activity and research purposes (Çömlekçi and Başol, 2019; Acun et al., 2017; Akıncı Vural and Bat, 2010; Taşçı and Ekiz, 2018). The reason for the spread of social media usage in daily life can be the diversity of social media usage purposes.
The concept of addiction for social media, which has an increasing number of users and hours of use, has become a popular topic in the literature. Social media addiction has similar findings to substance addiction. These findings include delaying jobs and not fulfilling responsibilities due to social media usage (Ayeni, 2019; Hootsuite & We Are Social, 2020). This may occur as academic procrastination among students. Therefore, this study focused on a measurement model that was developed with the literature to test the effects of the addiction and usage of social media which is thought to affect students’ academic procrastination behaviors.
1.2. RESEARCH PURPOSE AND IMPORTANCE
The study aimed to reveal the effect of social media addiction and social media usage purposes on academic procrastination behaviors. Academic procrastination behaviors have negative effects on academic achievement (Balkıs et al., 2006). Testing the presented measurement tool is important to enlighten some parts of the reasons for the students’ academic procrastination behaviors.
1.3. CONTRIBUTION of the ARTICLE to the LITERATURE
The study contributes to the literature with the enlightenment of the reasons for academic procrastination and the diversity of the application area of the structural equation model.
2. DESIGN AND METHOD


2.1. RESEARCH TYPE
This quantitative study examined the correlations between the variables and included the statistical analyses.
2.2. RESEARCH PROBLEMS
The study problem was to test the validity of the formed measurement model and reveal the correlations between the variables to enlighten the reasons for the students’ academic procrastination behaviors.
2.3. DATA COLLECTION METHOD
Data were obtained through an online questionnaire form including four sections. The first section of the form included statements to obtain demographic data. The second section included the Social Media Usage Motives which was developed by Solmaz et al. (2013) and had 11 items to reveal the social media usage purposes of the participants. The third section included the Social Media Addiction Scale-Adult Form which was developed by Şahin and Yağcı (2017) to determine the level of social media addiction of the participants. The last section of the form included the Academic Procrastination Scale which was developed by Aitken (1982) and was adopted to Turkish by Balkıs (2006) to measure the academic procrastination behaviors of the participants. All statements except the statements used in the first section were presented to the participant with a 5-point Likert structure.
The population of the study included Osmaniye Korkut Ata University associate degree, undergraduate and postgraduate students, and the sample included 399 students who were reached by simple random sampling.
2.4. QUANTITATIVE / QUALITATIVE ANALYSIS
The Structural Equation Modeling (SEM) is a statistical method which enables the testing and verification of the effects between the observed and potential variables and which has a wide range of application areas. In other words, it examines the compatibility of a measurement model, which is considered to be conceptually valid with the obtained data and helps to verify or reject them (Şimşek Kandemir, 2019; Gürbüz, 2019).
SmartPLS 3.2.9 software was used for the application of variance-based structural equation model that works with the partial least squares method to test the presented model in the study. The validity of the study model and testing of the hypotheses were realized with this software (Doğan, 2019).
2.5. RESEARCH MODEL
This study examined the effect of social media addiction on social media usage purposes and academic procrastination behavior and the effects of social media usage purposes on academic procrastination behavior in the measurement model formed with the support of the literature. Additionally, the study checked the mediator effect of social media usage purposes on the relationship between social media addiction and academic procrastination behavior.
2.6. RESEARCH HYPOTHESES
The four determined hypotheses of the study are as follows;
H1: Social media addiction has a positive significant effect on social media usage purposes.
H2: Social media addiction has a positive significant effect on academic procrastination behavior.
H3: Social media usage purposes have a positive significant effect on academic procrastination.
H4: Social media usage purposes have a mediator effect between social media addiction and academic procrastination.
3. FINDINGS AND DISCUSSION


3.1. FINDINGS as a RESULT of ANALYSIS
The validity and reliability analyses of the scales were performed at first. Thus, the combination validity, internal consistency and disintegration validity were examined. Cronbach's alpha coefficient and Composite Reliability (CR) coefficient were taken into consideration for the internal consistency reliability. The factor loads and average variance extracted (AVE) were regarded for combination validity (Hair et al., 2017: 109). Three criteria were examined for disintegration validity. These criteria were cross loadings, Fornell-Larckerr table and HTMT values. The structures in the study received values at reference points of all criteria, and the internal consistency reliability, combination and disintegration validities of the structures were reached (Hair et al., 2017: 113; Doğan, 2019: 46).
PLS algorithm was used to determine the VIF values, effect size (f2) and R2 values of the study model while Blindfolding analysis was used to determine its predictive power (Q2). Considering R2 values in the analysis results, academic procrastination behavior and social media usage purposes were respectively explained at the rate of 37.5% and 27.7%, and there was no linearity problem according to VIF values. Additionally, the model had predictive power on the variables of academic procrastination and social media usage purposes according to Q2 values. Considering the effect size (f2) values, social media addiction had high effects on social media usage purposes while social media purposes had high effects on academic procrastination behavior; however, social media addiction had no effect on academic procrastination. Mediator effect analysis was made based on this finding. In the analysis, the effects were checked using 5000 sub-samples in bootstrap technique and VAF coefficient was calculated (Zhao et al., 2010; Hair et al., 2014). VAF values for the related correlation was calculated as 1.04 (indirect effect=0.326 and direct effect=0.013) and social media usage purposes were found to have complete mediatory effect on this correlation.
3.2. HYPOTHESIS TEST RESULTS
The study supported H1 hypothesis which states that social media addiction has a positive significant effect on social media usage purposes (β= 0.527, p<0.000) and H3 hypothesis (β= 0.619, p<0.000) which states that social media usage purposes have a positive significant effect on academic procrastination. However, the study did not support H2 hypothesis (β=-0.013, p=0.854) which states that social media addiction has a positive significant effect on academic procrastination. The result of mediator effect analysis (VAFSMA-SMUP-AP=1.04) supported H4 hypothesis and showed that social media usage purposes had complete mediatory role on the relationship between social media addiction and academic procrastination.
3.3. DISCUSSING the FINDINGS with the LITERATURE
In line with the studies in the literature, the study results showed that social media usage purposes had a direct positive effect on social media addiction (Çömlekçi and Başol, 2019; Idubor, 2015; Filiz et al., 2014). The variable of social media usage purposes had a direct positive effect on the variable of academic procrastination behavior. Accordingly, the study stated that the quality of the students’ social media usages was also important. Unlike the literature, the study revealed that social media addiction had no direct positive effect on academic procrastination behavior (Gürültü, 2016; Gür et al., 2018; Teyfur, 2017; Demir and Kutlu, 2017). Therefore, the study performed an analysis according to the mediator analysis conditions determined by Zhao et al. (2010) and determined the presence of a mediator effect.
4. CONCLUSION, RECOMMENDATION AND LIMITATIONS


4.1. RESULTS of the ARTICLE
The analysis findings showed that the proposed measurement model in the study can be used. The study data showed that students’ social media addiction is mainly due to reaching information and staying up-to-date and that their social media usage purposes tend to be accessing to information and staying up-to-date as well as spending free leisure time and communicating. The study revealed that social media addiction affected social media usage purposes while it positively affected academic procrastination behavior. Accordingly, the study deducted that complete mediatory effect of the variable of social media usage purposes was comprehensible.
4.2. SUGGESTIONS BASED on RESULTS
Future studies can perform supportive analyses that explain the mediatory effect in the model. Additionally, different variables can be added to the model and mediatory effect in the variables can be examined.
4.3. LIMITATIONS of the ARTICLE
The sample form which the data were obtained has the ability to represent the universe using the simple random sampling method to generalize to cover all students.