Problematic internet use and safety behavior: The moderating role of safety climate

Iran were invited to participate in the study. The table from Krejcie and Morgan [36] was used to determine the required sample size. Given the total number of nine thousand employees in the company, the sample size was determined to be 368. However, the response rate is rarely 100% and we had to invite more participants to the study. Consequently and to guarantee the required sample size, the determined sample size was conservatively multiplied by 2, so 736 questionnaires were randomly distributed among employees. Data gathering tool Standard and validated questionnaires were used to gather the required data. A 6-item scale was developed by Siciliano et al. [37] to measure PIU. The scale was later modified by Abuba- kar and Al-Zyoud [5] to be usable in occupational settings. In this study, both tools were adopted to develop a Persian scale appropriate for the Iranian population. Content validity index (CVI) and content validity ratio (CVR) of all items were assessed before application [38, 39]. “I feel nervous and restless when I am offline and these feelings are disappeared when I go back online”, “I think that I use mobile phone and internet more than what I really need”, “sometimes I forget or neglect safety considerations because of being busy with mobile phone and internet”, and “I prefer to use mobile phone and internet instead of going out with my friends or exercising” are some items used to measure PIU. All items were responded using a 5-point Likert scale ranging from 1 (totally disagree) to 5 (totally agree).

A 7-item scale from Ghasemi et al. [28] was used to assess perceived safety climate (accept- able content validity and Cronbach’s alpha coefficient = 0.870). “In my workplace, safety is as important as production” and “Managers actively participate in designing and implementing safety programs” are some items used for measuring safety climate [28]. All items were responded using a 5-point Likert scale ranging from 1 (totally disagree) to 5 (totally agree). A 6-item scale was used for measuring safety behavior. The items were adopted from Neal and Griffin [3] which were translated and validated into Persian by Mahdinia et al. [40]. The scale encompasses both safety participation and safety compliance aspects of safety behavior and demonstrates acceptable reliability and validity (Cronbach’s alpha coefficient = 0.8 and acceptable content validity). Data analysis First, the construct validity of the variables was assessed. To do so, a measurement model was built and the goodness of fit of the model was assessed using absolute and comparative indices such as RMSEA, chi-square/df, NFI, RFI, GFI, CFI, and so on. RMSEA lower than 0.08, chi- square/df lower than 3, and comparative indices higher than 0.9 were considered to be accept- able [28, 41, 42]. Convergent and discriminant validity are two main aspects to be investigated during construct validity assessment. The convergent validity of variables was investigated based on standardized factor loadings (SFLs) and average variance extracted (AVE). SFLs higher than 0.3 and AVE higher than 0.5

were considered to be acceptable, as recommended by Bagozzi and Yi [43]. The discriminant validity of variables was evaluated by comparing the square root of AVE with correlation coef- ficients among variables. The discriminant validity is confirmed if the AVE square root of a variable is higher than its correlations with other variables. Hierarchical regression analysis in accordance with the procedure explained by Helm and Mark [44] was conducted to assess the moderating effect of safety climate on the relationship between PIU and safety behavior. This approach is one of the most popular methodologies uti- lized for the assessment of moderation effect [44]. In this respect, two regression models were constructed and examined. The first model was the basic model in which PIU and safety PLOS ONE Problematic internet use and safety behavior PLOS ONE | https://doi.org/10.1371/journal.pone.0279767 December 30, 2022 4 / 13