Correlates of video games playing among adolescents in an Islamic country – BMC Public Health

Sample

This cross-sectional study was performed with a random sample of eight schools selected from all 26 middle schools in the city of Hamadan, which is located in central Iran, with a population of over 500,000 people [21]. Almost 1950 sixth to eighth graders were registered in the eight randomly selected middle schools. One classroom from each school which includes 477 of adolescents (almost one out of four students) was selected to participate in this study. Determining the sample size for this study, we specified the difference between the largest mean and the smallest mean for GHQ-28 as the effect size (delta) and we hypothesized that all means for GHQ-28 other than the two extreme ones (non-gamers and excessive gamers) are equal to the grand mean. We calculated the sample size for three different effect size delta values (0.25, 0.75, 1.25) corresponding to “small”, “medium”, and “large” effects, according to Cohen & Cohen, Statistical Power Analysis for the Behavioral Sciences. The initial calculation indicates that using a medium effect size, we need at least 45 subjects in each group (180 participants) in order to detect a reasonable departure from the null hypothesis (α = 0.05 and 1 – β = 85%). However, our preliminary investigation showed that only 10% of students were non-players, therefore, if we would had selected 180 participants only 18 (180 × 10%) of them would be the non-players. Therefore, we increased the sample size by 250% (n = 180 × 250% = 450) to have 45 non-players in our sample.

This survey was self-administered in the absence of the instructor and collected by research associates of this project who were unaffiliated with the schools during regular classroom hours. To increase the validity of the responses, efforts were made to guarantee complete anonymity. Students were given a brief introduction and were asked not to write their name or any other identifiable information anywhere on the survey. The survey was conducted from February to March 2008. This study was conducted with approval from Hamadan University’s Institutional Review Board. Informed assent and consent were obtained from participants and their parents/guardians. The survey instrument was pilot tested with 20 students and modified accordingly. The questionnaire was administered in Farsi, the official language of Iranian schools. Regardless of ethnicity or background, all Iranian students speak and write Farsi.

Measures

Demographics and video-game playing

In addition to demographic characteristics, survey instruments included several items that were specifically designed to capture the amount of time students spend playing video games during weekdays and weekends. Commitments to video game use, the average and longest duration of play, and the type of video games were assessed.

Mental health status

The validated 28-item Farsi version of the General Health Questionnaire (GHQ-28) [22] was used to assess the mental health status of participants. The GHQ refers to subjective symptoms of psychological distress, somatic manifestations often associated with anxiety and depression, relationship difficulties, and social, family, and professional roles [23]. The GHQ-28 is composed of 4 subscales (score range, 1-7): somatization, anxiety, social dysfunction, and depression. Both subscales and summated total scores were used [23, 24]. All items have a 4 point scoring system using Likert scoring (0-1-2-3). A higher score on the GHQ-28 represents poorer mental health status. In our sample the Cronbach’s alpha coefficients of reliability of the subscales vary around 0.75 and the internal consistency of the total scale is 0.90. The participants were classified into higher and lower GHQ-28 groups based on a cutoff score of 8.

Aggression

The Orpinas’ aggression scale was used to measure aggressive behaviors of students [25]. The scale consists of 11 items designed to measure self-reported aggressive behaviors among school-aged students that might result in psychological or physical injury to other students. The scale requests information regarding the frequency of the most common overt aggressive behaviors, including verbal aggression (teasing, name-calling, encouraging students to fight, threatening to hurt or hit) and physical aggression (pushing, slapping, kicking, hitting), as well as information about anger (getting angry easily, being angry most of the day) [25]. The internal consistency of the scale showed coefficient alpha of 0.89.

Perceived side effects of video-computer games

A 5-item rating scale was used to gauge student beliefs on the side effects of playing excessive video games. The following are examples of the 5-item rating scale: “I believe that excessively playing video games has negative affects on educational performance;” and “I believe that obesity could be a side effect of excessively playing video games.” Each of these items was measured on an ordinal 5-point Likert-type scale (1 = certainly disagree, 5 = certainly agree). Higher scores on the scale indicated a greater perceived threat of video-computer games. The internal consistency reliability of this scale was examined using Cronbach’s coefficient alpha (α = 0.72).

Video violence exposure

Based on previous content analyses of popular video games among Iranian children, video games that were reported to be played by participants were categorized into two major groups: violent and non-violent games. Additionally, one of the co-authors of this study personally examined the content of each game to verify whether games are correctly categorized.

Statistical Analysis

Descriptive statistics were calculated to identify the distribution and frequency of all items which were subsequently used to construct the independent scales and indices for examination. In the bivariate analysis, chi-square tests were performed to measure the association between independent and outcome variables. A series of logistic regression and descriptive analyses was computed. Interrelationships between independent variables were examined to assess potential multicolinearity between independent variables. All statistical analyses were performed using the statistical software package SPSS (SPSS Inc., Chicago, Illinois, USA).