CHAPTER THREE - METHODS

CHAPTER THREE: METHODS

CHAPTER THREE: METHODS

Overview

Postsecondary students are using technology for academic work on a regular basis, resulting in more access to, and use of, Internet technologies. This increased use and access, also exposes students to unsafe interactions that can put their well-being at risk (Musharraf & Anisul-Haque, 2018). Although studies indicate a decrease in traditional bullying, incidents of cyberbullying continue to rise, resulting in both emotional and/or physical damage to the student victim (Alqahtani et al., 2018; El Asam & Samara, 2016). Furthermore, there are limited studies on cyberbullying prevalence among college students attending faith-based universities; therefore, the purpose of this quantitative, causal-comparative study was to gather self-reported data on the prevalence of cyberbullying experiences, as a victim and as an offender, among traditional undergraduate students attending faith-based universities. The study sought to determine if the individual’s biological gender or level of religiosity had an effect on the prevalence of cyberbullying experiences. This chapter addresses all components of methodology involved for the study.

Design

This study utilized a quantitative, causal-comparative research design. A quantitative, causal-comparative design was appropriate because this study was a non-experimental investigation that sought to determine cause-and-effect relationships in groups where an independent variable was, or was not, present (or present at different levels), and to determine if groups differed on the dependent variable (Gall et al., 2007). The dependent variables for this study are the prevalence of cyberbullying victimization experiences and the prevalence of cyberbullying offending experiences. For this study, prevalence of cyberbullying experiences scores measured how many times someone was cyberbullying victim and how many times someone was a cyberbullying offender (Hinduja & Patchin, 2015). Cyberbullying victimization is when someone is “repeatedly harassed, mistreated, or made fun of (on purpose to hurt them) online or while using cell phones or any electronic devices” (Hinduja & Patchin, 2015, p.11). Cyberbullying (offending) is when someone “repeatedly mistreats, harasses, or makes fun of another person (with the intent to hurt them) online or while using cell phones or any electronic devices” (Hinduja & Patchin, 2015, p. 11).

The independent variables in this study are biological gender (female/male) and level of religiosity (higher/lower). For this study, the definition of biological gender is female and male, determined by biological sex, as created by God (Gen. 1:27). Level of religiosity is generally defined as the frequency of church attendance, time spent in private religious activities, experiences with the Divine (God), and how one’s beliefs influence one’s behavior (Koenig et al., 1997). A higher level of religiosity indicates the individual reported a high frequency of church attendance, strong belief in God, and significant time spent in private religious activities. A lower level of religiosity indicates the individual reported little to no church attendance, little to no belief in God, and little to no time spent in private religious activities (Storch et al., 2004). This study examined the differences between the independent variables as they related to the prevalence of cyberbullying victimization experiences and cyberbullying offending experiences.

Research Questions

The research questions for this study were as follows: RQ1: Is there a difference in the prevalence of cyberbullying victimization experiences, based on biological gender and level of religiosity, among traditional undergraduate students attending faith-based universities?

RQ2: Is there a difference in the prevalence of cyberbullying offending experiences, based on biological gender and level of religiosity, among traditional undergraduate students attending faith-based universities?

Hypotheses

The null hypotheses for this study are:

Ho1: There is no difference in the prevalence of cyberbullying victimization experiences scores between female and male traditional undergraduate students attending faith-based universities.

Ho2: There is no difference in the prevalence of cyberbullying victimization experiences scores, based on level of religiosity (higher/lower), of traditional undergraduate students attending faith-based universities.

Ho3: There is no difference in the prevalence of cyberbullying victimization experiences scores, based on biological gender (female/male) and level of religiosity (higher/lower), of traditional undergraduate students attending faith-based universities.

Ho4: There is no difference in the prevalence of cyberbullying offending experiences scores between female and male traditional undergraduate students attending faith-based universities.

Ho5: There is no difference in the prevalence of cyberbullying offending experiences scores, based on level of religiosity (higher/lower), of traditional undergraduate students attending faith-based universities.

Ho6: There is no difference in the prevalence of cyberbullying offending experiences scores, based on biological gender (female/male) and level of religiosity (higher/lower), of traditional undergraduate students attending faith-based universities.

Participants and Setting

The participants for this study included 284 participants (N = 284) drawn from a convenience sample of traditional undergraduate students attending two faith-based universities. The postsecondary schools are private, Christian universities where traditional undergraduate students participate in chapel services, or convocation, as part of their educational studies. Both universities also have an implicit Christian mission clearly stated in their mission statement. Both schools are located in the southern part of the United States. All traditional undergraduate students who attend chapel, or convocation, as part of the curriculum were invited to participate in the survey during the fall 2020 semester.

The traditional undergraduate students for University A comprise the following demographics: 81% are Caucasian, 6% are two or more races, 4% are African American, 2% are Hispanic/Latino, 1.0% are of Asian descent, and 5% represent other nationalities. The studentfaculty ratio for University A is 12:1. The total undergraduate enrollment for University A is over 1,700 students, with a biological gender distribution of 60% female students and 40% male students. At University A, 99% of full-time undergraduate students receive some type of financial aid. Students at University A are required to attend a 50-minute chapel service three times a week.

The traditional undergraduate students for University B comprise the following demographics: 44.2% are Caucasian, 37.5% are African American, 10.4% are Hispanic/Latino, 4.5% are nonresident alien, and 3.4% represent other nationalities. The total undergraduate enrollment for University B is over 900, with a biological gender distribution of 54% male students and 46% female students. Traditional undergraduates at University B are required to have 21 spiritual growth credits per semester and the student-faculty ratio is 14:1.

The gender demographics of the sample included: 104 male participants, 78 males from University A and 26 males from University B; and 180 female participants, 140 females from University A and 40 females from University B. The sample size of 284 (N = 284) is more than the required minimum (N = 144) assuming a medium effect size with a statistical power of 0.07 at the 0.05 alpha level (Gall et al., 2007). The reason for increasing the sample size was to maximize the degree to which the results are generalizable to a wider population. Gall et al. (2007) explained that to achieve population validity, quantitative researchers must randomly select the sample from the population from which they wish to generalize the results. All students who received the survey had an equal chance to complete the survey and participate in the research.

Instrumentation

This study utilized two surveys, combined for distribution, and administered through Qualtrics XM software to obtain data from participants. The surveys used were the Cyberbullying and Online Aggression Survey (COAS) and the Duke University Religion Index (DUREL).

Cyberbullying and Online Aggression Survey

The COAS is a self-report questionnaire that measures cyberbullying victimization and offending among the population surveyed (Hinduja & Patchin, 2015; see Appendix A). This tool measures a wide range of behaviors related to cyberbullying and has demonstrated adequate psychometric properties in 10 different studies from 2009 to 2019 (Hinduja & Patchin, 2019b). The instrument was valid for this study because it measures the differences in the prevalence of cyberbullying and online aggression experiences and includes both the perpetration (offending) and victimization of the construct.

The COAS consists of two parts and includes 20 questions measured on a Likert scale from 0 to 3. The response format used to assess how often each behavior had occurred was as follows: 0 (never), 1 (once), 2 (a few times), or 3 (many times). Students were asked to respond to the question with the answer that best fit their feeling toward that behavior. The survey questionnaire began with the definition of cyberbullying that the student was to use as a guide for their answers. The students had to acknowledge that they had read and understood the definition of cyberbullying used for this research before taking the survey.

There are two scales of measurement on the COAS. One scale measures cyberbullying victimization experiences, and one measures cyberbullying offending experiences. This research study utilized both scales to measure how many times the student had been a victim of cyberbullying in their lifetime and in the past 30 days. The survey also measured how many times the student had been a cyberbully (or offender) in their lifetime and in the past 30 days. The victimization scale included 10 self-report items with an internal consistency (Cronbach’s alpha) from the 10 most recent studies ranging from 0.867 to 0.935 (Hinduja & Patchin, 2019b). The cyberbullying offending scale included 10 self-report items with an internal consistency, from the 10 most recent studies, ranging from 0.793 to 0.969 (Hinduja & Patchin, 2019b). Cronbach’s alpha coefficient measures internal consistency, or how closely related a set of items are as a group (Warner, 2013). The COAS met the standard of an average of 0.80 Cronbach’s alpha; therefore, the survey was reliable and appropriate for this study (Gall et al., 2007).

Scoring for this instrument utilized a summary scale for each of the two sections. Each student’s response score was calculated by adding the scored response of the Likert scale, with a possible score ranging from 0 to 30 for each section; a higher score indicated greater involvement with cyberbullying as a victim or as an offender. For each scale, scores closer to 0 indicated a low prevalence of cyberbullying victimization or online aggression, and scores closer to 30 indicated a high prevalence of cyberbullying victimization or online aggression. A score of 0 on the victimization questions indicated the student reported never being the victim of cyberbullying. A score of 0 on the offending questions indicated that the student reported never being involved as a cyberbullying offender.

Duke University Religion Index (DUREL)

Koenig et al. (1997) created the Duke University Religion Index (DUREL) to measure an individual’s level of religiosity based on his or her organizational religious activity, nonorganizational religious activity, and intrinsic religious activity (see Appendix B). According to Koenig and Büssing (2010), the instrument has been used by researchers in over 100 studies and is available in 10 languages. The DUREL has a high internal consistency (Cronbach’s alpha = 0.71–0.91), a high overall test-retest reliability (intra-class correlation = 0.91), and high convergent reliability when compared to other measures of religiosity (rs = 0.71-0.86; Ameri, Mirzakhani, Nabipour, Khanjani, & Sullman, 2017; Koenig and Büssing, 2010; Sharma et al., 2017). The DUREL meets the standard of an average of 0.80 Cronbach’s alpha; therefore, the survey is reliable and was appropriate for this study (Gall et al., 2007).

The DUREL is a five-item survey consisting of three subscales that measures the following dimensions of an individual’s religious involvement: (1) organizational religious activity (attendance at religious services/activities), (2) non-organizational religious activity (personal religious activities/prayer and Bible study), and (3) intrinsic religiosity (experiences the presence of the Divine (God; Koenig & Büssing, 2010; Storch et al., 2004). Question 1 measures organizational religious activity (Subscale 1) with a Likert score ranging from 1 to 6, with 1 being high involvement and 6 being low involvement. Question 2 measures non-organizational religious activity (Subscale 2), including time spent in prayer or other private religious activities. This question uses a 6-point Likert scale from 1 (more than once a day) to 6 (never). The final three questions used a 5-point Likert scale to measure the individual’s personal beliefs or intrinsic religiosity (Subscale 3) from 1 (definitely true) to 5 (definitely not true; Sharma et al., 2017). For this research, an overall score of 5–13 represented a higher level of religiosity, and a score of 14–27 represented a lower level of religiosity. The composite scores from the three subscales were used to determine an overall level of religiosity (organizational, non-organizational, and intrinsic) for this study (Storch et al., 2004).

Procedures

The researcher obtained permission to conduct the study from both universities (see Appendix H) and secured permission to use both instruments (see Appendices A and B) from the appropriate individuals. Following all approvals and a successful proposal defense, the researcher secured permission from the Liberty University Institutional Review Board (IRB) to conduct the research study (see Appendix C). Following IRB permission to begin the research, the principal investigator contacted the designated faculty representative at both participating universities to finalize survey distribution plans.

The researcher sent an email to each university with the student recruitment letter for them to disseminate to all traditional undergraduate students via their official school email address (see Appendix D). This email clearly explained the study and the length of time necessary to complete the survey (5–10 minutes). The researcher requested that the faculty representative at each university send the email to all traditional undergraduate students who attend class on campus and who are required to attend chapel or religious services as part of their degree program. The initial recruitment email was sent on Monday, September 28, 2020.

All traditional undergraduate students, 18 years of age and older, attending one of the participating universities were eligible to participate. Once the students received the recruitment email, if they chose to participate, a link provided took them to the Qualtrics XM site to take the survey. The first page of the survey contained the approved consent form, and clearly stated that the student could stop taking the survey at any time if they decided they did not want to participate in the research (see Appendix E). As part of the survey, the researcher asked students to identify their biological gender (female, male, or prefer not to answer). Upon completing the survey, the student received a thank-you email that appeared on the last page of the survey (see Appendix F). Two follow-up emails were sent to the students during the approved three-week window for research. The first follow-up email was sent on Monday, October 5, 2020, and the second follow-up email was sent on Monday, October 12, 2020, to ensure the study had the maximum amount of participation possible (see Appendix G). Students had three weeks to take the survey before the principal investigator ended the data collection portion of the research. The principal investigator sent a thank-you email to the administrators from each school who initially granted permission to conduct the study.

Data Analysis

Two separate two-way ANOVAs (analysis of variance) were utilized to analyze the data collected for this study. This analysis was appropriate because it assumes a continuous dependent variable (prevalence of cyberbullying victimization and offending experiences), a categorical independent variable (biological gender and level of religiosity) with two or more independent groups (female and male/higher level of religiosity and lower level of religiosity), and independence of observations (Warner, 2013). H01 and H04 tested the main effect of biological gender, H02 and H05 tested the main effect of level of religiosity, and H03 and H06 tested the interaction effect of the two variables (biological gender and level of religiosity).

Once all data were collected, all submitted surveys were sorted, and only completed surveys were used for data analysis. The researcher identified 284 participants who had completed the entire survey, which is significantly more than the necessary N = 144 to assume for a medium effect. All data were reverse recorded and entered into the SPSS program for data analysis. Data were sorted by biological gender (female/male) and level of religiosity (higher/lower) for the victimization scale and the offending scale respectively. All data have been kept confidential and were coded to ensure students were grouped by biological gender (female – 1, male – 2) and level of religiosity (higher level of religiosity – 1, lower level of religiosity – 2).

The researcher analyzed descriptive statistics using the mean and standard deviation to determine if the groups differed on the prevalence of cyberbullying victimization and offending experiences. Data were visually screened for missing data points and incorrect entries. Data sets with missing data points or having obviously incorrect entries were omitted from the overall data set. Box-and-whisker plots were used to screen the data for extreme outliers.

A two-way ANOVA requires that the assumptions of normality and homogeneity of variances be satisfied. A Kolmogorov-Smirnov (p > 0.05) was run for each group of the independent variable, for each hypothesis, to test for normality because N = 280 > 50. A Levene’s test of equality of error variances (p > 0.05) was also run for each hypothesis to test for homogeneity of variances. The effect size for this analysis on each hypothesis, which is the proportion of the total variance that is attributed to an effect, is reported in Chapter Four. Due to running two tests of significance, a Bonferroni correction was also used to guard against Type I error. The alpha level was calculated as: PCα = EWα/k where PCα is the experiment-wise α, typically α =.05 and k is the number of significance tests performed, which is two in this study. Therefore, PCα = .05/2, PCα = .025, and rounded to α =.03.

Summaries of all collected data and the resulting analysis will be provided to University A and University B once the researcher has defended the dissertation. Each university will then be able to use the data to assist with the process of creating new cyberbullying policies, awareness campaigns, and prevention programs based on the needs of their students. The researcher used the data collected to complete dissertation findings and recommendations.