Student Perceptions of Cyberbullying Interventions

Creating Effective Interventions to Reduce or Eliminate Cyberbullying

Unit 1 MBA 5652, Research Methods

David E. Couch Jr.

The proposed research study explores the perceptions of secondary students with regard to cyberbullying interventions. The hypothesis of the study is that students' perceptions of cyberbullying will be influenced by the instructional and curricular intervention as measured by increased empathy and willingness to report or intervene when they learn about or witness cyberbullying. The research findings will be of interest to educators and parents, and it will add to the literature on interventions for cyberbullying.

Key terms: cyberbullying, bullying, interventions, student perceptions

Creating Effective Interventions to Reduce or Eliminate Cyberbullying

Introduction

Cyberbullying is something new to the youth of the modern world today. In years past, bullying was something that happened within school contexts and at events that students attended where other students their age. In the contemporary high tech world, it is common for students to be online 24/7 with computers, tablets, and cell phones. There are many types of social media websites and apps that students use today to communicate and connect with peers and others (see Figure 1. Teen Use of Technology). The Common Sense Media website lists common dating applications and websites such as Omegle, MeetMe, Skout, and Tinder." (Conway 2015). Texting applications, such as Kik Messenger, ooVoo, and WhatsApp." (Conway 2015). Young people may also find out about secret apps and/or self-destructing applications, such as Snapchat, Burn Note, Whisper, and Yik Yak." (Conway 2015). A wide variety of chatting, meeting, micro-blogging applications and sites also exist, such as Instagram, Tumblr, Twitter, and Vine." (Conway 2015).

The Cyber Bullying Research Center (mediastampede.wordpress.com) illustrates the technology that is being used by tweens and teens from 10 to 18 of age.

Figure 1. Teen Use of Technology

Source: Hinduga, S. & Patchin, J.W. (2010).

Educators have a duty to act and must work with students to develop citizenship and healthy personalities, and to improve students' self-responsibility. Educators are aptly placed to conduct interventions to alter behaviors and attitudes that are linked with victimization or bullying in person or online. There is opportunity to facilitate education programs to address the issues of bullying and victimization with students in school settings -- which is of critical importance before they become adults.

Cyberbullying vs. bullying in schools.

Experts recognize how bullying has changed from being picked on in school or in the neighborhood to being a non-stop assault that can cause an otherwise well-adjusted child to take their own life (Rice, et al., 2015). Cyberbullying can be more insidious than traditional bullying, since cyberbullying can reach wide audiences (e-mails sent to everyone at a school), can be done in most cases anonymously. Cyberbullying has received widespread media attention as students have killed themselves, often posting online before doing so with no one taking action. Anti-bulling programs exist at all levels of education, including higher education. A strength of a number of the current programs is that they include students who once bullied or harassed other students, either in person or in digital environments. The literature on bullying and cyberbulling personalities has grown right along with the problem:

"Cyberbullying perpetrators are more likely to have problems with their behavior, peer relationships, and emotions, and are less likely to be prosocial than their peers who are neither cyberbullying perpetrators nor victims of cyberbullying. Specifically, female cyberbullying perpetrators express greater anxiety and depression than their female peers who are not cyberbullying perpetrators" (Rice, et al., 2015).

Proposed Study on Bullying vs. Cyberbullying

Definitions of bullying may vary but generally the unacceptable behavior is seen as having a repetitive aspect that is perpetrated by only a few individuals -- whether students or co-workers -- and consists of a highly adverse actions or communications (Harris, 2012). Hirsch (2014) stated, "cyberbullying is the use of technology to harass, threaten, embarrass, or target another person. By definition, it occurs among young people." Bullying can be thought of as an artifact of the existence of groups -- in other words, a long-standing problem -- whereas, cyberbullying is has only been around for a few decades -- following the trajectory of technology innovations -- and has only really been studied in the past 11 years. One of the first reported cases of cyber bullying was in 1998 when a student in middle school had designed a website that contained derogatory statements and threatening messages about his teachers and school administrators (Citation, XXXX).

A gap in the body of knowledge exists about how to prevent cyberbullying and how to help the victims of cyberbullying. Ferlazzo (2010) conducted a study to ascertain the number of students who been the victims of bullying and who have bullied other students (see Figure 2. Cyberbullying Offending and Figure 3. ). Many students have been the victims of but thy have also been the one doing the bullying as well (Hinduja & Patchin, 2010). Hinduja and Patchin's (2010) study particularly contributes to the theory as it illustrates that a wide range of types of bullying exist that are just a step away from what one would think of as bullying of the in-person form.

Figure 2. Cyberbullying Offending

Source: Hinduga, S. & Patchin, J.W. (2010).

Figure 3. Cyberbullying Victimization

Source: Hinduga, S. & Patchin, J.W. (2010).

Intervention Effectiveness for the Prevention of Cyberbullying

Cultural imitation can play a powerful role in the propensity of people to behave in particular ways, a factor that research on the influence of television on the behavior of children -- especially -- and adults clearly shows (Bandura, 1961; Bandura, 1973; Centerwall, 1992; Emmons, 2013; and Philips, 1983). Bandura and other social psychologists have devoted their professional careers to understanding the influence of imitation on human behavior. Occasionally, an outstanding bit of research reveals an unappreciated association between behavior and a seemingly cultural artifact, as did the seminal research of Philips (1983), discussed below.

Much of the research conducted on the influence of television viewing took place in the period from the 1970s to 1990s, which roughly marks a heyday for TV broadcasting. According to Centerwall (1992), "All Canadian and U.S. studies of the effect of prolonged childhood exposure to television (2 years or more) demonstrate a positive relationship between earlier exposure to television and later physical aggressiveness" (Centerwall, 1992). Centerwall concluded that preadolescent childhood is the critical period for strength of effect when exposing children to television violence, and that an "aggression-enhancing" chronic effect lasted well into adolescence and even adulthood. Televised violence may have more pervasive effects than intuition or conventional wisdom would indicate.

A seminal study of the effect of mass media violence in the U.S. showed that publicized prizefights were associated with a 12.46 increase in homicide rates between 1973 and 1978 when the research was conducted (Philips, 1983). It is important to recognize that the brief and sharp increase was statistically significant, accounting for the rise in homicides after correcting for seasonal variables, secular trends, and other factors found to be extraneous (Philips, 1983). Philips concluded that the cultural approval and excitement of fans surrounding the heavyweight prizefights can serve to "stimulate fatal, aggressive behavior in some Americans (Philips, 1983).

These findings, taken together, indicate the suggestibility of human behavior and point toward interventions to reduce cyberbullying need to be directed at children and the adults who care for them and teach them (Bandura, 1961; Bandura, 1973; Centerwall, 1992; Emmons, 2013; and Philips, 1983). Indeed, a recent study published in Pediatrics, the medical journal of the American Academy of Pediatrics, demonstrated that when three to five-year-olds watched television shows that emphasized empathy and cooperation rather than shows that depicted aggression, their behavior improved after just six months (Saint Louis, 2013).

By extension, cyberbullying may be viewed as harmful, aggressive behavior that is influenced by cultural sanctions, even when those sanctions are an aspect of pre-adolescent or adolescent subculture. As Emmons (2013) and Saint Louis (2013) indicate, the behavior of young children is influenced by the type of television programs they view, yet Centerwall's review conducted in 1992 points to a critical age -- a period in human lifecycle that can be likened to Lorenz' (1988) observations on imprinting. However, Bandura's social learning theory and research suggests greater degrees of fluidity do occur people's behavior. Bandura emphasized several models of learning, including observation of live models, verbal instruction models, and symbolic models. An important tenet of Bandura's theory is that learning does not always bring about changes in overt behavior, which is partly due to the importance of the mental state of the learning and the type of external (positive or negative) reinforcement and internal reinforcement (feelings of competency, pride, belonging) that is associated with the learning (Bandura, 1973; Bandura, et al., 1961).

Research Question

The research question is whether the perceptions and behaviors of middle school age students can be measurably altered through exposure to curriculum and instruction that emphasizes empathy for victims of cyberbullying, and elicits strong disfavor and a willingness to engage in whistleblowing with regard to students who initiate or participate in cyberbullying.

Methodology

The proposed research will use a school-wide curriculum designed to reduce cyberbullying that exposes middle school students to media exemplifying empathy toward victims, indignation in response to occurrences of cyberbullying, and engages the students in discussions about alternative behavior such as whistleblowing. The research will be a quasi-experiment in that it does not entail the use of randomization. The primary reason for not randomly assigning the students to a group is that the curricula will be used with entire classes in which students have previously been assigned. This is a condition imposed by the schools participating in the study. Some threats to internal validity are associated with this configuration, but the researcher will make every effort to control the variables that could alter the experimental findings. For this reason, all independent variables that have been associated with higher risk of experiencing bullying will be considered factors in the statistical analysis.

Hypotheses

Ho = Students' perceptions of cyberbullying will be influenced by the instructional and curricular intervention as measured by increased empathy and willingness to report or intervene when they learn about or witness cyberbullying.

Ha = Students' perceptions of cyberbullying will not be influenced by the instructional and curricular intervention, with essentially no changes in measures of increased empathy or willingness to report or intervene when they learn about or witness cyberbullying.

The dependent variables include measurable changes in student perceptions about cyberbullying, actual occurrences of reporting cyberbullying to school officials and of actions that can fairly be described as intervening when bullying is observed. The independent variables include the introduction of the anti-bullying curriculum, classroom based instruction that follows the curriculum, and exposure to media that illustrates the importance of empathy and positive social interaction. Blocking variables in the proposed study would include gender, gender preference, ethnicity, race, and other factors that research has shown increase the risk of bullying, such as being overweight, underdeveloped for age, wearing glasses with thick lenses, or having a speech impairment, or socioeconomic inequality. Extraneous variables would include achieved grade in middle school, or the teacher providing the anti-cyberbullying instruction.

A survey has been developed to explore consumer attitudes about the use of technology for content distribution. Subjects will be selected by using a process known as random sampling. Probability statistics rely on random sampling, which effectively eliminates bias -- or at least reduces it to acceptable levels. The survey will be conducted over a two-month period to collect data on how technological change and implementation influences the perceptions, attitudes, and behaviors of the survey participants.

The survey participants will use a 5-point Likert scale to record their responses. A Likert scale is a version of a summated rating scale, which is configured in a way that enables the conversion of text responses to quantifiable categories which can be summed to reflect the relative differences of the individual or aggregate responses (Jupp, 2006). Even though there are no correct answers attached to the items, the summated rating scale results in better reliability than a single rating scale tends to provide (Martinez-Martin, 2010). A primary drawback of a summative score of the ordinal numbers from a Likert scale is that it the score imparts a sense of meaning that is not truly representative of any real magnitude (Martinez-Martin, 2010). Statistical analyses will be utilized to determine relationships between the responses to the questions. Accordingly, statistics will be used to determine acceptable rates of reliability, validity, and sensitivity.

All subjects will agree to and sign an informed consent agreement in order to participate in the study, and subjects will be informed that they may withdraw from the study at any point in time. The identity of the respondents will be kept confidential by assigning a number identifier to each respondent and not using their names or any identifiable information in association with their survey responses. No one who has a conflict of interest with the study will be permitted to engage in the inquiry in any way. Debriefing of individual responses on the open-ended survey question will take the form of a member check to ensure interpretation of the data is accurate. Any human subjects paperwork and processes required by the academic institution will be completed and adhered to absolutely.

The survey questions that constitute the questionnaire are all written in a manner that enables survey participants to respond according to a Likert scale. Since the points of the Likert scale directly translate to numbers, quantitative data analysis is supported. Since a Likert scale is actually one version of a summated rating scale, the participants' responses can be summed to reflect the relative differences of the individual or aggregate responses (Jupp, 2006; Martinez-Martin, 2010)). A summated scale is used to measure a construct, which means that the correlation of the measures to characteristics that re related and unrelated to the construct are of central importance. A summated scale is expected to exhibit consistency as demonstrated by test-retest reliability and across the questions included in the survey, which is an indication of the internal consistency of the questionnaire (Spector, 1992).

A summated scale that has not been validated needs to undergo a pilot test with a sufficiently large sample, say 100 to 200 people, in order to determine if reliability and validity measures are within acceptable limits (Spector, 1992). Factor analysis can be used to demonstrate that questions on the survey are related, and to help identify the questions that provide the most robust answers (Darlington, 1997). The development of a summated scale is a complex, challenging endeavor -- done correctly -- and would not be practical for a survey of the proposed size and purpose (Darlington, 1997).

Because a summative score of the ordinal numbers from a Likert scale does not truly represent any real numerical magnitude, the scores will be analyzed using descriptive statistics. This will help to provide contextual information and relative perspectives of the composite and individual scores (Martinez-Martin, 2010). The two main types of descriptive statistics are measures of central tendency and measures of spread (Punch, 2003). Measures of central tendency, as the term implies, provide a description of the central point of a frequency distribution for a data set (Punch, 2003). The most commonly known measures of central tendency are the median, mean, and mode (Punch, 2003). Measures of spread show how the scores in a data set spread out, generally across a normal curve (Punch, 2003). The measures of spared provide information in terms of quartiles and range, and -- importantly -- absolute deviation and standard deviation (Punch, 2003). Both of these descriptive statistics types can be summarized in graphical and tabulated form, as well as through a statistical commentary on the results (Punch, 2003).

The descriptive statistics that will be used in the study include the following: 1) The mean for each question item; 2) the median for each question item; 3) the range of scores for each question item; and, 4) the absolute deviation of the scores for each question item (Punch, 2003). The graphical displays that are the most appropriate for this data include frequency tables, histograms, and pie charts with percentages of overall scores and scores for each question item.

Since the mean is essentially a model of the data set, in that it includes every value in its calculation set of data, the mean is associated with the least amount of error. This suggests that the mean will minimize error when used to predict any single value in the data set. But since the data from the summated scale is ordinal, not interval, the median is actually the most appropriate measure for the data (Punch, 2003). The median is the middle score in a data set that has been arranged according to the order of magnitude (Punch, 2003). A primary advantage of using the median is that it tends not be affected by skewness and outliers, which can be especially important when a small sample is used (Punch, 2003). Since the standard deviation is used for continuous data and not categorical data, the measure of absolute deviation is of more interest, and it is not as affected by outliers or skewed data as the standard deviation (Punch, 2003).