Measurement: Why It Matters

Prevalence rates of cyberbullying vary widely across studies, ranging from 6% (Ybarra, Mitchell, Wolak, & Finkelhor, 2006) to 72% (Juvonen & Gross, 2008) per (school) year. This wide variation is often the source of confusion among politicians and the public, and it can lead to distrust in the data. Reasons for the variance in rates are unclear. It seems likely, however, that much can be attributed to differences in measurement and methodology. Data are collected using national random telephone surveys (Finkelhor, Mitchell, & Wolak, 2000; Wolak, Mitchell, & Finkelhor, 2006), online random surveys (Ybarra, Leaf, Diener-West, 2007), online self-selected surveys (Patchin & Hinduja, 2006), and offline convenience samples (Juvonen & Gross, 2008; Raskauskas & Stoltz, 2007). Sampling frames vary from national, to international, to local, and from random to convenience (See Chapter 8, this volume). Furthermore, some studies include youth as young as 10 years of age (Dehue, Bolman, & Vollink, 2008; Finkelhor et al., 2000; Ybarra et al., 2007), while others focus on high school (Raskauskas & Stoltz, 2007). Given noted age differences in prevalence rates (Dehue et al., 2008; Kowalski & Limber, 2007; Slonje & Smith, 2008; Ybarra et al., 2007; Ybarra & Mitchell, 2008), this alone would explain some of the variance. Another variation in measurement is the time frame queried: The longer the time frame used, the greater the number of people who will have had the experience in question. For example, if items on the survey inquired whether the respondent had ever experienced cyberbullying, we would expect more people to say yes than if the survey asked respondents if they experienced cyberbullying in the past couple of months. In this section, chapters will focus on issues specific to measurement because it is essential to quality research in the field of cyberbullying.