Methods
This study relied on a variety of methods.3 Using anonymous survey questionnaires, data was collected from 317 undergraduate students (mean age 21.64, SD 3.68) from several colleges and institutions of higher education in Eritrea. Students were selected randomly from enrolment lists provided by the institutions and written consent was obtained from all respondents. Inter-alia, the questionnaire covered demographic characteristics, exposure to and consumption of pornography, as well as attitudes toward women, using the 15-item Attitudes Toward Women Scale (AWS) (Spence and Helmreich 1978).
The AWS is the most widely used instrument to assess attitudes about women’s rights and roles in society, and it has previously been used within research in many countries (Byrne, Felker, Vacha-Haase, and Rickard 2011; Chia, Allred, and Jerzak 1997; Delevi and Bugay 2013; King, Phillips, Walker, and O’Toole 2014). Previous work has suggested that viewing pornography may be associated with sexist and negative attitudes toward women (Bowen 1987; Brown and L’Engle 2009; Flood 2010; MacKinnon 1984, 1989; Peter and Valkenburg 2007; Peter and Valkenburg 2009; Skorska et al. 2018; Tylka and Van Diest 2015). In the present study, the AWS was utilized to help explore the potential relationship between viewing pornography and users’ attitudes toward women.4
In addition to surveys, between August 2018 and March 2019, in-depth, semi-structured interviews were conducted with key informants from across Eritrea. Key informants, selected through convenience and snowball sampling techniques, were from educational institutions, national ministries, and several entertainment video retail outlets and Internet cafes. The wide array of sources helps provide a diversity of views and perspectives, ultimately broadening and enriching understanding of the topic.
Interviews were open-ended and semi-structured (Hammer and Wildavsky 1989), a particularly useful format since the study retained many exploratory features. Furthermore, openended questions offered respondents an opportunity to organize responses within their own framework, thus potentially increasing the validity of responses. While the interviews and questions were semi-structured, they involved a range of follow-up queries and creative locutions (e.g., “why” and “what else”) in order to further probe issues of merit or pursue clarity (Hammer and Wildavsky 1989).
All interviews were conducted in person, generally lasted between thirty minutes and two hours, were conducted in either Tigrinya or English (two of Eritrea’s three working languages) and began after obtaining consent. During all interviews, hand-written notes were taken, with transcription occurring shortly afterward in order to minimize errors and ensure a high degree of accuracy and detail (Newing 2011: 113-14).
Additionally, the paper is guided by numerous informal conversations and several focus group discussions. Focus groups are practical and effective as they can reveal “how several people work out a common view, or the range of views, about some topic” (Fielding 1993: 141). Moreover, the utilization of focus groups helps to stimulate fruitful discussion and bring to the surface responses and perspectives about pornography in Eritrea that otherwise might lay dormant.
3 The study was approved by the College of Arts and Social Sciences Research Committee, and the Department of Sociology and Social Work Research Committee, both located in Adi Keih, Eritrea. All ethical requirements were followed in accordance with appropriate local and international procedures.
4 Each item on the AWS uses a four-point Likert Scale with anchors from strongly disagree to strongly agree. Total scores on the scale can range from 0-45, with higher scores indicating a more positive, egalitarian attitude toward women. Scores for items that were negatively worded are reordered.
For example, participants, on the basis of engaging with others, may be able to articulate more clearly their thoughts or views than they otherwise might if alone (Babbie 2007: 308).
Data analysis is based on data triangulation: obtaining, comparing and contrasting evidence from a wide range of data sources (Bieri 2010; Creswell and Miller 2000: 126; Yin 2003). Relying on multiple sources of data allows for the convergence of various lines of inquiry and strengthens validity (Creswell and Miller 2000: 126 ff.; Yin 2003), and is important in gaining cumulative insights and acquiring a broader, richer perspective of pornography in the context of Eritrea. In addition, descriptive statistics, Chi-square tests and one-way analysis of variance (ANOVA) procedures are utilized to report survey results, explore associations between different variables and determine possible differences in attitudes toward women between respondents that viewed pornography and those who did not.