Data collection procedures

The participants were recruited via convenience sampling. The survey used in this study was available online on http://kerdoiv.magtud.hu as long as the study was active. The study was circulated on university and college e-mail lists from all around the country.

Colleagues and Ph.D. students of the Institute of the Behavioural Institute were also asked to forward the e-mail to their contacts introducing the study and the link to the website. Teachers in the Institute and in other universities were also asked to disseminate the study in their seminars. Furthermore, the study was also shared with the researcher‘s students and acquaintances and they were also asked to forward this information to their acquaintances. Alexandra Béres, who is a fitness world champion

and has her own website, also disseminated this study on her e-mail lists as part of her newsletter to her subscribers. The ―Media and mental health‖ study was advertised on various health, diet, fashion and, mostly, university-related websites as well. Related Facebook pages (with keywords of university, college, student, diet, fitness, lifestyle, health) were also approached and they were asked to submit information about the study and links to the website in new posts on their own Facebook walls. No pages were approached to share the study that seemed to be sharing unhealthy or harmful messages to their users. In addition, those pages whose Facebook content could not be seen were also not approached.

Finally, an article was presented on one of the Hungarian news pages online, which was the http://444.hu. This was the only news page portal that responded to the request to advertise the current study with a neutral article about media and health. The article focused more on the positives when being active online, e.g. interacting with friends on social networking sites, and being able to reach up to date information. This news website is thought to be generally popular among Hungarian young adults. However, no details of the readers are available regarding this websites. The link to the study was present in the article at the beginning and at the end of this article.

Most of the dissemination methods were online methods, except the posters and flyers, which were advertised in local universities, gyms, and workplaces. Results showed that most participants (N=346, 42.2%) heard of the study on some website. The second most often cited source was Facebook (N=300, 36.6%), over hundred people (N=106, 12.9%) heard about the study from emails, and fifty-four people (6.6.%) reported that they got information from an acquaintance. Ten people (1.2%) said that they heard from the survey on a health promotion and prevention website (ProYouth), and four people (0.5%) reported getting information regarding this study via flyers. All participants had to provide their informed consent prior to filling out the questionnaire.

The subjects were not compensated for their participation; however after they finished the survey they were offered a short feedback in case they wanted to leave their e-mail address. No methods were used to prevent participants from filling out the questionnaire multiple times, however, it took approximately 20-25 minutes to fill out the survey and it is unlikely that many people would have wanted to finish the questionnaire twice or more times. Questions could not be skipped as participants were warned when they

missed a question and they could only move to the next page until they answered everything on the current one. The questionnaire was available online on April 23, 2013, and was offline on June 18, 2015.

4.2.2 Online data collection

Online data collection was implemented in the study and most of the dissemination methods were online methods as well, however, we also had a few offline techniques, but these appeared less successful. Based on the report from the Hungarian Central Statistical Office in 2014, in Hungary, 74% of all households had a broadband Internet connection (KSH, 2014). The online survey did not require any specific knowledge to fill in. Unfortunately, when creating the survey we did not have the sufficient technical resources to track all those who received the information regarding the study but decided not to take part. The Internet can function as an effective medium for collecting and exchanging information in psychology-related research (Birnbaum, 2000). In recent years, the use of internet-mediated questionnaire assessment has expanded and constantly growing on the provision of behavioural telehealth (Buchanan, 2003).

Studies on web-based psychological assessment indicate that such measures can be reliable and valid. Studies report that participants are more honest when filling in a test online (Davis, 1999; Joinson & Buchanan, 2001). The Internet as a way to collect data has its advantages because it‘s cheap, can reach various areas, large populations, it holds no time limitation, it is no harder to fill out than a paper-pencil test, and participation is completely voluntary (Riva, Teruzzi & Anolli, 2003).

Due to the perceived anonymity, Joinson (1999) found lower levels of social desirability in filling in online questionnaires compared to paper-pencil test versions.

Furthermore, Buchanan (2003) argued that literature on web-based assessments indicates that these online measures can be reliable and valid in the context of psychological testing (Carlbring et al., 2007; Barak, Buchanan, Kraus, Zack, Stricker, 2004). Many researchers reported that when comparing online and offline platforms for the same set of questionnaires, almost all the time no significant differences were found in the psychometric characteristics of these measures. Studies documented that online data collection can be a valid way to assess people in case of e.g. Internet attitudes (Riva, Teruzzi, & Anolli, 2003), panic/agoraphobia research (Carlbring et al., 2007),

depression (Holländare, Andersson, & Engström, 2010), assessing sexual boredom (Meyerson & Tyron, 2003) and obsessive-compulsive symptoms (Coles, Cook, &

Blake, 2007). However, Buchanan and colleagues (2005) found some contradictory results in their analysis. They found that in case of a prospective memory test the original factor structure of the questionnaire could not be recovered when the test was administered online. They highlight the importance of testing the equivalence between the offline and online measures, as the testing medium, namely the Internet may influence the outcome. Filling out questionnaires online comes with an increased feeling of anonymity, disinhibition effects and reduced socially desirable responding (Buchanan, 2005). When filling out questionnaires online, people have a lack of supervision or no social presence at all (Bartram & Brown, 2004); they might present increased self-disclosure (Joinson, 1999) and different types or degrees of motivation (Buchanan & Smith, 1999). During an online test the surrounding environment and time of submission are also unknown and uncontrollable (Reips & Musch, 2000). In an online survey people are usually self-selected, therefore they do not form a representative sample. Those who like to fill out questionnaires online skew toward the high end of the socioeconomic and educational spectrum (Birnbaum, 2000). Davis (1999) found that those who participated in the online version of a test had higher levels of self-focused negative thought compared to the offline group. Based on the found differences, there are some important implications when one considers using the Internet to collect data, especially in case of normative data (Buchanan, 2003).

Buchanan (2003) implies that when normative data was gathered in a traditional setting it might not be suitable to use as a base for interpreting the scores from online data collection. However, the mentioned differences between online and offline measures are usually minor. Even when the same questionnaire is administered to different samples one might get slightly different factor loadings, and therefore sometimes the difference may as well be the difference between samples. Especially that most measure is tested originally on student samples (Buchanan, 2005). Buchanan and Smith (1999) argues that since online samples are more heterogeneous they might provide a clearer picture of the factor structure of tests than traditional samples, especially when they were filled out mainly by students. Due to the available sizes and diversity in case of online assessing, the Internet could be an important tool. In fact, its weakness can be its

strength and promise a way to recruit a more diverse, representative sample for studies (Reips, 2000).