In the qualitative stage of this research and based on the opinion of experts and theoretical bases, health media messengers were categorized into 6 components and 15 items, and then by using exploratory factor analysis, the validity and reliability of the instrument were examined. A summary of the results are illustrated in Fig. 3.
Psychometric Analysis of Instrument Components
In the initial analysis of items of the questionnaire, based on face content validity, which was conducted by a survey of experts, 15 items were confirmed and all the items in the qualitative section were also verified. In the content validity phase, the results showed that CVI was confirmed favorably with the indicators of simplicity (0.973), clarity (0.978) and relevance (0.973) and the cumulative average (0.978). In addition, item analysis revealed that all the items had a score of over 0.80, which manifested the content validity of the items.
Inasmuch as the number of experts who participated in the confirmation of content validity was 10, based on the content validity indicator provided by Waltz and Bausell (1981), the validity of the whole questionnaire and each questionnaire item have been confirmed to a great extent [32]. In the CVR index, the degree of necessity of each item was reviewed, and based on Lawshe's content validity ratio index, when the opinions of 10 experts are collected, the average agreement of the opinions is expected to be at least 62%31.
In the current research, the total average was 0.830, and more than 80% was confirmed for each item except for item number 13. The only item with a little agreement on its necessity was "Using Iran's domestic social networks" where in the agreement rate was 0.4. Of course, due to its confirmation using CVI, there was a necessity to compare it with other social networks, we kept the item in our study to examine it separately in the construct validity phase.
Regarding the reliability of the tool, we checked the internal consistency of the questions using Cronbach's alpha, and the internal consistency of the whole questionnaire was 0.881. Cronbach's alpha is one of the most common tests to measure the internal consistency of questions and to determine the reliability, especially in Likert scale questionnaires [37 & 38]. The amount of Cronbach's alpha index ranges from 0 to 1, and the closer it is to 1, the greater internal consistency it has with the dimensions of the questionnaire [39, 40, 41].
The reliability of the whole questionnaire is expected to be at least 0.70 and the amount between 0.8 and 0.9 is excellent [43, 44]. Moreover, in the If Item Deleted mode of the SPSS software, each item was deleted one by one in order and the reliability of the rest of the items was calculated. If the reliability increases by removing an item, the item is problematic and we can edit or delete the option. As shown in the reliability results of the questionnaire, by removing each item, the reliability of the rest of the questions decreased which indicated that each items had a positive effect on the overall reliability. Further, when the reliabilities of sub-groups and sub-scales were calculated, the highest reliability was related to the subscale of Mobile-based social networks (0.761) and the lowest was connected with mass media (0.633). Another significant point in the analysis of Cronbach's alpha was that the value of Cronbach's alpha is not only affected by the internal consistency of the questions, but also by the number of questions, namely the less the number of questions, the lower the Cronbach's alpha [44, 45]. If the number of questions in the present questionnaire increases, it is expected that the reliability will also increase.
But the most important part of determining the validity of the instrument was calculating the construct validity using factor analysis. The validity of the instrument was confirmed based on the significant statistical indicators and the factor analysis. With KMO = 0.928 and Bartlett's sphericity < 0.001, the adequacy of the sample size and the appropriateness of the test were corroborated. Basically, the value of KMO > 0.90 is an excellent indicator [34, 35, 36].
Likewise, according to Table 5, communalities explained by each of the variables exceeded 0.6. The minimum expected value for this index is 0.5 35 and 0.4 and 0.3 values are also acceptable in some articles [37]. So the value 0.6 is a suitable index. In accordance with the initial classification of the quantitative phase which was based on the essence of the items, 15 items in 6 categories of academic messengers (books and articles), official messengers (website and text messages of the Ministry of Health and Universities of Medical Sciences), internet search, Social media, common mobile-based social networks (WhatsApp, Telegram, Instagram and Iranian social networks) and informal methods (conversation with colleagues, classmates and family) were categorized. Severin & Tankard (2010) also put formal and informal methods in two separate categories [28], but in the analysis of the exploratory factor analysis of this research, the two formal and informal categories were merged together and placed in one category.
Although, at first glance, these two categories look dissociated in terms of title, the information of the government sources of the Ministry of Health (the website of the University of Medical Sciences and the notification messages of the Ministry of Health) is included in the Medical Sciences category. On the other hand, the category of informal sources in this research was defined by two items (talking with friends, colleagues, and classmates, and talking with the family, friends, and relatives). It should be noted that our population and statistical sample are from students of medical sciences. Accordingly, most of their friends, colleagues and classmates are also students of the University of Medical Sciences. Therefore, the combination of these two categories (Formal /Informal) can probably be delineated by the construct "The Context of the University of Medical Sciences". In fact, if our statistical population is not from the students of the aforementioned university, these two areas may be separated. Once again, we conducted the exploratory factor analysis with the assumption of 6 factors. Interestingly, these two areas were separated according to theoretical bases and finally 6 predicted categories were confirmed. Additionally, the present questionnaire covers more than 0.64 of the concept of health messengers. We checked the messaging methods that were more common, and some methods such as digital boards installed in the city are methods that were not mentioned in the study.
Analysis of Students' Preferences in Selecting a messenger (Factor Priority)
The second part of the results deals with the analysis of students' preferences in media selection, which was based on the factor load of the components. On the whole, five factors were extracted in the factor analysis of the questionnaire which explained about 64% of the construct according to Table 6.
Social Networks and Messenger Applications
Messengers based on social media was the first messenger factor that had the highest factor load and was the students’ first choice. Similar research has reported that medical students are busy either in the classroom or in clinical settings and have access to smartphones more than anything else. Therefore, it can be predicted that they mostly have access to the news of health messages through mobile phones rather than anything else [46, 47, 48].
Raiman et al. (2017) found that one of the reasons for the preference of social media among the students is the availability and convenience of using instant messages; the immediate benefits of messaging to enhance understanding and learning, and the ability to access recorded discussions and using “voice” to ask questions.
In fact, in addition to being available, it is a two-way communication with feedback [47]. The factors of age and generation are also effective. In his research, Kubrick showed that the preferences of old people are different when compared to young people regarding the choice of messengers [49]. Inasmuch as students mostly belong to the digital age, they prefer electronic devices and feel comfortable with them. Moreover, mobile-based approaches are their priorities and had the highest factor load.
Among the items of this component, four messenger of Instagram, WhatsApp, Telegram and Iran’s national messenger were examined. The highest factor load was related to Telegram (0.85) and the lowest to the items of national social network messengers (0.47). If we deleted national messengers from the questionnaire, the impact of this factor would be greater. The reason why Telegram was preferred was probably due to the capabilities of the social network of Telegram in transferring different pictorial and textual contents and the recovery of files in Telegram. For example, there are some membership limits for channels and groups in WhatsApp, while Telegram enjoys a better status. Moreover, speed and technical problems of national social networks influences people’s choice. The preference of international over national networks may be because of the open virtual environment which is not limited to a particular time or space, is used by all and is not localized. It seems that mobile social media will not disappear in the near future. Therefore, we should ponder extensively over mobile infrastructure as well as how individuals incorporate them into their everyday lives [50].
Formal and Informal Messengers
The sources of health knowledge are divided into two categories; formal (official) sources (doctors and health service providers) and informal sources (family members) [1]. However, in the current research, the second factor in the health messaging questionnaire was based on the opinions of medical students, government messengers affiliated with the Ministry of Health, and informal messaging through colleagues, classmates, friends, and family.
Though at first look the two groups look far apart, as conversational information which is exchanged among friends and families of students of universities of medical sciences come from resources affiliated with those universities. Therefore, integration of the two factors can be labelled as ‘University of Medical Sciences setting’. It is crystal clear that in the aforementioned setting and the health affiliated fields the websites of the Health Ministry and universities of medical sciences are well-liked and placed as the main page on students’ personal computers as most of the university news is broadcast via them. On the other hand, university students follow many protocols, guidelines and news from the university site; thus, this component is among the highest priorities and the foregoing students share the news with their classmates, colleagues and families. It seems that in those studies the population of which come from universities other than universities of medical sciences nor from the related fields, the two areas are separate and independent. This point needs to be further checked in other studies. For example, in a research study it was illuminated that the first health messenger used was TV and Iran health networks were preferred followed by specialists, public newspapers, radio and internet. Satellite channels were on the bottom of the list. They also found that, although mass messengers may be the first source of news via which many people receive information, they do not suffice and people prefer face to face interpersonal communication to supplement or ensure the accuracy of information [3]. In other words, the combination of collective information (one-way) along with individual face to face communication is considered a more complete method, and people are eager to investigate practical experiences of others on the official news they hear.
According to the results of the research, the fifth messenger factor is mass media such as radio, television and publications. Today, due to the availability of electronic publications, social media and social channels related to health which are more accessible, the use of paper and printed methods has diminished. Since our sample was recruited from medical students who have greater access to more reliable sources such as professors, specialists, the and websites of the university and the Ministry of Health. In healthcare environments such as hospitals and clinics, access to televisions or radios may be limited for medical students and trainees. However, with the widespread availability of internet access and mobile devices, searching online has become a more accessible and convenient option for obtaining information and learning. As such, digital resources and online platforms have become increasingly important for medical education and training.
However, different groups have different preferences. For example, in Razavi et al.'s (2016) research, among health messengers, radio, television, individual face-to-face training, newspapers and magazines, CDs training, holding training courses in the form of pamphlets, were announced to be the highest priority [60]. It seems that "availability" in time and space is an important point in choosing media. For example, in the case of ordinary people like housewives, TV is the highest priority, but for a driver or an employer with an average education, radio is a suitable choice. Among people with specialized education, news is more important due to greater availability of specialists and reliable sources.
Interestingly, in our research, students used both official and unofficial news. In Khaniki's research, after receiving news from television and radio sources, individuals’ next priority was interpersonal communication [3] accounting for the fact that we need face to face communication as communication is an important act of interaction [61] and we have an interest in sharing with others what we hear (dissemination) or testing its validity. Exchanging messages may be a means to interpersonal communication.