**2. Materials and Methods**

The results of this study are based on a quantitative consumer survey conducted between 11 July and 14 August in 2018 with 1002 respondents. For data collection, personal sampling method was used with a questionnaire designed to be suitable for self-administered completion. Research was conducted at crowded tra ffic junctions in di fferent Hungarian cities: Budapest, Dombóvár, Eger, Füzesabony, Gy˝or, Kiskunfélegyháza, Miskolc, Siófok, Szeged, Székesfehérvár, Szolnok, and Veszprém. In terms of sex, age, and geographical distribution (NUTS-2) of the respondents, the sample is representative of the total adult population of Hungary, based on the latest census [53] at the time of the data collection (Table 1). To ensure representativeness, we employed quota sampling. During the research design, besides general socio-demographic characteristics, we aimed to collect data on some further particular conditions that may a ffect food consumption directly according to literature [29] (Table 2).

In the beginning of the interview, the respondents were informed about the aim of the research and the managemen<sup>t</sup> of anonymous data. If the respondents were willing to participate, before the research questions were asked, the quota parameters (age, sex, geographical location) had been recorded, that allowed the quota numbers to be tracked by the interviewers to ensure an appropriate level of representability. Although the questionnaire was designed to be self-administered, interviewers provided help to fill the questionnaire, which was important in the case of older respondents.

The questionnaire contained 288 variables, from attitude-related questions through to nutritional claims and carrier foods to questions focusing on diseases. The questionnaire employed closed-form questions predominantly. Many questions were measured on five-point Likert scale, where grade 1 meant "strongly disagree" and grade 5 meant "strongly agree." Table A1 in Appendix A shows the content of the questionnaire in terms of all variables used in this study.

Statistical analysis of the data was carried out by IBM SPSS Statistics 22.0 software package. Beyond descriptive analysis, Kruskal-Wallis test and Pearson's chi-square test (CI: 95%) were used to analyze data on ordinal scale when the distribution of data did not meet the criteria for normal distribution [54]. Factor analysis (principal component analysis—PCA) was used to explore overlaps and to combine correlated variables [55].


**Table 1.** Representative socio-demographic characteristics of the sample (% of respondents, n = 1002).

\* Latest census data of Hungarian Central Statistical O ffice to adult Hungarian population [53].


**Table 2.** Further socio-demographic characteristics of the sample (valid % of respondents).
