j_connections-2019_045_tab_001
| Response rate | 71%–100% in schools, 41%–100% in classrooms; see Section 2.4 of the Technical Report in Vörös et al. (2022) for exact figures. |
| Non-respondent bias | Some classrooms have high rates of non-response and should be used with caution or omitted from certain analyses. We had no means to test whether non-response was random or systematic due to lack of data about non-respondents. Based on informal discussions with teachers, students with attendance problems were likely overrepresented among non-respondents. In many cases, these students were not active members of the class community and networks. This suggests that our response rates are conservative. |
| Theoretical grounding | Our project was inspired by classic studies of school community dynamics using network questionnaires (e.g., Coleman, 1961). The study design aimed at creating a dataset that is suitable for analyses by state-of-the-art statistical network models, with focus on network-network and network-behavior co-evolution (Steglich et al., 2010). The network item battery was created based on earlier research into the multiplexity of peer perceptions in school settings (see Vörös & Snijders, 2017 for a review). The individual items were based on widely used questions in sociological and education research, which were adapted to our empirical context. |
| Publications using the data | There is a growing list of research publications using the dataset. We are currently aware of over a dozen peer-reviewed research articles in English and Hungarian, five defended PhD theses in English, and an edited volume in Hungarian. We provide highlights of this body of work in Section 5 of this article. For a list of publications as of 2022, see Section 4 of the Technical Report in Vörös et al. (2022). |
| Empirical context | Secondary-school classrooms (9–11th grades) in Hungary. |
| Respondents | Secondary-school students (9–11th grades) from seven schools. |
| Longitudinal | Yes. Data was collected in four waves in 2010–2013: October/November 2010, April 2011, April 2012, and April 2013. |
| Temporality | Low. While student values and behaviors may be characteristic of the time and place, general social network processes are likely comparable in student communities elsewhere and in other periods. |
| Analytic utility | The data can be especially valuable for studies into the multiplexity and dynamics of social networks and social influence processes in school settings. We highlight a few key substantive areas in Sections 5 and 6 of this article. Further pointers can be read in Section 5 of the Technical Report in Vörös et al. (2022). |
| Known issues | Some classrooms may not be suitable for complete-network analyses due to low response rates. Some may not be appropriate for longitudinal analyses due to high student turnover. See Sections 2.4–2.6 of the Technical Report in Vörös et al. (2022) for details. |