Validation of networks derived from snowball sampling of municipal science education actors

ABSTRACT

Social network analysis (SNA) has been used in many educational studies in the past decade, but what these studies have in common is that the populations in question in most cases are defined and known to the researchers studying the networks. Snowball sampling is an SNA methodology most often used to study hidden populations, for example, groups of homosexual people, drug users or people with sexually transmitted diseases. By use of a snowball sampling approach, this study mapped municipal social networks of educational actors, who were otherwise hidden to the researchers. Subsequently, the resulting networks were validated through interviews with central respondents and by using prior investigations of the municipalities. Our results confirmed that the majority of the designated respondents recognized the resulting sociograms and their own position in these. It was also found that not all predictions based on existing knowledge of the municipalities aligned with SNA data. However, these discrepancies could be explained by development in the municipalities in the time following previous investigations. This study shows that snowball sampling is an applicable method to use for mapping hidden populations in educational settings, but also that qualitative studies are needed in order to interpret the networks in real-life contexts.

KEYWORDS:

• Social network analysis
• snowball sampling
• mixed methods
• data validity
• Science Municipalities

Acknowledgements

We owe great thanks to the colleagues at Department of Science Education who have qualified the study. A special thank to all respondents including MSE coordinators in Science Municipalities and lastly a special thank to industrial supervisor Hans Colind Hansen.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1 In the cases where no vertex was named by at least 25% of the positive responses, the vertex with the highest indegree was selected.

Additional information

Funding

This work was supported by the Innovations Fund Denmark (Innovationsfonden); the Lundbeck Foundation (Lundbeckfonden).