1.3 Methods of data collection
The study has combined qualitative and quantitative methods for data collection and analysis. Data collection was conducted by Oxford Research’s consultants in their respective native language using templates with data collection instructions, research questions, and interview guides. It was organised in three distinct phases: a contextual mapping of the language training systems, a survey with participants and providers, and case studies of services considered to be examples of best practices.
The purpose of the contextual mapping of language training systems was to gain an overview of the language training systems in Denmark, Finland, Norway and Sweden, mapping and identifying both formal and non-formal training and how they relate to different sub-categories of immigrants. The mapping was based on desk research of available reports, studies, websites, legal documents etcetera, at national and Nordic levels. In addition, we conducted interviews with ministries and authorities, subject matter experts, employer and employee organisations, and providers of formal and non-formal training in the studied countries (see Table 3 for breakdown). We also collated quantitative data concerning participation in language training and results of formal language training services. Public data was, however, only available for Norway and Sweden. For Denmark, the cost of purchasing data was considered to be too high and for Finland no national-level data was available.
Based on findings from the contextual mapping, we conducted a survey with participants and providers. The survey focused on how providers and participants perceive the quality and benefits of the language training offered on a wider scale. The survey was sent to selected providers of formal and non-formal language training services, who were asked to distribute it to current and former participants in their services. In total, 98 participants and 77 providers responded to the survey. However, the response rate to the survey was uneven, particularly at participant-level. Participants from Norway were underrepresented, whilst participants from Sweden were overrepresented. More than twice as many women as men responded. In addition, the level of education among the respondents was high, with 67 percent of respondents having completed more than 12 years of education. The survey was distributed in Danish, Finnish, Norwegian, Swedish, and English. This meant that the respondents were required to have adequate literacy, and either a basic understanding of their new language or sufficient proficiency in English to be able to respond. Altogether, this means that the sample of participants cannot be considered representative of immigrants in the Nordic countries. To validate the results, we complemented the findings with interviews with participants, providers, and experts. The responses from providers were more balanced. Nevertheless, Denmark stood out has having a high proportion of responses, whereas Sweden’s was low.
Based on an analysis of the findings from the contextual mapping and survey, we determined a list of tentative quality criteria. The criteria were used to select two language training services per country for case studies to gain a deeper understanding of best practices and what seems to make particular types of language training successful for particular immigrant sub-groups. The case studies explored the short- and long-term results of the selected training services, how the processes work and why they seem to be successful, challenges for the providers and participants, success factors, and which learnings could be disseminated to other providers. Data collection was conducted in different ways depending on the case. It included studies of documentation, interviews with providers, participants, and (where relevant), employers, as well as site visits. The case studies are summarised in Annex A.