Job Grant: analysis of consultation responses

2 Methodology

This chapter details the methodology used to analyse the responses to the Job Grant consultation.

The data was first cleaned. There were no campaign responses that had to be dealt with. However, one respondent completed the consultation twice. In the later submission, they gave a different answer to questions 1 and 2, but the answers to all other questions were the same. In order to avoid double counting, we worked on the assumption that this respondent had changed their mind about questions 1 and 2 and took only the later response into account for the data analysis. This changed the percentages of yes and no answers to questions 1 and 2 only marginally and did not affect the overall tendency. In the interest of transparency, we have added a note to the data analysis where relevant.

The subsequent data analysis had a quantitative and a qualitative component. The quantitative analysis covered:

• The number and proportion of individuals and organisations who answered the consultation
• The number of respondents who answered each yes/no and each open-ended question, broken down by respondent type
• The number of respondents who answered each multiple-choice question by yes or no, broken down by respondent type.

The main findings from the quantitative analysis are presented in the Findings Chapter and all findings from the quantitative analysis are included as Appendix 2.

The qualitative analysis of open-ended questions involved the following stages:

• Stage 1: Based on 10 randomly selected sample responses, an initial thematic grid was developed for the analysis. This grid was further refined during the analysis.
• Stage 2: All open-ended responses were "coded" using the software NVivo and the thematic grid developed in stage 1.
• Stage 3: NVivo was then used to analyse the full range of views and issues raised by respondents and, where possible, quantify how often certain views were raised by different respondent types. As part of this, areas of consensus and disagreement were identified as well as any differences and similarities in views between respondent types.

The full qualitative analysis is presented in the Findings Chapter. Quotes are included where possible without breaching confidentility, in other words, only if a relevant respondent has consented to their response being published.

It is important to note that, given the number of respondents and the fact that they were not selected randomly, the views emerging from this consultation are not representative of the view of the whole population. But they give a good indication of the range of opinions held by a notable number of interested organisations and individuals and the main issues they identify as important.

Another limitation of the analysis is that, given the available data and its level of detail, it was not possible to break down responses by geography (especially urban/rural) or distinguish between types of organisation.