Women in farming and the agriculture sector: research report

2 Methods

The research findings presented were derived from a combination of qualitative and quantitative data collection approaches and analyses, undertaken between June 2016 and March 2017. All procedures were evaluated and approved by ethical review panels at Queen's University Belfast and the James Hutton Institute.

2.1 Qualitative approaches

The research questions were explored through interviews and focus group discussions with women and men involved in farming, crofting and the agricultural sector across Scotland. Specifically, the interviews and focus groups included women who are new entrants to farming, as well as those who are involved in agricultural industry leadership, estates and large-scale farms, crofting, and farm diversification. A sample of men involved in farming, crofting and the agricultural sector were also included as interviewees and focus group participants. Interviews and focus groups were also arranged with women who work in non-farming family businesses, in order to provide a comparison to farming businesses.

A purposive sample of interviewees and focus group participants were identified through contacts with local gatekeepers, as well as from previous research contacts, web searches, and recommendations from earlier interviewees/participants ( i.e. a snowball sample). Interviews were conducted face-to-face, and where necessary, in small groups of two to three interviewees (plus one or two interviewers). Focus groups were held in local venues in the early evening, with catering provided to encourage participation. All interviewees signed consent forms, and all but one agreed for the interview to be recorded. Interviewees were guaranteed anonymity, and any names used are pseudonyms. Sometimes particular details are omitted to protect the anonymity of the participant. The number, type, and location of the interviews and focus groups is detailed in Table 1.

The # codes at the end of quotations throughout the document are to indicate which category of interviewee is being quoted. For example, # young new entrant woman #3 indicates the third new entrant woman interviewee. The participants in focus groups are not individually identified with the exception of the 'Men in farming' group. This group had five participants and are distinguished by number, for example: #1 Men focus group #1; #1 Men focus group #2, etc.

Table 1 Description of interviewee and focus group participants

All interviews and focus group discussions were recorded by digital dictaphone, fully transcribed, and qualitatively analysed using MAXQDA but predominantly through repeated listening to interviews and rereading transcripts. Thematic coding was undertaken using the five subject headings that were identified by the Scottish Government, plus key themes that emerged in the research. The guides used in interviews and focus groups as well as informed consent forms are included in this report ( Appendix C).

2.2 Quantitative approaches

The quantitative elements of data collection involved two questionnaire surveys.

Main Survey

Firstly, an online questionnaire survey (using Limesurvey software) was launched on 24 ^th June 2016, following pilot testing, and remained open until 1 ^st October 2016. This questionnaire sought responses from women who live and/or work on farms in Scotland, and included questions with conditions that avoided responses from people who did not fall within these criteria. The questions included in the online questionnaire are presented in Appendix A. The link to the online questionnaire was distributed widely via social media and through direct communication by industry groups, individuals ( e.g. by passing on postcards with the survey details and web link), and through contacts of the Research Advisory Group. This questionnaire received 1,543 responses. These responses were then filtered, and 54 suspected duplicate records were removed from the main survey. Other responses removed were incomplete surveys i.e. when respondents indicated they did not identify as women, did not work on a farm etc, they were thanked for their interest in the survey and not given the opportunity to complete the remainder of the survey. In total, 1,118 responses were used within the analysis who met the following five criteria:

• lived and/or worked on farms
• identified as women
• were resident within Scotland
• gave consent to participate in the study
• were not potential duplicate records (based on information on IP address, age and education, and farm characteristics).

The total number of responses represents a substantial database on which a large number of statistically significant cohort based assessments could be made ( e.g. contrasting the responses of women of different ages, located on different sizes of farms etc). However, as an on-line survey, respondents were self-selecting, and participation was dependent upon internet access. The total dataset is therefore not necessarily representative of the total population of women living and working on farms in Scotland. We estimate that there were approximately 19,602 women working on farms in Scotland in 2016 (based on the total number of female working occupiers, regular staff, part-time staff and casual and seasonal staff). Survey respondents thus represented approximately 5.8% of this population.

Student and Alumni Survey

Secondly, a further online questionnaire survey was launched on 3 ^rd October (closed on 15 November) that sought responses from women who are current students or alumnae of agricultural courses at colleges and universities. The questions included in this second online questionnaire are presented in Appendix B. The link to this questionnaire was send by email by Scotland's Agricultural College ( SRUC) to 201 current female students and 479 alumni, and advertised in the alumni newsletters of SRUC and the University of Aberdeen. About 18% of the students responded. The link was also distributed widely on social media, through direct communication by industry groups, with individuals, and through contacts of the Research Advisory Group. This second questionnaire received 212 responses, which were filtered to exclude those which did not meet all of the following criteria:

• had not completed the main survey
• identified as women
• had completed, or were completing, formal education at college or university level relating to agriculture
• gave consent to participate in the study
• were not potential duplicate records (as per the main survey, based on IP address, age and agricultural education information, and farm description)

16 suspected duplicate records were removed from the alumni/student survey. Therefore, the data of 148 respondents in the alumni/student survey were taken forward to the analysis. As this is a smaller cohort than for the main survey, fewer statistically significant comparisons could be made (as per R Core team, 2016). There was considerable duplication in questions between the two surveys, allowing student and alumni responses to be presented alongside main survey responses.

Analysis

The analysis of these questionnaires used standard descriptive statistics and bivariate statistical tests. Some responses within the 'raw' data were logically recoded, and some new variables were created from the existing data before the statistical tests were carried out in the second part of the analysis. The following are examples:

• within the 'age' variable in the main survey, for some analyses the "18 or under" and "19 - 25" responses were combined into one category of "25 or under", and "I prefer not to say" responses were excluded from the data (recoded as ' NA'). (Similar recoding was carried out in the alumni/student survey)
• in the main survey, farm areas of "less than 2ha", "2-4.9ha" and "5-9.9ha" were recoded to a category of "less than 10ha", and the "I prefer not to say" responses were removed for some analyses
• "Strongly agree" and "Agree" responses were combined to "Agree", and "Strongly disagree" and "Disagree" responses were simplified in a similar way
• in the main survey, a variable was created to flag whether or not respondents lived within the crofting counties: defined within this research as the local authorities of Argyll and Bute, Highland, Moray, Na h-Eileanan Siar, Orkney and Shetland
• in the alumni/student survey, a variable was created to classify respondents as either alumni or students, based on responses given on course completion

From the two surveys, all variables of interest were identified. To give an overview of the characteristics of all respondents, descriptive statistics ( e.g. a breakdown of responses to questions) were produced. The total number of responses on which these figures were based depended on the structure of the questions within the online survey, and for some questions, respondents could select more than one response option. Next, for variables which were particularly important to the research questions and themes which developed from the research (three examples of these from the main survey were age, farm area and 'new entrant' status), cross-table analysis and bivariate tests ^[2] were used to assess whether these variables were significantly associated (at the 95% confidence level) with other variables (in both surveys).

Analysis of the main survey demonstrates a spread of geographical and socio-demographic coverage: respondents were from all Scotland's census regions, all age categories (from 18 and under to 65+), and all educational achievement levels. However, there does appear to be lower representation amongst older respondents: the 55 to 64 age category contained only 15% of respondents and the 65 and over category 4.7%. This is to be expected in on-line surveys (It has been well demonstrated in the academic literature, e.g. Pocewicz et al., 2012, that older people are less likely to respond to an internet-based survey). However, owing to the high number of survey respondents, it is still possible to statistically analyse for age-related differences ( i.e. the 53 women over the age of 65 who completed the survey represent a sizeable cohort).

The student and alumni survey elicited responses from 26 Scottish regions, Northern Ireland, England, and internationally, ranging in age from 25 and under (25%) to 65 and over (1.4%). Approximately 75% of respondents to the student and alumni responses were alumni.