6 The Survey: Method
6.2 Data Collection 8
As no existing measures could be found which explored the handler and guide dog partnership, a new survey instrument was constructed. A copy of the questionnaire along with the introduction and culmination to the interview appears in Appendix A.
6.2.1
Survey design
A mixed design, non-experimental approach was taken, comprising of a single observation enabling between subjects (conditions between different groups of participants) and within subjects (conditions within the same group of participants) comparison. The information collected was a mix of retrospective, current and/or prospective data depending on whether the participants (who were all either currently using and/or had previously used a dog/s) were being asked about their experiences before they used guide dogs or about their first or subsequent dog/so To illustrate, the participants were asked to remember their experiences before they acquired their first dog and when using each dog thereafter. Thus, responses to the questions pertaining to
before a dog was obtained yielded retrospective data, while responses to the time when a dog was used yielded current data if the participants were currently using a dog, and a mix of retrospective and current data for those who had used more than one dog. Prospective data was obtained from those who were currently using their first dogs and who were asked to project how they might feel at certain times, for example, at the end of the partnership.
6 -The Survey: Method The survey instrument consisted of a structured, self-report questionnaire that was delivered via telephone (78%), face-to-face interviews (20%), or a combination of both (2%). The telephone survey technique is recommended by the RNZFB as a useful means of obtaining opinions from its members, and in many surveys of exclusively face-to-face interviews rural residents are under-represented (RNZFB, 1990). Face to face interviews are also more expensive in terms of travel time and costs. Use of the telephone facilitated a nation-wide survey and, additionally, as around 12% of the RNZFB 's aDS' clients are 'deaflblind' (people who have combined hearing and sight impairments), it offered a viable means of communication with some deaflblind people, whose hearing was improved by the telephone hand-piece blocking background noise. However, two deaflblind participants asked not to be interviewed by telephone, as one was only able to lip-read, and the second preferred to communicate with the researcher via an interpreter. Some participants, who lived within a short distance of the researcher, also requested a face-to-face interview, which was granted.
Despite the lengthy nature of the questionnaires, a postal survey was disregarded for several reasons. Some participants would have had trouble in reading a regular (font size 12) or enlarged printing format, and if the questions were read out and answered by proxy, for example, by a friend or family member, this may have affected the response. The creation of Braille documents would have been prohibitively bulky and expensive, and aside from the impracticality, most New Zealanders who have a visual impairment or are blind do not read Braille.
6.2.2
Survey instrument
The questionnaire contained six sections, which are described later in this chapter. These were (a) general, (b) information on each guide dog used (c) guide dog rating, (d) mobility (e) quality of life and (f) service delivery and miscellaneous. The questionnaire, which was administered in this order, took between one and six hours to complete, depending on how many dogs each participant had used and how much they wished to discuss. Most of the sample (64%) had used a second dog, and the average interview took around 2 hours to deliver. Participants were informed that they could stop the interview at any time and postpone the remainder if they wished. Although the interviews were lengthy, most were completed in one sitting, and participants found them to be an enjoyable andlor cathartic experience. Several of the participants
continued to call the researcher after completion of the interview, as they enjoyed discussing life with their dogs.
As part of the initial planning of this study, the researcher corresponded with J. Nicholson (personal communication, October 1 998) who had investigated distress arising from the end of a partnership with a guide dog (Nicholson, 1993, Nicholson et al., 1995) regarding methodological issues when designing a study for a vulnerable population. J. Nicholson advised to put in place procedures to support a handler who might find discussing the end of a partnership distressing. Hence, provision was made for referral to an appropriate councillor if necessary. However, this option was not required.
Open-ended and forced-choice questions were used to gather the data. The inclusion of the open-ended questions obtained the participants' views of reality, thoughts, ideas and memories, rather than any preconceived notions the researcher may have. According to Reinharz (1992), this approach reduces the possibility that participants feel alienated from the research. The data obtained from open-ended questions2 was subjected to content analysis and coded into categories and themes. Forced-choice format included differential rating scales3 (continuous scale from 1 to 10 with opposite attributes at either end), Guttman scales (cumulative), ranking order, multiple choice or a simple
'yes' or a 'no' response.
6.2.3
Delivery
The questionnaire was tested for clarification of wording and concepts with four sighted Massey University students, in face-to-face and telephone interviews before being piloted on some of the participants. It was appreciated that the students were dissimilar to those in the main enquiry, and hence the exercise was of little relevance and sensitivity concerning visual impairment issues. However, this run through provided an estimate of how long the questionnaire would take to deliver and improved the researcher's interviewing skills.
2 Note: Multiple responses resulted in many percentages not adding to lOO.
3 A IO-point scale with no mid-point was chosen as scales with more than seven categories are considered to be more reliable and the lack of a mid-point negates JlOssible respondent bias towards the central option (Foddy. 1993).
6 -The Survey: Method The questionnaire was then pilot tested with the four, now familiar, participants that had attended the focus group, whose suggestions were incorporated into the measure. Once the fourth person was interviewed, the previous three were contacted and asked to respond to any additional or revised questions. Data collected from pilot tests is not commonly included in the final analysis. However, in this study data from all 50 participants was incorporated as the pilot data fitted descriptively with the rest of the dataset and was considered too valuable to disregard in terms of amount, quality, overall small sample size and length of time taken to collect. The pilot data was not analysed separately on a statistical level.
Suggestions that were incorporated included improving ambiguous wording, and instead of asking participants to rate their travel performance from their overall experience with guide dogs, they were asked to rate the dogs they considered unsatisfactory mobility aids separately from the satisfactory ones. This was due to pilot participants, who had experienced dogs that were unsatisfactory as mobility aids, not wishing to provide only one rating. Another revision that was made was the order in which participants were asked about their dogs. Initially the questions had been arranged to discuss the situation with the current, or most recently used dog first, however those who had used more than one dog found it easier to consider their earlier experiences and how these changed over the time they used their subsequent dogs.
The remaining 46 interviews were initiated by the researcher telephoning the participants, introducing herself and reiterating the purpose of the study. Participants were asked if the time and venue were suitable for them to be interviewed and if not, an appointment was made for their convenience. Participants were reminded that the information gathered would be used in such a way that they could not be personally identified (unless they specifically requested to be), and as the study was independent of the RNZFB, they may be as open as they wished. They were also reminded that they did not have to answer every question and that they could withdraw from the study at any time without giving a reason. The topics to be covered were reviewed and participants were encouraged to ask questions at any time during the interview. At the close of the interview, the participants were thanked and informed that they would receive a
summ
arl
of the results in due course, and were invited to contact the researcher if there were issues they wished to discuss further.The interviews were conducted, recorded, and the data coded into numerical categories by the researcher alone. Hence, consistent definitions of terms were provided, and probing of ambiguous or incomplete answers to clarify issues was standardised. The interviews were not audiotaped, as it was feared that a recording device might inhibit response.
6.3 Description of the Survey Instrument