Allied health professions LifecurveTM survey: report on respondents with communication support needs

4. Data and Analysis Limitations

This analysis was undertaken during the period January – April 2019, when an intern from the Scottish Graduate School of Social Science was available to undertake a three-month internship for SG's Assisted Communication policy team. In order to take advantage of her availability, and to accelerate the provision of the survey findings to the policy team, this analysis was completed before the definitive cleaning and analysis of the whole dataset. The findings are therefore provisional and may be superseded by later published reports from the main survey data controllers. This analysis does provide an early insight into the profile of people with communication support needs who are seeing AHPs across Scotland.

A number of other specific data and analysis limitations are shown below.

a) The survey did not set out to reach a random sample of the population, but was completed by people who happened to see an AHP on the day of the survey, and whom the AHP considered to be typical clients. For this reason, and because of the provisional nature of the data analysed as discussed above, it has not been weighted or otherwise adjusted, and the results have not been tested statistically.
b) The questionnaires were completed in written format and there was some variation in the information provided and the completeness of responses received.
c) Each survey response relates to one AHP appointment only, and does not capture information about the respondent's other appointments with AHPs or other health and social care needs.
d) Data for respondents who answered 'yes' to the communication question has been 'cleaned': for example, obvious errors such as postcodes without space have been amended (e.g. EH13DG changed to EH1 3DG). However data for the other survey respondents has not been cleaned, so comparisons with that group are provisional but are provided for information.
e) Some data which was submitted late has not yet been incorporated into the dataset. While it is considered unlikely that the profile of the missing data will vary significantly from the data which has been analysed, this cannot be ruled out.
f) Information about respondent gender was not requested, so male vs female breakdowns are not possible.