Using the Washington Group Short Set in a Telephone Survey
Surveys are not always delivered face to face, and especially in times like the current Covid-19 pandemic situation, the option of delivering a survey over the phone can be attractive – both in terms of not having interviewers be a vector of disease transmission and for improving the response rate from people reticent to speak face to face with someone they do not know in the middle of a health-related lockdown. Unfortunately, we have not done field tests of the Washington Group questions specifically for administration in a phone survey, so we do not have any direct experience or evidence with how they work in that modality. However, a few issues come to mind when considering this method.
1) Sample bias. As with all telephone surveys there is a chance of sample bias because not everyone has a phone. This is more of an issue, of course, within countries that have lower rates of phone ownership. But the issue is exacerbated when it comes to disability because as we have seen in at least a couple recent national disability surveys (e.g., Thailand and Vietnam) people with disabilities have less access to phones and to the internet. This is true at the household as well as the personal level. That is -- at least in those two countries -- households with a disabled member are (a) less likely to have a phone or the internet, and (b) within households that do have access to a phone or the internet, people with disabilities in those households have less access. The latter issue is not as important if a respondent is responding for the entire household.
2) General considerations for telephone surveys. In general, simple more straightforward questions work well in phone surveys and the Washington Group questions were designed to be simple and easily understood. The Washington Group Secretariat is aware of surveys administered by phone and the web that include the WG-Short Set, WG-Extended Set and the Child Functioning Module and have not heard about any administration issues.
3) Phone interviewing people with hearing difficulties and other communication difficulties. People with hearing or communication difficulties may have more problems responding over the phone which could bias the survey results if they are not included in the survey. Therefore, there must be some way to interview deaf persons - either by using a proxy, sign interpreter, video relay service (VRS), paper questionnaire or a mechanism where voice is transformed into written communication. In these cases, the question can be asked as written with all the answer categories presented for the interviewee to choose. Another concern is that the interviewer might feel foolish asking a question that they know the answer to (asking a deaf person if they cannot hear at all or asking a person who has demonstrated hearing ability if they cannot hear at all), but there are standard techniques for dealing with these situations like this, such as allowing interviewers to tell the respondents at the beginning that they are instructed to ask all questions and read the response options. These same concerns could also apply to the communication question.
4) Stigma. In some cultural contexts a lot of stigma exists around disability. In fact, people have been known to hide household members with disability from view. This is even easier to do during a phone survey where the interviewer is not able to see any evidence of an additional person and prompt their inclusion in the survey. For that reason, it is even more important to not refer to the word “disability” in any way, and to gently prompt the respondent to make sure they are including everyone in the household.