> What is an 'acceptable' response rate from a postal survey?
It all depends what kind of survey it is. I find most books on research
are misleading about the topic: they're written by people who've never
tried to do a postal survey, because they 'know' they don't work.
The first distinction you ought to make is between classes of respondent.
There's a substantial difference in attempting random surveys, surveys of
administrative practice, and inquiries from key respondents. I've found
them pretty helpful for making administrative enquiries. I began doing it
with housing allocations policy, where local authorities have a legal duty
to provide information to any member of the public who asks and a
non-response can also be recorded as a result. It yielded so much
information, though, that I couldn't see any reason not to try it for a
range of later work, and I've used the same approach in different kinds of
study. Overwhelmingly, by the way, simple, short, open questions yield
the most information.
The second distinction relates to qualitative and quantitative research.
Qualitative sampling is a vexed issue, but many researchers seem to settle
for convenience samples; if it's acceptable for group interviews it should
be acceptable for postal enquiries. You trade depth for a wider range of
respondents. I've undertaken several trawls for problems, asking
voluntary groups or professionals. If the aim is simply to identify
whether problems exist within a process or administrative system, the
number or proportion of people who complain is not centrally what matters,
and a postal enquiry can reach a lot of people who interviewers can't.
The main loss is the opportunity for interaction and investigation.
The third issue relates to the method of the survey. Is it
exploratory, investigative, descriptive, or something else? One of the
most famous, and successful, postal surveys was the PEP inquiry on racial
discrimination, which sent similar letters from people of apparently
different ethnic backgrounds to prospective employers. The question it
addressed was simply one of whether there was discrimination, and it
succeeded spectacularly.
I've only ever had one poor response rate, by the way. This was while I
was doing work for the local authority here and used the local authority's
address for an enquiry about poverty, instead of my university address.
The respondent agencies saw a local authority logo and didn't bother.
Paul Spicker
University of Dundee
Dundee DD1 4HN
Scotland
Tel: + (44) 1382-344929 (direct line)
+ (44) 1382-344656 (office)
Fax: + (44) 1382-344675
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|