I'm sitting in a hotel bar at 3pm trying to work out what's wrong with the respondent across the table from me. She was down on my list under 'physical disability', but for the life of me I can't see anything there. Her choice of spectacle frames may be questionable, but I'm almost certain that doesn't count.

With the exception of a few specialists, I suspect that most of us encounter a disabled respondent once in a blue moon. A recent project about disabled workers for Manpower, as part of that company's activities for the European Year for People With Disabilities 2003, did, however, cause me to question our attitudes to disability in mainstream qual. With one in ten people in the UK disabled and an EU total of 37 million, how come we see so few disabled respondents?

It is, perhaps, to be expected. After all, it is in the nature of qualitative project design to be selective; we seldom need to reflect the wider population in any systematic sense. Within groups, we often ask recruiters to seek a degree of homogeneity, to help the respondents gel with each other and be more productive. And there will be good practical reasons to exclude some disabled people: for example, those hearing-impaired people who struggle to follow conversation in group situations.

So why include disabled people at all, if it's unnecessary for the project ­ and may even jeopardise group dynamics? It's not as if they have a fundamental human right to spend two hours in a front room in Birmingham talking about epoxy adhesive brands.

Most of us do want society to be more inclusive and fair towards disabled people. Qualitative research, of all professions, ought to be ahead of the game. If so, perhaps we have to address the disabled-free environments we unwittingly create. The question is, does this mean exercising more social responsibility in our recruitment?

In a way, the principle has already been conceded: we often do seek a fair representation of minority ethnic groups, for example. At some level, we have accepted that such inclusiveness is 'a good thing', even if it might have been easier to conduct the research without it. Not so, it seems, with disability ­ though, admittedly, these barriers are of a different order entirely.

My concern is that we end up excluding some disabled people for no better reason than our assumptions ­ often ill conceived ­ about what they can or cannot do. Why not have a visually impaired respondent in a group, if the subject matter allows? You can read out the stimulus; and a lot of non-disabled respondents forget their glasses and can't see it anyway. Not that we're guilty of making the average focus group some kind of eugenicist dystopia full of physically fit, highly socialised specimens ­ not any I've moderated, at least. But it struck me how often in our industry we just assume that all the respondents ­ and clients ­ are non-disabled. How many viewing facilities require them to negotiate flights of stairs? How many recruiters' homes have a disabled loo?

Another aspect of the disability agenda will be with us soon, like it or not. Already, UK companies with 15 employees or more have a duty to make 'reasonable adjustments' to the workplace to accommodate disabled people. However, by the end of 2006 this will apply to all companies, whatever the size.

What to do? Draw up a set of guidelines? That would be a start. An interesting suggestion from our respondents was for employers to appoint disability champions ­ in the same way that one of the team may be designated a first-aider.

And the respondent in the hotel bar? After a brain tumour a few years ago, she relies on pills to replace essential hormones her body no longer produces; the surgery left her for some months unable to move parts of her face, though this has now recovered; she lost her sight in one eye. She said the experience fundamentally altered her approach to life. But you would never have known.