I made the mistake of presuming the GP was contributing a substantial comment. This was not the case. In that case, he was not trying to pretend that three singular cases were representative of a trend.

> False data is still data isn't it.

No, it is not. Data has to come from reality. False data does not.

> Many fields use self-reports as a central part of their data gathering.

Yup. And then they recognize the potential errors this can introduce and have systematized ways of reducing them. For instance, surveys and questionnaires have to be carefully designed so that the self-report actually reports what we want them to report, and can be usefully synthesized into numerical data. Interviews are distilled into impressions and discussed or replayed with a colleague to mitigate bias. Repeatability makes sampling errors harder to remain hidden. Methodologies are written up with exacting detail so that they can be scrutinized and criticized when there's any doubt.

Over and over, they make up for the problem of anecdote and readily admit that their data can be faulty if an assumption is not recognized and accounted for.

> The worth of particular data needs to be understood, for sure, but this chastising of people presenting anecdotal information needs to be reined in IMO.

I am going to chastise people who pretend that their single cases have wider implications than the specific case they cite. If you'd like to rein me in, then perhaps I should start subscribing to the GP's paranoid fantasies of an Orwellian thought police.