We belatedly worked out that they were just using jargon to sound technical but hadn’t actually calculated any effect sizes. Hence our annoyance – if you had wasted a lot of time trying to generate useful SDs that would approximate to anything like what they claimed were effect sizes – you might be less generously disposed. Particularly when you might strongly suspect that they didn’t realise that it is a technical term – despite all their ludicrous posturing.

This isn’t particularly coherent – I might have a bash some other time.

i can clarify that those are also definitely NOT the responses he sent to the BBC, although the responses he did send were equally long and flawed.

To be fair to Holford and co, it isn’t completely mad to look at the proportion giving a categorical rating of “very good” – similar methods have often used in pain research and other subjective fields (although you’d usually group all ‘good’ and ‘very good’).

I took a bit of a back seat for this part of the programme and it was all handled by the Executive Producer from the BBC Radio Science Unit, Rami Tzabar.

I can tell you absolutely for definite that those are NOT the questions which the BBC sent to Patrick Holford asking for a response.

And furthermore, those are NOT the answers Professor Holford sent back to the BBC.

This would seem to be further clear evidence of the spectacular unreliability of Patrick Holford’s claims.

Absolutely astonishing. There are no depths etc.

If Holford is shown to be misleading his readers about the questions he was asked, and the answers he gave, then that would be a remarkable new discovery.

This really is very funny. My professional advice to him would be to keep veeery quiet, but of course, his hole-digging is why he is such an instructive joy to watch.

So all the data analysis and pretty (if mis-leading) bar graphs in the report are based on 74 kids?!! Or, to put it another way 0.72% of the respondents.

If I was a parent of one of the 99.28% who have been ignored, I’d feel a bit miffed at having wasted my time.

0.72% in his outliers??? And that’s the basis of comparison? He is talking about p values when he should be discussing statistical noise?!?

Words fail.

If I was a parent of one of the 99.28% who have been ignored, I’d feel a bit miffed at having wasted my time.

Although, I do wonder if PH’s original hypothesis was that the vast majority of kids would be in the v. poor/poor diet category, and, when the results showed they were almost all “neutral”, faffing around with a few outliers was all he had left to “prove” his points. Hey, why change your mind when the data doesn’t support you when you can change the analysis post-hoc?

Ho Hum.