The FastDay Forum

I haven't followed your entire conversation far too nerdy, but are you considering age as a significant factor?

I haven't followed your entire conversation far too nerdy, but are you considering age as a significant factor?

We're collecting the data, the analysis will say if age is significant or not.

[carorees]

Ok so is the questionnaire good to go?

PhilT, am I right that I include the dropbox link and people download the questionnaire and then email it to the google docs address?

What is the consensus on inclusion criteria?

* On diet for more than 1 month
* Weight loss either <1.36 kg/month or >2.86 kg/month (women) or <2.0 kg/month or >5.22 kg/month (men). This is based on the 25th and 75th centiles in the group who submitted their data in last month's weigh in.
* OR we could go for those whose weight loss is outside 1SD of the mean from last month's weight in data which would be for women <0.76 kg/month or >3.8 kg/month and for men <1.21kg/month or >5.94 kg/month so we would get a much smaller study population.
* OR we can let anyone who likes fill in the questionnaire and then choose our study populations afterwards depending on the stats we do?
* 5:2 only or all diets?

Please give me your thoughts on the above and we can get this study underway!

PhilT, am I right that I include the dropbox link and people download the questionnaire and then email it to the google docs address?

Sounds good, if you email it to the address I can put up a Google drive download link too.

Can I ask that you have a plain text version for those without proprietary software, and that word versions are .doc files and not .docx (Save As... Word 97/2003 Format) for those with older versions of word or compatible products.

On inclusion, I would offer it to all men on account of small numbers.

I'm no stats guru, but selecting on results has some issues. Personally I would be ok with the upper and lower quartile and looking for differences in those populations that are statistically significant.

If we want a regression based approach we should select randomly as weight loss is an outcome variable rather than a subject characteristic - obese people would be a more logical group than biggest losers, for example.

In the end it comes down in part to workload for translating the responses into data. As a first shot the upper / lower quartile is good for me but we can invite anyone else to submit with the rider that we may not analyse it in the first pass ?

[carorees]

As a first shot the upper / lower quartile is good for me but we can invite anyone else to submit with the rider that we may not analyse it in the first pass ?

Good idea...let's do that.

I'll put a link to a plain text version of the questionnaire as well as a .doc version and email both to the google docs address.

Can you PM me about how I would access the responses? (Can you tell I've never used google docs?!)

I'll pm you with the draft invite thread...

What is the consensus on inclusion criteria?

* On diet for more than 1 month
* Weight loss either <1.36 kg/month or >2.86 kg/month (women) or <2.0 kg/month or >5.22 kg/month (men). This is based on the 25th and 75th centiles in the group who submitted their data in last month's weigh in.
* OR we could go for those whose weight loss is outside 1SD of the mean from last month's weight in data which would be for women <0.76 kg/month or >3.8 kg/month and for men <1.21kg/month or >5.94 kg/month so we would get a much smaller study population.
* OR we can let anyone who likes fill in the questionnaire and then choose our study populations afterwards depending on the stats we do?
* 5:2 only or all diets?

Please give me your thoughts on the above and we can get this study underway!

Hi,

Sorry, I've been very busy with work and fallen behind on this discussion.

Be it quartiles or SDs, using arbitrary inclusion criteria causes misleading and skewed (and biased) data. I suggest anyone who is interested in participating should be allowed. As long as the variables are there, a proper analysis will reveal the significant predictors of higher weight loss anyway.

Thanks for all your efforts.

Best,
e.