If you are reporting stats in a paper - for publication - would you bother putting in info about normality, sphericity, homogeneity i.e. the assumptions of tests, or would you just say "all assumptions were met"??

as in putting in my own stats in a paper i was writing? In my field it tends to be just 'samples were tested for normality using (whatever test) and if they are transformed or not according to what subsequent tests will be'. But if I was just reporting stats I read somewhere I wouldn't include normality stuff unless I was disputing them really...

yeah, this is for my own work. For my thesis I've gone into all the tests, but I guess for the paper(publication) I should just give a very brief sentence to indicate I bothered to look at them.

Most of the time they are not mentioned at all, papers won't even mention if they tested diagnostics it is just assumed (all though I suspect many people don't pay the assumptions as must attention as they should). I've never mentioned it in a paper, but have it all in lots of detail in the thesis.
:-)

I'm a bit confused now, I've decided to leave them in for now - I'll take them out if the word count is too long. I would ask my sup but she knows nothing of stats and would probably ask me to do a paragraph explaining "a t-test compares two means" etc etc if she had the chance

well if its for a paper, just look at what the other papers in your field do when reporting stats, and look especially at papers in your target journal. If they have it all in then leave it in, if they don't then just stick in a line saying you tested for normality and assumptions were met. I think loads of papers don't bother with it cos they don't know about it! I know of several student who've recently passed vivas with stats which were wrong or inappropriate, and it wasn't picked up on.

problem is, some of them weren't met and then I justify why I went off and did the test anyway :p

lol! You could try, all samples were tested for normality and t-tests were used... :D just ignore the outcome of the normality test! :) or if you want to stick in why you justify using the test then go for it, it's whatever you think is most relevant i suppose!