All was fine - you were all right! I'm still learning these things! Nerves now 100% settled. :-)

It's not needy that you want to improve at what you do. I have had a similar situation throughout my PhD. It could be one of the following:
- Poor supervisors who are not able to offer you critical feedback on your work, although there are areas where it could be improved
- Poor supervisors who have the ability to give you critical feedback, but who aren't engaging in your work and providing it
- Your work is very good and there is very little that your supervisors can comment on for it to be improved

It could be a mixture of all three. If you realise that it is the first or second one, it might be worth changing supervisors... after all, you aren't going to develop from having your typos pointed out. One good thing you can do is get submitting your work to journals asap - sent it to high up ones in your field and get the reviewers' feedback. I had a rejection from a great journal, and the comments from one of the reviewers were like gold-dust to me. They actually caused me to think about my work at a higher level, which is what I believe good feedback should do. And normally, even if your work is to a very high standard, there is some scope for constructive feedback - even if it is perhaps comments on how you can expand your thinking in a future study. But the person supervising you has to have the ability to think at a different level too, and be willing to invest and develop you in that way.

I hope this helps. Don't settle for less even though it's great that your work is already good!

Just to follow up on my original question (in case anyone encounters it in the future and it is helpful) - I decided to just show the untransformed values in the models because the differences were so minimal and it wasn't worth the trade off of the results being super difficult to interpret vs slightly better accuracy. I have noted simply noted where the differences occurred with the transformed data and what they were.

I've transformed a few DVs for my regression models and all is cool - the residuals now look nice and normal and homoscedastic etc (before they looked less good because the variables were quite positively skewed). But... now I'm wondering how on earth to interpret the betas.. normally for one unit increase in your IV there is an increase of X amount (beta) in the DV. But how would one interpret this when the betas are based on log and square root transformed variables?

I'm half thinking it is better to just go with the slightly biased non-transformed data. The pattern of results is quite similar, and at least then the interpretation is dead straight forward!

Any advice appreciated!

Cheers. I've sent the email now. It should be fine. The results aren't actually that different (just a bit of a shocker when you realise you should have done something but didn't!)

Phew - thanks. That is half my nerves settled (I have no idea how I actually slept last night).Now it's about having the balls to tell my mentor (at least as I see her) and co-author on this that I got it wrong / didn't do it thoroughly enough (i.e., checking out the models and trying other things) and here are the new results...

Makes sense. What about the conference abstract? They're not usually published at this conference, but this year it says they'll be made available in the online program.

I'm thinking I might have to withdraw it, if they won't let me change it.

Hi all,

I recently submitted a conference abstract (which can't be changed once submitted), but I have done some further analyses using a different statistical model that yields some quite different (but more valid) results.

Any tips on what is best to do / anyone know how this kind of thing works?

Thanks!

Ha! Can just imagine it! :-D

Could you visit the two cities? You might just way prefer the feeling of one of them, and kinda see yourself living there! All best!

He sounds like a creepy specimen! Just a bit of advice... because the things he is doing that make you uncomfortable are quite subtle, you may not be able to point them out to him directly. He can easily deny anything and make it seem like you are the one with a problem.

So - another way to address this is to change your behaviour - e.g., stop smiling, stop being friendly, become ultra professional, and avoid him at any cost. When he says something like about saying hello because it makes his day, just pull a face that says "this is pathetic / I have no time for this", and whatever you do, don't smile. When he texts you, do not reply. Tell him that email is the preferred method of contact - and stick to that. Delete his number and never reply to a text (if he texts about something related to your work then just reply by email).

Actually, I think that just doing the above might be enough to get the message through that you are not interested/going to tolerate it. It may be more effective (and easier) than trying to have a conversation about it.

If you do change your behaviour in this way, then don't be surprised if he asks you about it (e.g., "what's wrong?" "have I upset you?" "I liked it when..."). But you can be prepared for that. Again, no smiling or unnecessary friendliness. It might sound harsh, but those things are just nonverbal behaviours that can encourage unwanted attentions.

If you are anything like me, you might have to practice not smiling, as it can be a natural thing to do when communicating. But it truly does send a strong message if you can just look busy, frowning, looking away and glancing at the door or your watch, and like you are not at all interested in his flirtation/overtures. He will soon find someone else to satisfy his boredom/fantasies!

Good luck! :-)

And somehow, that helps! (not trying to be dramatic - just ran out of space)