PhD Postgraduate Forum - My supervisor said this is correct

PhD Discussion Forum

The following thread is brought to you by our sister Web site PostgraduateForum.com. If you wish to reply or post your own thread, you will be redirected to this site.

Author	Message
Sharn773 [Registered User] 22 December 2020 10:06	My supervisor said this is correct - is it? User: Sharn773 - 22 December 2020 10:06 we performed one-way ANOVAs with test order as the dependent variable and group as fixed factor. There was no significant interaction for the modified test, F(2,102) = .82, p=.443, η2p = .016 or the EG test, F(2,102) = .29, p=.751, η2p =.006.
rewt [Registered User] 31 December 2020 21:41	User: rewt - 31 December 2020 21:41 I think that makes sense. For anova's, a p value of less than 0.5 usually indicates no correlation, 0.5-0.7 is moderate correlation and above 0.7 is strong correlation. So there is no strong interaction however I do not know what an EG test is.
abababa [Registered User] 06 January 2021 00:00	User: abababa - 06 January 2021 00:00 Be wary of asking on internet forums for a correct statistical analysis. The root here, is that you give the impression you don't understand your own stats, and are attempting to 'blame' the supervisor for okaying them if they're wrong. This will not help you get a PhD. To correctly infer if the numbers are entirely correct, you'd need the raw data and method. It is your job as a PhD candidate to understand this - it is completely reasonable if you consult a statistician - but not an internet forum. It will not hold up in viva if you say 'PhD forums said it was right' or 'supervisor said it was right' - you need to understand and answer this question yourself.
bob86 [Registered User] 06 January 2021 19:10	User: bob86 - 06 January 2021 19:10 I'd eco what abababa says here. Even if statistics is not your strong point/emphasis of your PhD, you really need to take the initiative to understand and solve the problem (validate the data) by yourself. By all means consult a specialist. But fundamentally you must go through the process as by doing so you will gain insight into the problem/issue which will only help you when you come to discuss the data/results later in your thesis. Getting out of your comfort zone is part of the PhD process and attempting to justify decisions/steps that you've taken throughout with ``My supervisor said so...'' will only get you into bother at the viva stage. Good luck.

Author

Message

Sharn773

[Registered User]
22 December 2020 10:06

My supervisor said this is correct - is it?

User: Sharn773 - 22 December 2020 10:06

we performed one-way ANOVAs with test order as the dependent variable and group as fixed factor. There was no significant interaction for the modified test, F(2,102) = .82, p=.443, η2p = .016 or the EG test, F(2,102) = .29, p=.751, η2p =.006.

rewt

[Registered User]
31 December 2020 21:41

User: rewt - 31 December 2020 21:41

I think that makes sense. For anova's, a p value of less than 0.5 usually indicates no correlation, 0.5-0.7 is moderate correlation and above 0.7 is strong correlation. So there is no strong interaction however I do not know what an EG test is.

abababa

[Registered User]
06 January 2021 00:00

User: abababa - 06 January 2021 00:00

Be wary of asking on internet forums for a correct statistical analysis.

The root here, is that you give the impression you don't understand your own stats, and are attempting to 'blame' the supervisor for okaying them if they're wrong. This will not help you get a PhD.

To correctly infer if the numbers are entirely correct, you'd need the raw data and method. It is your job as a PhD candidate to understand this - it is completely reasonable if you consult a statistician - but not an internet forum. It will not hold up in viva if you say 'PhD forums said it was right' or 'supervisor said it was right' - you need to understand and answer this question yourself.

bob86

[Registered User]
06 January 2021 19:10

User: bob86 - 06 January 2021 19:10

I'd eco what abababa says here. Even if statistics is not your strong point/emphasis of your PhD, you really need to take the initiative to understand and solve the problem (validate the data) by yourself. By all means consult a specialist. But fundamentally you must go through the process as by doing so you will gain insight into the problem/issue which will only help you when you come to discuss the data/results later in your thesis. Getting out of your comfort zone is part of the PhD process and attempting to justify decisions/steps that you've taken throughout with ``My supervisor said so...'' will only get you into bother at the viva stage. Good luck.