Dr. Todd Daniel, this video is incredible. Thanks a lot! The video helped me to tackle this issue I was experiencing for preparing a class. I just found out your channel and now I it will be my top 1 for stats classes. Thanks for your contribution.
Glad it was helpful! The question about negative t-values comes up frequently, so I thought it deserved a video. Just let the numbers explain themselves. Thanks for the comment!
Dear Dr Daniel Thanks for this great video🙏🏻 When we are going to decide about a null hypothesis vs. an alternative one through a lower-tailed test , if calculated t-value is less than critical t which leads to consequent rejection of the null hypothesis, then " switching" of the means can potentially result in a t value which is beyond the rejection area toward the mean and so fails to reject the null. How would we look at these different end results ? Could you please clarify this for me? Many Thanks
Good day, Doctor Todd. Where did you get the Cohen's d values when in the first set of tables of paired t-test, the Cohen's d table is nowhere to be found?
Hi, I am calculating paired sample T-test for Pre-arrival awareness with post-arrival challenges i.e., shortfall in awareness increase the challenges. It is calculated for various paired for e.g., Jobs, living cost, assessments, accommodation,... When I ran the paired sample T-test, for most of variable, mean was positive which means pre-arrival awareness information is more than post-arrival challenges with t level positive However for couple of variable i.e., Jobs and assessment, means was negative which shows pre-arrival awareness information is less than post-arrival challenges and t was negative. Out of Jobs and assessment, p value of job was significant i.e., < 0.05 So, how to interpretate and report the data?
Typically, with a paired sample t test, you subtract PRE minus POST. A positive difference means your PRE was higher, resulting in a positive t and d. However, you said the MEAN was positive, which is expected and tells you nothing about MEAN DIFFERENCE. But yes, if your, t, d, and mean difference were positive, then pre-arrival was higher than post.
The IV (grouping variable) should be 1 & 0, or 1 & 2. Yes, you COULD code with negative values, but it is unnecessary. If your DV has negative values and they are legitimate values, then yes, you can use an independent samples t test.
A negative Cohen's d does not mean that the effect is "less than zero". Interpret the same way as a correlation...an effect of -.8 is the same size as +.8, both are very strong. Good luck
When using one-tail test, then it depends on whether you were looking in the direction for increase or decrease. If you were looking for an increase and the TValue is negative, then your test is non-significant. If you were looking for a decrease and the T value is negative then you have a change in the right direction. You need to make sure that your T value exceeds the critical value. This one might help: ruclips.net/video/XZ_CD0Yz6l4/видео.html
Yes, exactly...the negative sign tells you the first group mean was lower, which is the direction of change. But if you switched the order of the groups, the magnitude would be the same and the sign would change
You just saved me f4om an embarrassing performance in my Biostat module!
Dr. Todd Daniel, this video is incredible. Thanks a lot! The video helped me to tackle this issue I was experiencing for preparing a class. I just found out your channel and now I it will be my top 1 for stats classes. Thanks for your contribution.
Glad it was helpful! The question about negative t-values comes up frequently, so I thought it deserved a video. Just let the numbers explain themselves. Thanks for the comment!
@@ResearchByDesign Indeed, I am teaching geography students t tests and it is in fact a very common question I get
A negative effect size has been significantly meaningful for me since I watched the video :) Thank you :))
Wonderful! Glad that you learned something valuable.
Your videos are saving my bachelor dissertation. Thank you!
I'm so glad! Wish you the best
I could cry you made perfect sense :D
Clear explanation. Thanks
what a great and informative vid! thanks a lot and greetings from germany! :)
Thanks. Very well-expalined!
Awesome, thanks for sharing.
Thank you so much
Thank you so much!
that helps a lot, thanks Doc
Happy to help
so cute! thanks man!
Thank you! 🙏❤️
You are so welcome
Dear Dr Daniel
Thanks for this great video🙏🏻
When we are going to decide about a null hypothesis vs. an alternative one through a lower-tailed test , if calculated t-value is less than critical t which leads to consequent rejection of the null hypothesis, then " switching" of the means can potentially result in a t value which is beyond the rejection area toward the mean and so fails to reject the null. How would we look at these different end results ?
Could you please clarify this for me?
Many Thanks
Good day, Doctor Todd. Where did you get the Cohen's d values when in the first set of tables of paired t-test, the Cohen's d table is nowhere to be found?
thanks dude
Hi, I am calculating paired sample T-test for Pre-arrival awareness with post-arrival challenges i.e., shortfall in awareness increase the challenges. It is calculated for various paired for e.g., Jobs, living cost, assessments, accommodation,...
When I ran the paired sample T-test, for most of variable, mean was positive which means pre-arrival awareness information is more than post-arrival challenges with t level positive
However for couple of variable i.e., Jobs and assessment, means was negative which shows pre-arrival awareness information is less than post-arrival challenges and t was negative. Out of Jobs and assessment, p value of job was significant i.e., < 0.05
So, how to interpretate and report the data?
Typically, with a paired sample t test, you subtract PRE minus POST. A positive difference means your PRE was higher, resulting in a positive t and d. However, you said the MEAN was positive, which is expected and tells you nothing about MEAN DIFFERENCE. But yes, if your, t, d, and mean difference were positive, then pre-arrival was higher than post.
veru informative and helpful.
can one run a model using independent variables which has negative values?
The IV (grouping variable) should be 1 & 0, or 1 & 2. Yes, you COULD code with negative values, but it is unnecessary. If your DV has negative values and they are legitimate values, then yes, you can use an independent samples t test.
Man ❤️😘
so negative cohens d in paired t test does not mean a small effect?
A negative Cohen's d does not mean that the effect is "less than zero". Interpret the same way as a correlation...an effect of -.8 is the same size as +.8, both are very strong. Good luck
What to do when the negative value was using one tailed?
When using one-tail test, then it depends on whether you were looking in the direction for increase or decrease. If you were looking for an increase and the TValue is negative, then your test is non-significant. If you were looking for a decrease and the T value is negative then you have a change in the right direction. You need to make sure that your T value exceeds the critical value. This one might help: ruclips.net/video/XZ_CD0Yz6l4/видео.html
But the negative sign shouldn't be interpreted as the direction of the effect?
Yes, exactly...the negative sign tells you the first group mean was lower, which is the direction of change. But if you switched the order of the groups, the magnitude would be the same and the sign would change
@@ResearchByDesign tks professor!
AM ASKING FOR YOUR EMAIL DR?