# Reporting regression

Announcements
#1
Hi,

I'm having trouble reporting regression, can someone help?

Regression analysis was used. With impulsive measures, a significant proportion of variance was attributed to age β =0.90, t=1.412, p=<0.00 R2 =0.008, F(1, 238)=1.993
0
7 years ago
#2
(Original post by Noodlzzz)
Hi,

I'm having trouble reporting regression, can someone help?

Regression analysis was used. With impulsive measures, a significant proportion of variance was attributed to age β =0.90, t=1.412, p=<0.00 R2 =0.008, F(1, 238)=1.993
Hi. If you've only got two variables, for the sake of brevity its worth doing a pearson's correlation instead. So what you'd write is, age signifantly correlated with impulsivity, r=.90, p<.001.

If your doing a multiple regression, then its slightly more complicated. F-values are only imporant if your doing a hierarchial model, and if your doing that, then we need more infomation. Also, your R^2 value should be .95 for the overall model.
0
#3
(Original post by iammichealjackson)
Hi. If you've only got two variables, for the sake of brevity its worth doing a pearson's correlation instead. So what you'd write is, age signifantly correlated with impulsivity, r=.90, p<.001.

If your doing a multiple regression, then its slightly more complicated. F-values are only imporant if your doing a hierarchial model, and if your doing that, then we need more infomation. Also, your R^2 value should be .95 for the overall model.
Great, thanks so much!

Do I still include t and beta?
0
7 years ago
#4
It depends what type of regression you have and how many variables. It sound like you are doing multiple regression? Or singular regression?
0
#5
(Original post by JamesManc)
It depends what type of regression you have and how many variables. It sound like you are doing multiple regression? Or singular regression?
singular i believe

regressing age and gender separately to impulsive scores
0
7 years ago
#6
According to my trusty textbook if it's a singular (correlational) analysis you would just write the r and p like this (r = ?, p = ?) and leave out the t, F, R2 altogether. So, with impulsive measures, a significant proportion of variance was attributed to age (r = , p = .012). That's if you're doing single regression btw, it sounds like you are from what you've said. You should never write p = 0.00 though so give the exact probability figure. Lecturers hate you giving 0 as the P and would rather p = 0.02 or whatever it is. Give the exact figure. A P of 0 means that if you did a hundred trillion replications of your experiment it would always have the same result which is unlikely. Also, for your other numbers (same for all statistical tests) give them to two decimal places, it's APA style.
0
7 years ago
#7
(Original post by Noodlzzz)
singular i believe

regressing age and gender separately to impulsive scores
Sometimes it might make sense to run correlation/regressions WITHIN genders too. I guess if your expecting an interaction effect...

I would run a correlation between impulsivity and age. A t-test between gender and impulsivity. You could run a mutliple regression with gender and age on impulsivity too, not sure what it would add.

As mentioned exact p values are better for text than <.05 or <.01, although if its below <.001 then just write that.
0
#8
(Original post by iammichealjackson)
Sometimes it might make sense to run correlation/regressions WITHIN genders too. I guess if your expecting an interaction effect...

I would run a correlation between impulsivity and age. A t-test between gender and impulsivity. You could run a mutliple regression with gender and age on impulsivity too, not sure what it would add.

As mentioned exact p values are better for text than <.05 or <.01, although if its below <.001 then just write that.
Thanks! Could you suggest what I could include in a table? The design is comparison between impuslivity and FFM vs. BPD and impulsivity with a regression of age and gender.

Thoughts?
0
7 years ago
#9
(Original post by Noodlzzz)
Thanks! Could you suggest what I could include in a table? The design is comparison between impuslivity and FFM vs. BPD and impulsivity with a regression of age and gender.

Thoughts?
Ohh so you'd do a multiple regression with two steps. report table link this:

Step 1:
Age (B, Beta, SE Beta, p)
Gender (B, Beta, SE Beta, p)

Step 2:
FFM (B, Beta, SE Beta, p)
BPD (B, Beta, SE Beta, p)

Could provide the R^2 change between step 1 and 2, with the associated F and p value. I'm not actually too sure what the norm is for reporting multiple regression in APA style, im just saying what I would do :P. I would also include 95% confidence intervals for beta actually for all of the above.
0
#10
(Original post by iammichealjackson)
Ohh so you'd do a multiple regression with two steps. report table link this:

Step 1:
Age (B, Beta, SE Beta, p)
Gender (B, Beta, SE Beta, p)

Step 2:
FFM (B, Beta, SE Beta, p)
BPD (B, Beta, SE Beta, p)

Could provide the R^2 change between step 1 and 2, with the associated F and p value. I'm not actually too sure what the norm is for reporting multiple regression in APA style, im just saying what I would do :P. I would also include 95% confidence intervals for beta actually for all of the above.
Thank you so much for this, I really appreciate it 0
X

new posts Back
to top
Latest

### Oops, nobody has postedin the last few hours.

Why not re-start the conversation?

see more

### Poll

Join the discussion

#### Do you know what you'll do if you don't get the grades you're hoping for?

Find something else in clearing (31)
27.43%
Take a gap year (17)
15.04%
Resit my exams (31)
27.43%
Look for alternate pathways to the career I want (17)
15.04%
I don't know yet (12)
10.62%
Something else (tell us in the thread) (5)
4.42%