# Thread: SPSS - Regression - interpretation and generalizing results

1. ## SPSS - Regression - interpretation and generalizing results

Hello,

Im really grateful you have taken the time to read my post. Thanks for your help.

Im doing a regression analysis trying to predict time to recovery from a treatment that causes transient pain. Duration of time in seconds spent in pain influences time to recovery. The result from regression is:

B = 0.236
Std Error = 0.098
Beta = 0.187
t = 2.403
Sig = 0.018

So, eac additional second in pain increased time to recovery by approximately 0.24 seconds.

And if thats completely accurate then how does one calculate the effect of 30 seconds in pain compared with 10 seconds in pain? The ideal outcome would be a statement similiair to:

We would expect the recovery of someone that experiences 30 seconds of pain to be approximately X seconds longer than someone that experienced 20 seconds of pain.

2. ## Re: SPSS - Regression - interpretation and generalizing results

Hello Adam: In your regression output you should have two betas. Beta0 is the intercept and beta1 is the slope. I'm assuming you have one independent variable (duration of pain in seconds) and one response variable (duration of recovery in seconds). If so, you can construct a linear regression equation:

Yi(predict) = beta0 + beta1*Xi , where Xi is your independent variable value

Then you can find the predicted value by entering your desired Xi into the equation, such as Yi(predict) = beta0 + beta1*(30) or Yi(predict) = beta0 + beta1*(20).

The values that you use to predict should be within the range of X values from your data. For example, if your data X-values only range from 1 to 10 seconds, your predictive model will have more error if you select an X-value outside that range. It could be that the relationship between pain duration and recovery duration is not linear outside your data range.

Your output should show a test of whether your beta1 slope is significant. This is a hypothesis test for slope=0 vs. slope not=0. If your beta1 p-value is not significant (>0.05), then there is not strong evidence, based on your sample data, that a linear relationship exists between your two variables. There are other things to consider to check if you have a good model, but I'll stop here. I hope this helps!

Steve

3. ## Re: SPSS - Regression - interpretation and generalizing results

Hi Steve,

Thanks for the help with that. It was very helpful.

best wishes,

 Tweet

#### Posting Permissions

• You may not post new threads
• You may not post replies
• You may not post attachments
• You may not edit your posts