Evalution of recored temperatures before and after

#1
Hi

I have recorded the temperature of a fluid in a container before and after heating, and want to evaluate the significance by looking at the mean differences of the temperature before and after using SAS jmp. I used 8 different samples and from them I recorded the temperature before and after.

Additional info: Moreover the containers had different sizes how can I account for that in my analysis to avoid deviation?

Thanks in advance :)

best
Zahra
 

staassis

Active Member
#2
You have too little data to study the effect of Size. Simply compare Before to After using Wilcoxon test.

If you collect more data, do the following. Regress Difference = After - Before on Size. Use bootstrap to calculate p-values of the regression coefficients. Based on the p-values, prune the model, leaving only statistically significant terms.
 
#3
Thank you very much!
but why can I not use paired t-test? I know that due to the small sample size it is difficult to assume whether the data is normal or non-normal distributed. So of course based on this, we can do a non-paramatic test, but is this the case every time we have a small sample size?
 
Last edited:

hlsmith

Less is more. Stay pure. Stay poor.
#4
Was the baseline temp the same in each sample? Was the heating period the same in each trial?

Please just post your data so we can see what you are working with. You can post real data or real data with a little noise added to deidentify if desired. I would be interested in seeing the magnitude of change we are discussing, along the other variables? What is considered your level of significance in this endeavor (i.e., alpha value)? Are you just wanting to say the post temperature is not equal to zero beyond chance? Why were the samples of different sizes?
 
#5
Was the baseline temp the same in each sample? Was the heating period the same in each trial?

Please just post your data so we can see what you are working with. You can post real data or real data with a little noise added to deidentify if desired. I would be interested in seeing the magnitude of change we are discussing, along the other variables? What is considered your level of significance in this endeavor (i.e., alpha value)? Are you just wanting to say the post temperature is not equal to zero beyond chance? Why were the samples of different sizes?
The heating period depends on the amount of volume in each bottle. We had 8 bottles with different sizes and volumes of fluid :)

Temperature before <-- x(32, 36, 41, 42, 37, 37, 38, 37)

Temperature after <-- x(32, 33, 35, 41, 38, 33, 37, 35)