Hello. I got the coverage factors by calculating the the Effective Degree of Freedom from Welcsh-Satterthwaite Formula and applying the result to Student's T-distribution Table to get the coverage factor.
For resistance my coverage factor is 2.776 at 95% confidence level because the result for Effective Degree of Freedom is 4.
For voltage my coverage factor is 2.571 at 95% confidence level because the result for Effective Degree of Freedom is 5 and for current is 3.182 because the Effective Degree of Freedom is 3.
Goodness knows I'm no expert, which is why I paid handsomely for software written by an expert to get expert results--well, I like the illusion of it, anyway . . .
Short version: You do not want those factors. You only want the sample standard deviations.
Long-winded version: You have taken a sample of size X to get a measurement average, 3,4 or 5 and so on. The df is n-1 for a sample average, so a sample of 5 has a df of 4. If I had samples describing a linear curve, I might have 10 points, and since the curve fit requires zero and slope, there are n-2 df; df=8, and so on. I'm not sure we have to get Mr.'s Welch and Satterthwaite out of bed for this. Somebody tell me if this is incorrect.
Your t-table will give the factors you have for the df you stated. Those factors multiplied by the standard deviation of the sample will give you the error limits for the sample average at the stated confidence, and perhaps you could use this if you were to compare different samples with each other. That and all the other characterizations like kurtosis and halitosis and skewness and so on just describe the quality of the sample or some-such. But I think this is the wrong path for us.
The standard deviation of the sample
is the standard uncertainty we want. This is the value we will use to combine with the others. Remember, we are combining standard uncertainties. Then I think we will use W-S to estimate the combined df from all of the budget elements.