How could I get the P-value from the mean, SEM, and n of different groups in the t-test without the original data?
For detail, I would like to know the P-value between BMP-2: 50ng/ml group and the control group.
Thanks a lot.
Hello Wei Lu,
1. Assuming these are independent groups (+ serum, - serum), compute the t-statistic as follows:
t = (M1 - M2) / SEp
where: M1 = mean for first group (+ serum); M2 = mean for second group
SEp = pooled standard error of the difference in means (see below)
SEp = sqrt(Vp * (1/n1 + 1/n2))
where: Vp = Pooled variance, n1 = N for first group; n2 = N for second group
Vp = [(n1 - 1) * SE1^2 + (n2 - 1) * SE2^2] / (n1 + n2 - 2)
where: SE1 = SEM for first group; SE2 = SEM for second group
2. Probability of t-statistic can be elicited from many statistical packages, or from spreadsheets (like Excel), and online calculators. Here is the command in Excel to give prob > |t| (e.g., as for a nondirectional test):
=T.DIST.2T(t-value, df)
t-value is the unsigned t-statistic; df = n1 + n2 - 2
Good luck with your work; I hope this isn't a homework problem!
Hello Wei Lu,
1. Assuming these are independent groups (+ serum, - serum), compute the t-statistic as follows:
t = (M1 - M2) / SEp
where: M1 = mean for first group (+ serum); M2 = mean for second group
SEp = pooled standard error of the difference in means (see below)
SEp = sqrt(Vp * (1/n1 + 1/n2))
where: Vp = Pooled variance, n1 = N for first group; n2 = N for second group
Vp = [(n1 - 1) * SE1^2 + (n2 - 1) * SE2^2] / (n1 + n2 - 2)
where: SE1 = SEM for first group; SE2 = SEM for second group
2. Probability of t-statistic can be elicited from many statistical packages, or from spreadsheets (like Excel), and online calculators. Here is the command in Excel to give prob > |t| (e.g., as for a nondirectional test):
=T.DIST.2T(t-value, df)
t-value is the unsigned t-statistic; df = n1 + n2 - 2
Good luck with your work; I hope this isn't a homework problem!
Hi, David Morse, thank you for your kind explanation.
I got this question from a published paper in which the author just showed the P-values of comparison between each group (just serum+) and the group of "100ng/ml( serum+)". (showed in the figure of the table)
But I need the results of each group (just serum+) with the control group (serum+). (individual groups)
At first, I found SPSS could do the calculation of P-value which needs n, mean, and SD. But the author listed the n, mean, and SEM.
Then I got your help here and I will try the method you provided. And I really appreciate your help. Thanks again.
You need to explain the p value for what, is it for the correlation or comparison.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Structural Equation Modeling