I am working with body temperatures in R, conducting a t-test.

It is widely believed that the average body temperature for healthy adults is 98.6 degrees Fahrenheit.

The body temperatures of $n = 130$ healthy adults were measured:

$\bar{x} = 98.249$ with standard deviation $s = 0.7332$

So, the null hypothesis is $\mu = \mu_0 = 98.6$

Two-sided test: $H_a : µ \neq µ_0$, with $α = 0.05$

Critical limit (c) given by $t_{n−1,1−α/2} = t_{129,0.975} = 1.979$ This is upper limit, lower limit is then $−1.979$

(98.249-98.6) / (0.7332/sqrt(130)) 
 [1] -5.458287

The test statistic $t$ falls in the rejection region $(−5.46 < −1.98)$, so we reject $H_0$ at $α = 0.05$

However, when we run the t-test in R, the p-value is $2.411e-07$:

t.test (bodytemp $ temp, mu=98.6, alternative="two.sided") 

One Sample t-test

data: bodytemp$temp 

t = -5.4548, df = 129, p-value = 2.411e-07 

alternative hypothesis: true mean is not equal to 98.6 

95 percent confidence interval:

98.12200 98.37646 ## sample estimates: 

mean of x


Therefore, if the p-value is greater than 0.5 (given by α), we would accept H0, but we know this is not correct given the test statistic t falling in the rejection region, and the mean of x being 98.24.

SO, shouldn't the p-value be less than 0.5? Why is the p-value 2.4?

  • 2
    $\begingroup$ Please fix the significance level you talk about at the end of your question. It is NOT 0.5 $\endgroup$ – Glen_b -Reinstate Monica Jun 25 '16 at 8:09

R gives you a p-value of $2.411\times 10^{-7}$ , thus much smaller than .05. The convention of writing this as 2.411e-7 is used in many contexts in programming and in no way specific to R.

A p-value is never bigger than 1.0


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.