0
$\begingroup$

i'm a graduate student (biology program) with minimal experience with statistics. I have two questions relating to my project that i'd love help with! The first question is how I can take a single measurement of many events (such as, 55 coin flips out of 100 yielded heads) and test if that is significantly different than a separate measurement (such as, now with a different coin, I see 40 coin flips out of 100 yields heads - so are the two coins significantly different? Can I put a p-value on that?). The second question is, if I use an R program that draws a best fit line on some data and gives i.e. the slope is 10 and the 95% confidence interval is from 8 to 12, then compare that to another slope that is 14 with a 95% CI from 12 to 16, are these results significantly different? And can I put a p-value on that as well? Thanks in advance for any help!

$\endgroup$

1 Answer 1

0
$\begingroup$

Let's do the coin flip first. Here is how to do this in R

heads<-c(55,40)
flips<-c(100,100)

prop.test(heads,flips, correct = F)

The heads array houses the number of heads we observed. The flips array houses the number of times each coin was flipped. The function prop.test performs a test of proportions. You can see the p-value in the print out.

As for the regression, I don't think you can do this without access to the data. You'd need to specify a model, and then look at the coefficients of that model. Do you have a sample dataset we can work with?

$\endgroup$
6
  • $\begingroup$ Am I missing something? OP asked about 55 vs 40 each out of 100. Then from prop.test(c(55,40),c(100,100), correct = F)$p.val, I get 0.03367207, so signif diff at 5% level. $\endgroup$
    – BruceET
    Sep 4, 2019 at 18:25
  • $\begingroup$ @BruceET Oops, must have thought OP said 50 and not 55. Is that what you're referring to? $\endgroup$ Sep 4, 2019 at 18:38
  • $\begingroup$ Yes. Thanks for edit. $\endgroup$
    – BruceET
    Sep 4, 2019 at 18:53
  • $\begingroup$ Thanks! That's perfect. $\endgroup$
    – Daniel
    Sep 4, 2019 at 22:17
  • $\begingroup$ As for the regression, I'm not sure how to present it on this site. Hearing that the statistical analysis must be done on a more upstream step is useful to know though. I'm going to hunt through the R program that I used to calculate the CIs and see if there are more comparative functions available. $\endgroup$
    – Daniel
    Sep 4, 2019 at 22:20

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.