1
$\begingroup$

I want to perform an ANOVA using R. I have three populations, represented by their respective means and SD: Pop.1: 5.5 +- 0.4 (n=100) Pop.2: 5.9 +- 0.3 (n=150) Pop.3: 6.2 +- 0.5 (n=200)

Which is the exact code using R to perform the ANOVA using exclusively these data?

Moreover, how can I perfom subsequently a post-hoc analysis?

Thanks in advance

Improve this question
$\endgroup$
5
  • $\begingroup$ What specific post-hoc comparisons do you want to make? If it's all 3 of them then your ANOVA is pointless. $\endgroup$
    – John
    Commented Jun 9, 2014 at 18:45
  • $\begingroup$ If you have population means and sds you don't need statistical inference at all - the population means plainly differ. Statistical inference (like ANOVA) is used when you want to make inferences about populations but you have samples. If you mean something different from what you asked, you'll need to edit your question. $\endgroup$
    – Glen_b
    Commented Jun 9, 2014 at 21:06
  • $\begingroup$ In the same way that a T-test can be performed by using exclusively mean, SD and n, I supposed that an ANOVA could be as well performed in someway. How about simulating those data? $\endgroup$
    – antecessor
    Commented Jun 10, 2014 at 7:53
  • $\begingroup$ Assuming that you refer to samples rather than populations, it seems that it is necessary to have the original data in order to fill the ANOVA table. A workaround probably is to create random data from the assumed underlying distributions of each samples and apply F test in these data. $\endgroup$
    – Epameinondas
    Commented Jun 10, 2014 at 7:54
  • $\begingroup$ Otherwise, John, what about a Bonferroni or Tukey tests? $\endgroup$
    – antecessor
    Commented Jun 10, 2014 at 8:13

1 Answer 1

Reset to default
1
$\begingroup$

R does basic math calculations (?Math) and you can just look up online the math to solve your problem. To get the the criterion of F value for the test statistic you can use the function qf. Post hoc tests are similar in the sense that the math and equations are easy to look up. Perhaps you should update your question with the kind of post hoc test you need and why you want that in particular.

Improve this answer
$\endgroup$
2
  • $\begingroup$ As Epaminondas say, I have 3 samples, with their means, sd and n. Which is de code to introduce in R and calculate de one-way ANOVA in the example from the top? I checked your link John, but couldn't open the programs. Thanks $\endgroup$
    – antecessor
    Commented Jun 10, 2014 at 9:50
  • $\begingroup$ @antecessor Check out the rpsychi package. $\endgroup$
    – Roland
    Commented Jun 10, 2014 at 10:59

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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