Skip to content

simple Q: required sample size & usage of power.t.test()

2 messages · osiander@24on.cc, Peter Dalgaard

Original
osiander@24on.cc
# 
Hi community,

Seems I am too stupid:

I try to calculate with power.t.test() an example from the statistics textbook [Jerrold H. Zar, Biostatistical Analysis, 2nd edition, Prentice Hall, 1984, Chapter 8, Exercise 8.4, (page 120)]:

-------
A sample of size 18 has a mean of 13.55cm and a variance of 6.4512cm^2.
(a) Calculate the 95% confidence interval for the population mean.
(b) How large a sample would have to be taken from this population to estimate µ to within 1.00cm, 
 with 95% confidence?
(c) 2.00cm with 95% confidence?
(d) 2.00cm with 99% confidence?

Result should be - according to the answers appendix - (a) 13.55 +- 1.26cm (b) n=29, (c) n=10, (d) n=11
-------

Here my questions:

(1) which functions in R can be used to solve (a)
(2) what has to be the syntax to answer (b), (c) and (d) ?


Thanks in advance,

Osiander
Peter Dalgaard
# 
osiander at 24on.cc writes:
Ah, this is not a standard sample size calculation, but a pedagogical
exercise. t tests and power.t.test are irrelevant for this.
Hmmm, now are you trying to learn R for yourself by trying it on some
textbook examples, or are you trying to get us to help with your
homework?

Either way: You'll need the formula for the confidence interval
(presumably a couple of pages earlier in the book). The qnorm()
function gives you quantiles of the Normal distibution, the rest
should be basic algebra.