Bioequivalence and Bioavailability Forum • non-central t, power, R:

non-central t, power, R: [🇷 for BE/BA]

posted by Helmut – Vienna, Austria, 2009-10-03 23:45 (6097 d 17:41 ago) – Posting: # 4292
Views: 45,886

Hi ElMaestro,

very interesting question! I'm too short in time to read Owen's paper again (and likely too stupid to understand it as well). nQuery Advisor 7 (ASA184, I guess) comes up with
The computed power was too low for the approximation to be considered reliable. Use a larger sample size.
The lowest sample size without an error message was 40 (power 5.871678%, more than 10times the value R comes up with)...

Have you considered contacting Russell Lenth (author of ASA243 and the VBA-routines in FARTSSIE)? He writes:

Accuracy

Formula accuracy
Most computations are "exact" in the sense that they are based on exact formulas for sample size, power, etc.
Machine accuracy
Even with exact formulas, computed values are inexact, as are all double-precision floating-point computations. Many computations (especially noncentral distributions) require summing one or more series, and there is a serious tradeoff between speed and accuracy. The error bound set for cdfs is 1E-8 or smaller, and for quantiles the bound is 1E-6. Actual errors can be much larger due to accumulated errors or other reasons. Quantiles, for example, are computed by numerically solving an equation involving the cdf; thus, in extreme cases, a small error in the cdf can create a large error in the quantile.
[…] in the case of quantile computations, no warning message is generated for extreme quantiles. If you want a power of .9999 at alpha=.0001, you can expect the computed sample size to not be accurate to the nearest integer! If you specify reasonable criteria, the answers will be pretty reliable.

—
Dif-tor heh smusma 🖖🏼 Довге життя Україна!
Helmut Schütz

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes

Complete thread:

non-central t, power, R: ElMaestro 2009-10-03 19:05 [🇷 for BE/BA]
- non-central t, power, R:Helmut 2009-10-03 21:45
  - non-central t, power, R: ElMaestro 2009-10-04 21:51
    - non-central t, power, R: Helmut 2009-10-04 23:58
- non-central t is not the question d_labes 2009-10-05 12:58
  - SAS -> R -> SAS Helmut 2009-10-05 14:53
    - Lost in translation d_labes 2009-10-06 14:31
      - Lost in translation Helmut 2009-10-06 14:42
        
        Macacus cynocephalus d_labes 2009-10-06 15:14
        
        Lost in translation Helmut 2009-10-06 21:46
  - non-central t is not the question ElMaestro 2009-10-05 17:32
    - Power of TOST in R d_labes 2009-10-06 16:45
      - Power of TOST (R package MBESS) Helmut 2009-10-07 02:18
        
        Power of TOST (R package MBESS) d_labes 2009-10-07 08:46
        
        Power of TOST (R package MBESS) Helmut 2009-10-07 12:38
      - The Power at limits d_labes 2009-10-07 11:32
        
        Power at limits 2 d_labes 2009-10-07 15:15
        
        package MBESS Helmut 2009-10-07 22:36
        
        Package MBESS and Power curiosity d_labes 2009-10-15 14:57
        
        Package MBESS and Power curiosity ElMaestro 2009-10-17 12:26
        
        Apfelstrudel and EFG 2.01 Helmut 2009-10-17 13:44
        EFG and another power curiosity d_labes 2009-10-20 11:35
        
        EFG and another power curiosity ElMaestro 2009-11-23 19:03
        
        EFG and another power curiosity Helmut 2009-11-23 19:17
- non-central t, power, R: yjlee168 2009-10-07 00:08
  - PASS Helmut 2009-10-07 01:16
    - PASS yjlee168 2009-10-12 23:17
- R code for power ElMaestro 2009-10-09 11:32
  - Belgian beer ? Ohlbe 2009-10-09 11:50
  - Another one - R code for power of TROST d_labes 2009-10-09 13:50