Imph
★

2023-03-06 15:29
(527 d 20:51 ago)

Posting: # 23487
Views: 2,488

## Sample size for replicate crossover studies [Power / Sample Size]

Hi,

For sample size calculation in the case of full replicate design using R software, when specifying the 2x2x4 in the script, do we get the same sample size whether the design is to be : TRTR|RTRT or TRRT|RTTR. And same question for the partially replicated design TRT|RTR vs TRR|RTT.

In the same way for manual computation, can we use the sample size formula presented by CHOW for the TRRT|RTTR design to calculate the sample size for TRTR|RTRT design. And same question, could the sample size formula of TRR|RTT design be used for TRT|RTR design.

Thanks a lot.
dshah
★★

India,
2023-03-06 16:40
(527 d 19:40 ago)

@ Imph
Posting: # 23488
Views: 2,089

## Sample size for replicate crossover studies

HI Imph!

The sample size would remain same because it is mainly driven by variability, desired power and ratio.
Regards,
Divyen
Helmut
★★★

Vienna, Austria,
2023-03-06 17:48
(527 d 18:32 ago)

@ Imph
Posting: # 23489
Views: 2,101

## Sample size for replicate crossover studies

Hi Imph,

❝ For sample size calculation in the case of full replicate design using R software, when specifying the 2x2x4 in the script, do we get the same sample size whether the design is to be : TRTR|RTRT or TRRT|RTTR.

Yes. See there for the covered designs.

❝ And same question for the partially replicated design TRT|RTR vs TRR|RTT.

These are 3-period full replicate designs. See there for partial replicates.

❝ In the same way for manual computation, can we use the sample size formula presented by CHOW for the TRRT|RTTR design to calculate the sample size for TRTR|RTRT design. And same question, could the sample size formula of TRR|RTT design be used for TRT|RTR design.

Why are you still referring to Chow’s approximations full of typos? At least when you deal with reference-scaling you need simulations anyway. A formula does not – and cannot! – exist.

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
Imph
★

2023-03-07 12:30
(526 d 23:50 ago)

@ Helmut
Posting: # 23490
Views: 2,074

## Sample size for replicate crossover studies

Hi,

❝ Why are you still referring to Chow’s approximations full of typos? At least when you deal with reference-scaling you need simulations anyway. A formula does not – and cannot! – exist.

I've been asked in my compagny to provide sample size formula for TRTR | RTRT and TRT | RTR designs.

Thank you all.