Coding for pairwise comparisons [General Statistics]
Hi John & Neera,
Alternatively (as suggested in EMA’s GL) you could drop the irrelevant administrations from the pairwise comparisons whilst keeping the sequence/period coding. Recoding the 6×3 Williams’ design…
…gives two Incomplete Block Designs allowing separation of CVs:
Edit: For details see this article.
❝ […] Yes you will end up with 1 set of intrasubject CV due to the pooled variance effect.
Alternatively (as suggested in EMA’s GL) you could drop the irrelevant administrations from the pairwise comparisons whilst keeping the sequence/period coding. Recoding the 6×3 Williams’ design…
Seq Per Trt
1 1 A
1 2 B
1 3 A+B
2 1 B
2 2 A+B
2 3 A
3 1 A+B
3 2 A
3 3 B
4 1 A
4 2 A+B
4 3 B
5 1 B
5 2 A
5 3 A+B
6 1 A+B
6 2 B
6 3 A
…gives two Incomplete Block Designs allowing separation of CVs:
Seq Per Trt Seq Per Trt
1 1 A 1 2 B
1 3 A+B 1 3 A+B
2 2 A+B 2 1 B
2 3 A 2 2 A+B
3 1 A+B 3 1 A+B
3 2 A 3 3 B
4 1 A 4 2 A+B
4 2 A+B 4 3 B
5 2 A 5 1 B
5 3 A+B 5 3 A+B
6 1 A+B 6 1 A+B
6 3 A 6 2 B
Edit: For details see this article.
—
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
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:
- Drug interaction Neera 2013-12-13 08:39 [General Statistics]
- Drug interaction jag009 2013-12-13 20:55
- Coding for pairwise comparisonsHelmut 2013-12-16 15:13
- Coding for pairwise comparisons Neera 2013-12-17 09:00
- Coding for pairwise comparisonsHelmut 2013-12-16 15:13
- Drug interaction jag009 2013-12-13 20:55