We are interested in testing
the effect of 5 different chicken feeds on egg production. We wish to test egg
production in a number of different breeds of chicken (we’re not interested in
the particular breeds, so we take a random sample of 4 breeds) to enhance the generalizability of our study. Some breeds of hens are
indeed known to produce more eggs per week than other breeds. Accordingly, we emply a randomized block design (with the breeds as blocks)
to minimmize this effect and estimate our treatment
effects more precisely.
Egg production data:
Breed (blocks)
a b c
d Mean
Feed I 5.00 3.25 3.00 4.00 3.81
Type II 7.00 2.00 4.00 2.50 3.87
III 4.00 2.50 3.25 4.40 3.56
IV 6.50 6.00 6.50 4.75 5.94
V 4.50 3.25 3.50 2.00 3.41
Mean 5.40 3.30 4.05 3.55
Test [alpha = .05] to
determine whether the five feed types tested differ in average egg production
per week.
Some numbers to aid in the
analysis: SS(Feeds) = 17.7 SS(Breeds) =
12.4 SS(Residual) = 12.7
Source
of Variation SS df MS F P-value
Feeds
(treatments) 17.7 4 4.42 4.18 .01<P<.025
Breeds
(blocks) 12.4 3 4.14
Residual
(interaction + error) 12.7 12 1.06