Design of the Month

Zebra finches

The following quotation is taken from page 11 of the New Scientist of 30 August 2008.

Zebra finches mate for life, so choosing the fittest partner is vital. ... Akshat Vyas at City University of New York and colleagues studied how early experience, quality of song and oestrogen levels all affected a female's response to male birdsong.

The team split the finches into four groups. One group had never heard male birdsong before the test, while the remaining groups were each tutored with one of three male songs: complex, simple, or long. The experiment was carried out three times, first when the finches had naturally low levels of oestrogen, because they were not ready to mate, second after artificially raising levels to mimic those normally seen prior to mating, and third, after they [sic] birds were drugged to lower oestrogen levels.

Ignoring the quality of song:

  1. What were the treatments and how many of them were there?
  2. Draw the Hasse diagram for the factors on the treatments.
  3. What were the experimental units?
  4. Draw the Hasse diagram for the relevant factors on the experimental units before treatments were applied. What extra information do you need in order to complete this?
  5. Making a plausible guess at the missing information, write down the skeleton analysis of variance table, showing stratum, source and degrees of freedom.
  6. Design a better experiment to achieve the apparent aims of the reported one.

Other designs

Page maintained by R. A. Bailey