Statistical modelling / Milk curve comparisons

As already indicated, the fitting of Wood's model to the single set of milk values obtained pre and post the longitudinal survey is not really appropriate. It is better to fit a simple linear regression and compare slopes and intercepts.

We have put the milk values for the pre- and post-surveys into separate worksheets in CS6Data4, having restricted the range of days post calving to between 1 and 270.

Calving dates for the monitored cows ranged from 3rd March to 8th October, 1999 - not for the full year. Thus, in any one month there was not a complete range of days across the lactation period.

In order to produce a similar data set to the other two years, it was decided to use the first milk yields recorded in June and December, respectively, for the cows that were sampled in those months, and then to pool these two data sets. These data have also been stored in CS6Data4.