We have a question which asks something strange, something which seems like it would prevent any conclusions being made from the data.
The question is as follows:
The subjects of a study came from a genetically pure strain of rats. From each litter, one rat was selected at random for the treatment and one for the control group. Each animal in the treatment group lived with 11 others in a large cage, furnished with playthings which were changed daily. Animals in the control group lived in isolation, with no toys. After a month, the experimental animals were killed and their cortex weights (in milligrams) were recorded.
We are then given a table of data - one column with the treatment group and one with the control group and several rows of recorded weights.
The first two questions involve finding if there is an association between the treatment and cortex weights, they were no problem. However, the third question is as follows:
After doing this analysis you are told that the lab assistant who recorded the weights had mixed up the labels that identified the pairs of siblings. Test whether treatment affects the cortex weight of rats using appropriate graphical methods and a parametric statistical test. Check that the assumptions of the test are satisfied.
I cannot think of how we can use the data if the labels of the rats have been mixed up. It is worded as if the labels aren't simply swapped but rather the results are mixed. Surely if they were mixed up then the data is jumbled together and we cannot make any meaningful conclusions as we cannot separate the control and treatment? I would really appreciate if anyone could shed some light on what this question is expecting.
All the best!
