Dear Members/Colleagues:
In a recent book review, I specifically commented on the following, which
authors and instructors should give attention to.
1. When it is emphasized that "correlation does not equal causation" in the
context of "Pearson's correlation" only, this quite likely can lead to
students believing that this notion does not apply to the other tests of
correlation/association/relationship/difference. Rather, students should be
told that it is the design and not the statistical test that can lead to an
outcome of association (observational design) or causation (true experimental
design).
2. The simple random sample does not assure representativeness. It addresses
selection bias.
What do you think?
Regards,
Rossi
|