Supplementary Exercise 3.94
---------------------------

(a) A sample survey.
The objective is to describe a distribution of opinions among people
about a particular question of interest. There is no comparison between
groups and therefore no use for experimentation. A survey could be
considered as being observational because no treatments are imposed,
but this is really because no treatments are studied.


(b) An experiment.
The objective involves a comparison between two types of learning, and
the best design (or study type) to compare treatments is an experiment.
One question is whether it will be feasible to conduct an experiment
with a sufficiently large group of participants to make these
representative for a population of wider interest. The participants
would have to agree to being assigned to one of the two learning types,
and for ethical reasons that could probably only happen on a voluntary
basis. Volunteers can differ in their characteristics from the
population of interest. On the other hand, if the study was done as an
observational study, one would have to deal with potential confounding
by factors related to the students' choice of learning type, and that
might be difficult as well.


(c) An observational study.
The objective does not involve a comparison of groups but is only a
description of current practices among the teachers. As the teachers
most likely do not have a well-defined rule for how long to wait for
answers, there is likely to be variation in these waiting times for each
teacher (across different instances). A survey would not easily be able
to capture that variation because it collects information at a single
point in time. An observational study whereby observers record the
waiting time at a specified number of instances for each teacher would give
better information. The study would be observational because there is no
intention to influence the teachers' actions, although the basic fact
that observations are being made may very well affect the teachers' actions!