Video Transcript
A doctor wants to find out
about some possible side effects of a common drug they have prescribed. Which of these samples is
unbiased? Is it (A) sending a survey to a
select group of patients? (B) Interviewing patients who
suffer from side effects of the drug. (C) Interviewing all the
patients who come for an appointment on a Saturday. (D) Interviewing patients who
come for an appointment during the week at random. Or (E) generating a list of
patients to interview by phone randomly from the patient registry.
In order to identify whether a
sample is unbiased or biased, we need to ask ourselves a question, does each
member of the population have an equal chance of being selected? If the answer to this question
is yes, then our sample is unbiased. In this question, the
population are the patients that have been prescribed the drug. We want each of these patients
to have an equal chance of being selected. Option (A) does not satisfy
this criteria, as only a select group of patients have been surveyed. Option (B) is also incorrect,
as this time we’re interviewing patients who have suffered from side effects of
a drug. This means that our results
will be skewed. Each member of the population,
those who have and those who haven’t suffered side effects, must be able to be
selected.
Options (C) and (D) are
incorrect as we’re not asking the correct members of the population. In option (C), we’re asking all
the patients who came for an appointment on a Saturday. Many of these might not have
been prescribed the drug. Also using this sample, we’re
not interviewing any patients from any other day. Option (D) has a similar
problem to option (C) in that we have no way of knowing if these patients have
been prescribed the drug.
Option (E), on the other hand,
is the correct answer. We are generating a list at
random from the patient registry. This will include all the
patients that have been prescribed the drug by the doctor. Each member of the population
has an equal chance of being selected.