Written Homework 1

Due Friday, September 17 by 11:59 PM.

1) Roughly every 10 years, the American Film Institute selects the top 100 movies of all time. From
the last such survey, suppose the top 100 movies are the cases (or observational units) in a study
to compare these movies. For each of the following, indicate first whether it can legitimately be
considered a variable or not. If it is a variable, classify it as categorical or quantitative. If it is not
a variable, explain why not. In order to decide whether the variable is legitimate, think about
what the cases are and review the definition of a variable outlined in class.
a. Length of the film - Variable Quantitative

b. Whether or not the film occurred after the year 2000. - Variable Categorical

c. The number of films which were longer than 150 minutes. - Not Variable – It cannot be
measured on the individuals in the population. If it was “whether or not the films were
longer than 150 minutes,” that would be measurable.

d. The type of film (Drama, Comedy, Science Fiction, Action, Documentary, other). -
Variable Categorical

e. The rank of the film in the top 100 list. - Variable Categorical

2) Suppose you are interested in the population of all elementary school students in Iron County.
Suppose also that we collect information from a (non-random) sample of 30 elementary school
students at a particular elementary school. Identify each of the following as either a parameter
or statistic and explain your reasoning.
a. The average score received on a standardized math test by the 30 students sampled. – A
Statistic: we are collecting the data from the sample, not the whole population

b. The proportion of elementary school students in Iron County participating in after-school

c. The proportion of the 30 students sampled who participate in after-school programs. – A

d. Explain how this sample may be biased. – This sample is non-randomized. All of its
individuals are from the same school. Different schools or areas may have differing
levels of involvement or differing average scores.

3) On July 15, 2004, the Harris Poll released the results of a survey asking whether people favored
or opposed abolishing the penny. Of a national sample of 2136 adults, 59% opposed abolishing
the penny. Explain what is wrong with each of the following statements:
 a. The population is the 2136 adults contacted by the Harris Poll. – The population was all
adults. The 2136 adults were the sample.

b. The sample is the 59% who oppose abolishing the penny. – The sample is the 2136
adults. The 59% is a statistic.

c. The variable is the 59% who oppose abolishing the penny. – The 59% is not a variable, it
is a statistic collected from the sample.

d. The observational units in this survey are pennies. – The observational units were the
American adults, as they were the ones observed.

e. The parameter consists of all American adults. – The parameter would be the result if
ALL American adults were polled. In this case however, only a sample was surveyed.

f. The statistics is the average number of adult Americans who oppose abolishing the
penny. – The statistic is the percent of people that opposed abolishing the penny, not
the average number (because only one sample was polled, you cannot get an average
of samples)

4) In order to study the effects that wildfire smoke has on human health, a researcher plans to
randomly sample 100 people from each of four states in the US: California, Montana,
Washington, and Virginia and do a biometric screening on each person. Since Virginia does not
get frequent wildfires, the researcher will think of people from that state as a “control” group.
a. What are the variables that will be measured on (or collected from) each observational
unit (person) in the sample? – The health effects of smoke. I.e., the change in health
from before the study to after the study.

b. Is this an observational study or an experiment? Explain your answer. – Observational.

c. Can this sampling design be best classified as a simple random sample, a stratified
random sample, a cluster sample, or a systematic sample? Justify your answer. –
Stratified Sampling. The population was divided into groups by state, and a random
sample was taken from each group.

d. If it turns out that people in Virginia were, on average, in better overall health than those
in California, Washington, and Montana, is that evidence that wildfires are the cause of
the health differences? Why or why not? – Not necessarily. There may be other,
confounding variables that may have influenced the health of the individuals, and
therefore have influenced the results.