| This project is an opportunity
for you to apply the statistical methods you will learn in this class to a
topic of your own choice. Below are the design features of the project
which are required; the topic you choose to research is your choice. You
will be spending a lot of time gathering and analysing the data, so it’s
best to select a topic of personal interest to you - perhaps relevant to
your major, or even just ‘campus trivia’ which interests you. you may
extend the topic any way you wish.
There will be two preliminary phases to your project which will be due
during the block; your completed statistical report will be due on the
last day of class. On the second Tuesday you will turn in a statement of
your topic selection. On the third Tuesday you will turn in your data
together with basic summary statistics.
It is important that you start to think about your project topic soon
(mostly because the data collection takes time). The design must include
two types of variables - one quantitative and one dichotomous.
The Dichotomous Data
The dichotomous data can be viewed as a YES/NO response to some
question. For example: Are you male? Do you drink coffee? Are
you an upper-classman? Do you take recreational drugs?
The Quantitative Data
The quantitative data has numeric values. The differences between the
variable values should carry comparitive meaning. Examples: What is
your GPA? How many cups of coffee do you drink per day? How many hours do
you sleep per night? An example which is not good would be What is
your major? Although you can assign numeric values for students
majors, the relative sizes of those numbers would not mean much.
One aim of the project is to determine what proportion of your ‘population’
falls into each dichotomous group.
Examples:
For example, suppose that your
population is the CC student body, your dichotomous variable is the
response to "Do you drink coffee" and your quantitative
variable is "How much sleep do you get each night". You
would be able to determine the proportion of coffee drinkers on campus.
Another goal of the project is to compare the two groups determined by the
dichotomous variable. Suppose that the coffee drinkers get 9.6 hours sleep
per night on average, and that the non-coffee drinkers get 9.8 hours of
sleep per night. Is this a real difference? Is it just some random
fluctuation due to chance error? A hypothesis test will enable you to make
such decisions.
More Examples from previous years:
C (From 1998): What proportion of
the population believe that President Clinton should be impeached, and
is there a difference in educational level between the group which is
for impeachment, and the group which is against?
C What proportion of students
attended religious services regularly as children, and how often do they
now attend religious services?
C What proportion of students are
living off campus, and do the students living off campus own more potted
plants than those living on campus?
C Do men
and women on the swim-team score differently on the Rosenberg
self-esteem scale? Is there a correlation with self-esteem and
performance?
C The
dichotomous variables for this project: gender, parental income, upper
vs. lower classmen. Quantitative question: How much involvement do
students have on campus in community service.
C Does
dating affect school performance?
C Are male
students really better at spatial perception, and faster to identify
geometric objects which have been rotated?
C Blood
Pressure project: This project had several dichotomous variables: sex,
alcoholism, smoking, stress, diet, and attempted to observed differences
in blood pressure between groups.
C Does the
practice of martial arts cause more violence, or does it teach self
control? Qualitative question: Have you taken a martial arts class?
Quantitative question: How many fist-fights have you been in on campus
each year?
C Of all
phone calls to the investigators dorm room, what proportion are for the
investigator and what proportion are for her room-mate? Does the
investigator's room-mate spend more minutes on a phone call than the
investigator does herself?
C What proportion of drivers have
red cars, and do they run the red light by a greater margin than other
drivers?
C Project on peoples perceptions and
beliefs about the use of animal research for developing drugs.
C What are
the different levels of alcohol consumption between various campus
groups (male / female, upper/lower classmen, etc.)
C By
interviewing people outside the DMV, determine the proportion of people
who drive a four wheel drive vehicle, and whether or not they are less
accident prone.
C Is there
a difference in education levels/GPA between creationists and those who
believe man evolved from Apes.
C What proportion of CC students are
in a sorority or fraternity, and do they have more tattoos?
C What proportion of
CC students have tattoos, and do the number of body piercings differ
between the tattooed and non-tattooed group?
C What proportion of
CC students use marijuana, and do they have more tattoos?
C Do in state or out of state
students spend more money on phone calls?
C Do
male students smoke more than female students?
C Do
people involved in organized campus activities which require 15 or more
hours a week of time have a poorer GPA?
C Are
people who do not label themselves as feminists really feel the same way
about women’s issues as people who do label themselves as feminists?
C Did women
or men spend more time in the Colorado State Mental Institution, 1879 to
1889?
C Do people
who claim to be at their ideal weight really weigh less than those who
do not claim to be at their ideal weight?
Examples of How to take Surveys:
C Telephoning students from a random sample of
100 registered students obtained from the registrar.
C Worner box survey ... expect a 20 %
response rate!
C Colorado Springs telephone survey.
C Design a survey
and ask people at Worner centre to complete it.
C Watching and
recording data directly - e.g. observing traffic patterns.
The size of the sample should be between 50 and 100, but it may be
smaller in special cases. Discuss it with me if a sample of size at least
50 is not possible. The method of making the random selection is up to
you, but you should decide on a method before you begin and then select
all of your subjects consistently. In your eventual write-up you will
comment on any biases which may have been introduced by your method. You
must select your own sample, but your sources may include college offices,
local businesses, government agencies, laboratory experiments or a survey
you conduct. Randomly generated lists of 100 CC students (first names and
phone numbers) will be available. Your grade on the project will be based
predominantly on your final report.
|