1.) My research project aimed to answer questions concerning Kalamazoo College’s
housing and living provisions for students. I believe that ones living conditions
is a very significant factor in their college experience. As a student, I’ve
heard many people’s opinions on the food we eat, the dorms we live in,
and the amenities offered for us. Finding out what students think about current
housing will identify the problem. In addition, the current housing policy that
of Kalamazoo College is critical to the schools unique character. Finding out
what students think about current housing identified the problem. From there,
I wanted to know what students thought would improve these conditions. This
provides insight for a solution. The college is not providing a positive environment
if students are not pleased with the current housing at this college and feel
forced to live on campus until their senior year. Kalamazoo College should aim
to make students want to live on campus as part of a close community. I believe
my study will show what would be good investments for Kalamazoo College to make
in order to improve student life.
2.) My population of interest was students at Kalamazoo College. Of the 1200
students, I sent surveys out to one hundred. I made 100 survey packets that
consisted of the survey, a return envelope labeled “Dayna Doman Box #
175” and a Dum-Dum sucker For my survey, I decided a systematic random
sample was ideal. Because my topic concerns student housing, using a convenience
sample, such as students in only my dorm, would have led to biased results.
I selected my sample by generating a random number between one and fifty to
be the starting person, and an additional random number between one and fifteen
to be the sequence. By using a random number generator (www.random.org), I started
with the 29th person on my list and picked each 8th person. The list I used
was the spring quarter 2006 student phone directory. This list includes all
enrolled students, both on and off campus. The list is alphabetical, and not
sorted by year. Participants were asked to complete the survey, put it in the
marked return envelope and send it to my mailbox through student mail. I received
thirty-six surveys back, which is a little lower than what I had hoped for,
but still significant enough to conduct a well- rounded statistical study. This
was a 36% response rate and 3% of the total population. Unfortunately, I cannot
say my resulting sample was representative of the population. A majority were
females, freshmen and sophomores. This may be because off-campus students do
not check their mail center boxes often. Although I did not get a large number
of upperclassmen or off-campus students, my sample still included both.
3. A) My descriptive statistics allowed me to see both what students thought
about current student housing and what alternative and improvements most appealed
to them. The lowest rated aspect of current student housing were, as I anticipated,
the cable service and meal plans. The highest rated aspect of current student
housing was our internet service, even though the demand for wireless was also
the highest. The most appealing alternative housing options were apartment-style
dorms and townhouse-style dorms. The least appealing were building on-campus
houses and adding programs to integrate off-campus life. The most appealing
housing options were cable and wireless internet. The least appealing options
were adding long-distance phone plans and putting kitchenettes in dorm rooms.
Additionally, the overall mean of current student housing on a scale of one
to five was 3.093. I see this as a fairly low overall rating of current housing,
and hopefully a good indication to our college that changes should be implemented.
For my results of what students would be willing to pay for alternative housing
and housing options, I had to change my data. Students were asked to check off
a range of price, such as $100-300. I thought asking them to just write down
a price would be confusing and too time consuming for participants. By giving
them a range, I only had one person not complete it. However, in order to calculate
descriptive statistics, I had to convert each range into one number. Each range
was converted into the median of it (so $100-300 became $200). The >$3000
range was converted to $4000 to prevent any extreme values. It is important
to note that only one participant checked that price range. I was not surprised
to see that people would pay more for new housing than options. I was most surprised
that even though students rate cable here very low and find cable very appealing,
they weren’t willing to pay too much for it. The mean for cable was $77/year.
This may be because students think should be a college investment and not something
they need to take full financial responsibility for.
3 B.) I created confidence intervals for the highest and lowest ranking categories
of current student housing, alternative student housing and alternative housing
options. Each of these has a 95% confidence. Since my sample size was 36, I
used the t-value for 40, which is 2.021. Each of these intervals confirms that
there is statistical significance in saying that it is the highest or lowest
ranking. None on my lowest ranking intervals cross over into the same range
of my highest ranking intervals. The standard errors for these six categories
were fairly close to each other, ranging from about 0.14 to 0.27. I had originally
feared that large standard errors and my smaller sample size would create confidence
levels that crossed over and proved that my highest and lowest ranking results
could be interchangeable.
Current Housing
HIGH Internet: 3.530-4.246
LOW Cable: 1.468-2.198
Alternative Housing
HIGH Apartment-Style Dorms: 3.906-4.482
LOW Off-Campus Life Programming : 2.866-3.746
Alternative Housing Options
HIGH Cable: 3.955-4.711
LOW Long-Distance Phone Plans: 2.230-3.324
One of my most interesting proportions I found was how many students believe
we should allow off-campus housing after their first year of school here and
allow students to have no requirement at all for the meal plans. My survey included
options for both ranging from after the first quarter of first year to senior
year (the current policy). Of thirty-five participants (one didn’t respond),
54.3% thought Kalamazoo College should allow off-campus housing either after
the first year or even before the end of first year. Additionally, 25.7% thought
the school should allow off-campus housing after sophomore year. 68.6 % of participants
thought the school should have no meal plan requirement at all, and 22.8% thought
the requirement should only be either before or by the end of the students first
year at K. Even after creating confidence intervals with a 95% confidence, my
results illustrated that a large proportion of students believed we should have
less housing requirements and no meal plan requirement.
Housing Requirement
After First Year 37.8%- 70.8%
After Second Year 11.2%- 40.2%
Meal Plan Requirement
No Requirement 53.2%- 84.0%
After First Year 8.9%-36.8%
3C.) For my first hypothesis testing, I tested my mean of satisfaction with
Kalamazoo College life to that of the Princeton review. The variables I used
were the ten variables in my “Current Housing” section. The Princeton
review uses a 60-99 rating scale and I used a 1-5 scale. This is because of
the complexity of my study as compared to that of the Princeton Review. The
mean rating of life at Kalamazoo College was 73, which is 13 points up, on a
39 point scale. This is at the 33.3% point along the Princeton Review scale.
If I converted this a 1-5 scale, the mean would be 1.67, which is at the 33.3
percentile of my 1-5 scale. I hypothesized that the mean of the Princeton Review
would be equal to my mean. Because The Princeton Review did not provide data
on a standard error for their calculation, I will be testing to see if I can
reject their mean from my own. H null is that the mean is 3.093 , and H alternative
is that the mean is not 3.093.
My equation was 3.093 +/- 2.021 (.63275) , which led to a critical region of
1.814 to 4.372. Although it came fairly close, I can reject The Princeton Reviews
calculation of 1.67. It is important to note that their study was probably conducted
with more participants in a population not exact to my own (it’s an older
study). Additionally, the factors used to measure a rating of campus life for
each study were similar, but not identical.
My second hypothesis was to compare my results of the proportion of students
saying they live off campus, the percent of males and females, and the mean
GPA from my study to that actually reported for the college. The Princeton Review
last recorded that 26% of students live off campus and that 57% are female.
A fiftey-student survey I found online stated that “Kalamazoo College
students surveyed reported having an average GPA of 3.33 with a standard deviation
of 0.37.” (Student Spending at Kalamazoo College and Harvard). My sample
showed a mean of 3.31, 48.6 % females, and 5.6% of students living off campus.
For percent of students living off-campus, my standard error was 7.46%. This
gave me a critical region with 95% confidence of -14.62 % to 14.62 %t difference
between the two calculations. Because 26-5.6 is 20.4, I can reject the null
hypothesis that P1-P2=O. For the percent of females on campus, my standard error
was 11.54% which gave me a critical region with 95% confidence of -22.61% to
22.61%. I can not reject the null hypothesis that P1-P2= 0 because 57-47 is
10 and falls within the critical region. For student GPA, my standard error
was 0.11. This gave me a critical region with a 95% confidence of -.23 to .23.
I can not reject the null hypothesis that x1-x2= 0 because 3.33-3.31 equals
.02 and falls within the critical region. I’m glad that I could get at
least some demographics close to other studies done. I had expected from the
start that I would not get a proportional amount of off-campus students. This
is simply because they are less likely to be checking mail in the Hicks Center
since they live off-campus.
3E.) I had two multiple regressions that I thought were important for my study.
The first was to compare the relation between the satisfaction of current housing
( a mean of all ten ratings for each participant), to the family income of the
student, the GPA of the student, and the age of the student. That is, with age,
GPA, and family income as a function of satisfaction with current housing. I
had 22 participants that had answered both their GPA and annual family income.
My equation was Y=.258 x1- 8.044 x2 -0.044 x3 -1.536. This means there was a
negative Y-intercept, a slightly positive relationship between age and the satisfaction
with the school (+.258), a very negative relationship between satisfaction with
the school and income (-8.044), and a negative relationship between GPA and
school housing satisfaction. . I expected that older students would be less
satisfied with the school, but was wrong. I did think that the more income a
student had, the less satisfied they’d be with housing, however not a
strong as the regression suggests. This is probably because many students families
make over $100,000 a year. I also thought students with a higher GPA would be
more satisfied, but was wrong. My R-squared was 0.43, which means the variation
isn’t fully explained by the equation, but isn’t completely unexplained
either.
My second multiple regression was to see the relation between the appeal of
on-campus apartment-style dorms, the amount students were willing to pay (using
my conversion earlier stated), and the annual income of the student. This included
26 participants. Apartment-style on-campus dorms were the highest ranking alternative
housing option. The regression I calculated was Y=.0011 x1 + 1.082 x 2 + 3.737.
There was a very slightly positive relationship between the appeal of the housing
option and the willingness to pay (+.0011). This was not what I expected. One
would think that the appealing the idea, the more willing a student is to pay
for it. I did find a positive relationship between annual income and the appeal
of apartment-style dorms. After seeing the negative relationship between satisfaction
with current housing and income, this doesn’t surprise me. However, the
R-Squared was 0.264, which means the line has a pretty “poor fit”
and has a lot of unexplained variation.