Está en la página 1de 4

Analytical Report

Do people do better on a math and


history quiz with their shoes on or off?

Matthew Walls
Statistics
Period 6
5-15-15
2014-2015

This was a very interesting topic to experiment for. When I was quizzing people they also
thought I was weird asking them to take their shoes off for a quiz. But at the end of the day this
experiment turned out to be significant and that was actually kind of surprising to me. This
experiment focused on if students have a higher combined score on a math and history quiz with
their shoes on and off. I ran a 2 sample proportion hypothesis test and setting the null hypothesis
at B1 = B2, where B1 is the true population proportion quiz score of subjects with their shoes on,
and B2 is the true population proportion quiz score of subjects with their shoes off. And the
alternative hypothesis at B1 < B2. First I tested the combined scores, and having a p-value of .
0448 was a very close call since I was trying to get under the alpha value of .05. I chose a
significance level of .05 because it is the normal level of significance. I had a t-value of -1.74
with the two averages and standard deviations of the control group(shoes on) and the treatment
group(shoes off) being 8.225, 9.95 and 3.1308, 3.1326 respectively. Then I tested the two
different quizzes individually with a 2-Sample T Test and for the history quiz it gave me a pvalue of .3122 with the t-value being -.4936 and the two averages and standard deviations of the
control group(shoes on) and the treatment group(shoes off) being 5.025, and 1.72, 1.80
respectively. That quiz did not turn out significant at all with the alpha level of .05 and it
wouldn't have even with an alpha level at .1. The next test I ran was another 2-Sample T Test but
for the math quizzes only and it gave me a p-value of .0156 and the two averages and standard
deviations of the control group(shoes on) and the treatment group(shoes off) being 3.2, 4.65 and
2.07, 2.03 respectively. This quiz turned out to be extremely significant and what helped the
combined test to turn out significant in the end. I had 20 people in each group totaling up to 40
experimental units. The calculator gave me a degrees of freedom of 37.9999. I collected my data
on Friday, May 8th 2015 in several AGS classes. I included mainly seniors, but with some

juniors, and sophomores here and there that I knew have at least taken pre-cal. I would give them
the two pages of math and history quizzes and tell them to not use a calculator and to tell half of
the people to take off their shoes. I used a convenience sample to collect my data. I would choose
two people to quiz each time and then choose the one with the shoes that were easier to take off
to take off their shoes for their two quizzes. On the history quiz, there was a question that had
two answers and the possibility of getting a half point for only answering one of them, on the
math quiz there was no such type of question. So this could've been some type of bias when I
only ran the test on the history quiz results since it was possible to get half points on one it and
not on the math quiz. Also there was no blinding since the subjects knew that they were being
experimented on. This project does not have much meaning to the average person at all, it was
mainly an interesting topic that I chose to explore more deeply on and luckily it came out
significant. My population was any person who is in some type of schooling. My sampling data
might not have been representative of the population because it only included 3 grades and was
not random. Ideally I would have had a math and history quiz for each grade level and would
have randomly assigned which students to take off their shoes and which ones to keep their shoes
on.
In conclusion, Since we found a p-value that is less than our alpha value then we reject
our null hypothesis that the true population mean combined score on a math and history quiz
with your shoes on is equal to the true population mean combined score on a math and history
quiz with your shoes off. This test, that is insignificant to anybody and everybody, turned out to
be significant in the end. Whether it was a mistake because of the sampling procedure or just that
it happened to be a very rare sample it does not matter because at the end of the day we can
reject the null and thats what we wanted to do in the first place. If I were to run this project again

I would take an SRS of the whole high school and randomized who had to take their shoes off
and who didnt.

También podría gustarte