James H. Austin V

1
Odd or Even Favorite Numbers Test

• James H. Austin V
• Allie Levine
• Heather Luntz
• Pooja Viswanath

2
Introduction

• Allie, James, Heather and Poojas favorite
  numbers are odd.
• Is this a random phenomenon or does the general
  public really favor odd numbers over evens?
• Could this be due to cultural attitudes that
  surround certain numbers, or does it relate to
  birthdays, home addresses, etc?
• Could this knowledge be used to make bets to
  guess peoples favorite numbers? Would knowing
  this give us a better chance of having less
  people pick the same numbers as us in the lottery?

3
Goals

• Our goal is NOT necessarily to determine every
  subjects favorite number, but rather the overall
  tendency to favor odds over evens, or evens over
  odds.
• We tried to figure out the probability of a
  random persons favorite number being odd.
• More precisely, do people really prefer odd
  numbers to even numbers?

4
Protocol and Testing

• For the test we tested all the students that
  agreed to take our survey in food court, February
  16th, from 630-730, and stopped when we had
  received 100 numbers. (The last person actually
  made it an even 100 numbers. We were not going
  to stop them if they gave us a few more than
  100).
• Protocol
• We walked to each table with the surveys at hand
  and asked the subject to fill it out. The
  subject did not know that our test was to see
  whether people prefer odds to evens so not to
  influence their opinions. After they completed
  the survey we glanced it over to make sure that
  it was completely filled out.

5
Questions

• We asked our subjects the following questions
• Question 1.List your favorite numbers (max of
  five numbers).
• We asked each subject to list their top favorite
  numbers, up to five. We did not want to require
  them to list five numbers because some people do
  not have five favorite numbers and we wanted them
  all to be legitimate. Also, by allowing people
  to choose five numbers it accounts strength of
  preference.

6

• Question 2 Did you choose any of these numbers
  based on some special significance to you? (i.e.
  birthday, anniversary, age, superstition, etc)
  If so, which numbers and why?
• We wanted to ask some background questions to
  see if any of this information affects
  preference. We contemplated that people who have
  birthdays that fall on odd numbers may be more
  likely to like their birthday number, just
  because it is odd.

7

• Question 3 What is your gender?
• We asked this question to make sure we were not
  completely biased with all female subjects or
  male subjects. We also wondered if there would
  be a significant difference between male and
  female preferences for odd numbers.

8
Our Pretest Beliefs

• We believed that people prefer odd numbers.
• Furthermore, we speculated that of the favorite
  numbers we collected, about 75 of them would be
  odd numbers.

9
Hypotheses

• Null Hypothesis Overall, the chance of a
  random persons favorite number being odd is
  equally likely to it being even.
• pnull .5
• Alternate Hypothesis The chance of a persons
  favorite number being odd is more likely than the
  number being even.
• palt gt.5
• Power Hypothesis podd ppow .75

10
HOLY COW!! We Were Right!
11
Results
12
73 of the favorite numbers were odd while 27 were
even!!!
13
A graph showing our power hypothesis versus our
actual results.
14
Analysis of Results

• There were 73 odds and 27 evens.
• There were an overwhelming number of 7s and
  21s.
• A majority of people had reasons for why their
  favorite numbers were their favorite. The most
  popular reason was sports numbers and the second
  most was birthdays.

15
26 people gave reasons while 18 didnt.
16
(No Transcript)
17
(No Transcript)
18
Our Calculations
19
Conclusion

• So it turns out that people really do prefer odd
  numbers to even numbers.
• According to our results, 72 of people prefer
  odd numbers over evens. This is very close to
  our power hypothesis of 75.
• Our test is so powerful that the chance that our
  .75 hypothesis (power) is true, and we fail to
  reject the null hypothesis is totally negligible.
  This ends up being about a one in 5,000 chance
  of committing a Type II error.

20
Confounding Factors

• Most of the culturally popular numbers (i.e. 7,
  21, 69) are odd numbers, so one might argue that
  these numbers should be taken out when we compile
  our data. On the other hand, we feel that one of
  the reasons why odd numbers may be more popular
  is because of these culturally popular numbers,
  therefore these numbers cannot really be
  disregarded.
• Also, we surveyed people in Food Court and due to
  the fact that we were probably annoying the
  diners, they may have made up arbitrary numbers
  to get us away. Who knows, they could have given
  us the price of their food, how many chicken
  nuggets they had, or how many people were at
  their table.

21
So, Turns outEveryones just an ODDBALL