Stats Paper

Mo 208 40 points Project 2 CONNECTIONS - REFLECTIONS The purpose of this second project is for you to connect what you are learning in this class to the world around you. This project will consist of three short papers and a project conclusion organized into a project that demonstrates your understanding of statistical inference. Over the next two weeks open your ears and eyes to watch for examples of statistical inference used in our society. Write a one page paper, connecting and reflecting, about once a week, and then organize your three examples with an introductory page and a closing page explaining the logic into a final project. More details: 1. Papers A) Your ideas for a paper may come from any media or professional source. (Please include the source

in your paper.)

B) Title each paper. Describe the example you found and clearly relate it to a specific model of statistical inference we have studied in class. (You may reflect on any problems, ethical issues, or societal values you wish to discuss.)

C) Hand in one completed paper (with the scoring sheet for papers

below) by the due date in your calendar for feedback.

An example paper is attached. The criteria used to grade papers is also attached.

D) Your three papers should connect with any three distinctly different models of statistical inference we have studied in class. For example:

Paper 1 - Confidence Interval (using model p ) Paper 2 - Hypothesis Test

(model: population mean µ ) Paper 3 - Comparing Two Groups p

2. Project - due ____ (see calendar)

A) Introduction, including a definition of statistical inference in your own words.

B) Three papers. C) Conclusion, closing statements. Be sure to include your personal

answer to "How do researchers use sample data to answer questions about large populations?". “What role does the significance level α play?”

D) Table of contents should be included after your title page. This project should be presented in a professional manner, of high enough quality that you could include it in your employment portfolio. (Word processing required.) If you have any questions e-mail or call me.

Sample paper 1: Holiday Hungry

It’s beginning to look like Christmas. Red, green, and white overwhelm every

store, sales and price tags become topics of conversation and cookies, pies, and fruitcakes

fill the table. It is no surprise that people gain weight during holidays. In a “growing”

country, holiday weight gain is either treated as something inevitable, or as a myth. A

study done by the University of Oklahoma found on pubmed.gov indicates that college

student participants considered overweight before thanksgiving, on average, gained more

weight over Thanksgiving than participants of normal weight. As the results state,

participants classified as overweight gained 1.0 kg (2.2 lbs) of weight during the

Thanksgiving holiday compared to a 0.2 kg (0.44 lbs) weight gain in the normal weight

group.

This study is an example of comparing two population means. The first

population is the group of participants considered overweight. The second population is

the normal weight participants. Each participant was weighed prior to and after

Thanksgiving. The study also indicates that the findings are statistically significant with a

P <0.05. This study used statistics to make an inference regarding the two populations

regarding holiday weight gain. Researchers set up a hypothesis test to compare the two

populations. The null hypothesis could be stated as follows: Normal weight participants

gain the same amount of weight over holidays than overweight participants. The study

used significance (alpha) of 0.05. Using hypothesis testing they resulted with a test p

value < alpha of 0.05. With a p value < alpha the null hypothesis is rejected and the

conclusion is made that normal weight and overweight participants do not gain the same

amount of weight over holidays. In fact, it was found that the overweight participants

gained on average 1.76 pounds more than normal weight participants.

Americans are in a losing battle of the bulge. Obesity rates are high, and despite

recent health awareness developments, rates continue to increase. A holiday weight gain

of even 0.44 pounds may seem innocuous, but it becomes significant if that amount is

gained for each holiday each year. Moreover, the average weight gain of overweight

people (2.2 lbs) puts this already vulnerable group at increased risk of weight gain and

weight associated diseases such as type II diabetes and heart disease.

While the findings from this study are alarming they also may be skewed. This

study was conducted with college students as participants. This may not make an accurate

inference about the general population. As a generality, college age people may drink

more alcohol than middle aged parents of young children. Alcoholic beverages are not

calorie free. Some of the extra weight gained during this period could have been acquired

at the local bar with friends. Secondly, the study only used college age students. In order

for the study to have more credibility as a population inference study, participants of

every age, race, and gender should be studied carefully. Regardless of the possible

problems with the study, holiday weight gain is a statistically significant phenomenon

and not just a myth. Therefore Americans should put down that 13th sugar cookie and opt

for some fruit instead. Or, at the very least, make “losing weight” a priority New Year’s

resolution.

Study used : The effect of the Thanksgiving holiday on weight gain.

Hull HR, Radley D, Dinger MK, Fields DA.

Source: Department of Health and Exercise Science, University of Oklahoma, Norman, OK, USA.

http://www.ncbi.nlm.nih.gov/pubmed?Db=pubmed&Cmd=ShowDetailView&TermToSe

arch=17118202&ordinalpos=1&itool=EntrezSystem2.PEntrez.Pubmed.Pubmed_Results

Panel.Pubmed_RVAbstractPlus

Sample paper 2: Newspaper Reports American Attitudes Changing Alternative life styles are a hot topic in the media these days. "Nearly six in ten Americans think that gays can be just as good at parenting as heterosexuals" is the claim stated in an article printed in the Saginaw News on October 28, 1996 (page A4). The article details a poll taken of 929 adults by Newsweek magazine where individuals were asked questions about their opinions of children being raised in nontraditional households. This is an example of statistical inference because the magazine is claiming most Americans support gay parents, based on a sample of 929 Americans. The model being used is the population proportion or percentage ρ ("six in ten" is claiming 60% of all Americans). It sounds like they asked each participant a question like, "Do you think gays can be just as good at parenting as heterosexuals?". They then probably recorded the number of people who responded "yes", "no" or "uncertain". The article reported that 57 percent said "yes". This article may be accurate and then again it may not be. If the sample was taken on a liberal college campus, or in a gay community, the results could be biased, not reflecting the views of the typical American. Hopefully Americans are becoming more tolerant of individuals living alternative lifestyles. Gays and lesbians have been discriminated against in this country for years and now those who want to marry, buy homes together, or raise children have fewer barriers. Public attitudes have much power over the type of legislation influencing the lives of gays and lesbians.

Math 208 Project 2 Name___________ Grading Criteria for Papers Expectations for Papers Student self-assessment(check last column) Instructor Evaluation (points)

10 points total

student self- assess (check)

Appropriate title Introductory paragraph(s) describing your example Summarizes main topic of your focus in source. Identifies the question of statistical inference. Correctly references your source.

Primary paragraph connecting to a particular model Clearly describes what statistical model you see Explains why this model (What data did they collect? Did they calculate a mean? A percentage? One sample or two? You may need to improvise a bit here to support your choice of model because many media reporters gloss over research details as they summarize.)

A concluding paragraph discussing the importance of your example as you see fit. (How could the results affect our future? What other research could be pursued to answer their questions? How could a potential flaw you noticed be improved?)

Paper uses correct ESWE (edited standard written English). Paper is presented in professional format (word processed, written in third person, neat, etc.).

PARTIAL CREDIT Paper meets all above criteria EXCEPT the following

percentage of points lost

Title missing or not appropriate 10% Source not referenced properly 10% Article not summarized 20% Question of statistical inference not identified 10-30% Model of statistical inference incorrect or ambiguous 10-40% Concluding paragraph shallow 10-30% Concluding paragraph criticizes sample size inappropriately 10% Paper doesn't meet EWSE (edited standard written English) 10-40% Name____________ Project 2 – Grading Criteria (Please submit this with your project) PROJECT MEETS ALL BELOW CRITERIA

possible points

40 student score

Professional Presentation All writing uses correct ESWE All writing directed to a professional audience (written in third person)

Project Introduction Including personal definition of statistical inference (in your own words - not copied)

2

Paper 1 (see above criteria) 10 Paper 2 (see above criteria) 10 Paper 3 (see above criteria) 10 Project Conclusion Clearly explains how sample data is used to make conclusions. Clearly explains the role of the significance level in the process of using data to make conclusions

8

Last update 8/04