Field Project

Summary:  Your field project will apply what you have learned about inferential statistics. You will formulate a question, plan your project, collect or find sample data, perform a hypothesis test to answer the question, and submit a written report. This represents almost 20% of your course grade, and you may earn extra credit by presenting your project to the class.

Project Plan

Many past students put off their projects till the last minute, and then they didn’t have time to do their best work or they picked an unacceptable project. (For example, they did a project of descriptive statistics instead of inferential statistics.) To prevent such problems, please talk your idea through informally with Prof, Marvel in person or by email, then write out a Project Plan and get Prof. Marvel’s approval.

An approved project plan earns you the first 10 points out of 100 for the project. Without an approved project plan you don’t get those points, you can’t present your project to the class for extra credit, and you run the risk that your project topic won’t be acceptable at all, which means a zero on the assignment. These should be overwhelming incentives to plan your project in advance, which is crucial in doing quality work.

To give you an incentive not to put the project off till the last minute, no approvals will be given after the end of class on the date shown in the schedule. Allow yourself extra time, because most students need to make some revisions in their first Project Plan before it’s finally approved.

After reading through this protocol, please see the separate Field Project Plan.

Ideas

You should pick a question that matters to you, probably in your professional field of interest, or perhaps a cherished hobby. Finding it hard to get started? The following ideas from past projects may help, but be creative! Remember that originality is part of your project grade.

Numerical data — one mean, one sample

• Do people tip less than 15% in a restaurant where you are a server?
• Do students’ GPAs go up or down after they change programs?

Numerical data — two means, two samples

• Is there a difference between the number of hours that TC3 men and women work at a job?
• Are prices lower at Store X than at Store Y?

Attribute Data — one proportion, one sample

• What is the proportion of students that own a computer, have Internet access, believe a computer course should be required, etc.?

Attribute data — two proportions, two samples

• Are couples who sleep in king-size beds more likely to get a divorce?
• Do women in traffic stops get out of more tickets than men?

Attribute data — multinomial

• You could test almost anything having to do with preferences or activities — video rentals by day of week, preferred political candidate, class of car — against a reasonable model that you’ve come up with, or seen published for some other distant geographical area.

(Since colors of M&Ms are a class example, your project should not be about colors of candies unless it’s substantially different.)

Data Collection

Before you plan your data collection, make sure you’re doing inferential statistics. Remember Chapter 1? Inferential statistics is taking a known sample and asking what the unknown population is like. If you already have data on the whole population, that's descriptive statistics and is not material for a Field Project.

Sample Size

For your own original data, around 50 attribute responses or 35 numeric responses per sample are acceptable. (For Case 6 or 7, plan your sample size so that, according to your model, all the E’s are ≥5.) While a bigger sample size is almost always better, don’t make yourself nuts over this project. You’ll need time for designing your study and analyzing your data.

If you’re using published data, please discuss sample size with Prof. Marvel in advance.

Survey Guidelines

If you’re taking a survey to gather data, please follow the guidelines in this section very carefully. Failure to obtain required approvals will make your project unacceptable.

Survey form: At a minimum, you must include the following, prominently:

• Your name and class
• A statement that participation is voluntary
• A statement that all information is confidential

Make sure nothing on the form lets you identify the respondent personally.

Approvals: Before you distribute any survey, you must obtain all required approvals in writing; allow extra time for this.

• Prof. Marvel must approve every survey, without exception.
• If the survey will be distributed to anyone outside TC3, the Office of Institutional Research must also receive written notice in advance. Prof. Marvel will give you further information with his approval.
• If any minors will be given the survey, you must have their parents’ consent in advance. In practice this is so difficult to do that I advise against any project involving folks under 18.

Attach all approvals to your project when you hand it in.

Survey technique:

Have respondents fill out a form and put it in a closed container. This is essential both to protect respondents' privacy and to ensure accurate answers. Because people being interviewed tend to answer what they think the interviewer wants to hear, it’s not acceptable just to ask people a question and write down their answers.

If you want to distribute your survey during a MATH200 class, be sure to arrange this with the instructor at least a day in advance. If you want to distribute your survey in another class, ask well in advance but understand that the instructor may not have time. In any class where you distribute a survey, you must announce that this is voluntary and not a class requirement.

TC3 Survey Guidelines must be followed. Please see <http://www.tc3.edu/dept/ir/guidelines.asp> for the full guidelines.

Data from Other Sources

You need not collect the data yourself. In your report, cite your source (see the TC3 library’s Citation Guide if you need help) and attach a printout, photocopy, or tear sheet as appropriate. (If you collate data from many sources, please talk to Prof. Marvel before copying or printing many sheets of paper.)

The Office of Institutional Research provides anonymized data from the TC3 student database for many types of requests. You can obtain data for most requests yourself by visiting http://www.tc3.edu/dept/ir/math200.asp. If you need other forms of student data, please consult with Kristine Altucher at least two weeks in advance to find out whether your request can be filled and to allow enough time for her office to prepare the data.

When you use the student database, remember you must do inferential statistics. That means you must take a sample and use it to test a claim about a larger population. Think carefully about what population you are testing with your sample.

Project Report

Your project report will document what you did, and it is the basis for your grade. Treat this like any other important presentation: proofread it to make sure you actually said what you mean to, check spelling and grammar, use technical terms correctly without “snowing” your audience, and so on. Staple or bind everything together in order; don’t waste your money on a report cover.

Students often ask how long the report should be. Answer: long enough to cover the things listed below, and no longer. Don’t try to pad it to make it look more impressive. You’ll find additional details in the Grading Rubric later.

Use the following format or similar, and label each section.

1. Introduction
   In a paragraph or two, describe the problem and state (in English) the claim or question that you tested.

2. Population
   Describe the population you’re making an inference about. How large is the population?

3. Sampling Procedure
   How large was your sample and how did you decide on that sample size? Describe your method of collecting data. How did you select your sample? What might be possible biases in your sample? (If you think the answer is “none”, say so; don’t feel you must make something up. Check your notes from Chapter 1 to be sure you know what “bias” means.)

4. Raw Data
   Tabulate your raw data neatly, with appropriate sample statistics. If you obtained data from a published source, also cite the source and include a copy, as explained above.

   Examples: For numeric data, show the actual data in some appropriate form such as a table or graph, with x̄ and s. For attribute data, show a count of each response to each question with p̂, but don’t list all the individual responses. Think back to Chapter 2 for appropriate methods of presenting data.

5. Hypothesis Test
   State the case number and description from Inferential Statistics Cases. Choose an appropriate significance level, and explain your choice. Show that requirements are met and show all six numbered HT steps.

   Show hypotheses in symbols and in English. (Cases 6 and 7 will be in English only.) Example:

   H₀: p = 0.1, 10% of car trips are made in a Subaru

   H₁: p > 0.1, more than 10% of car trips are made in a Subaru

   In the calculation section, draw the curve, shade the appropriate region, and label the features as we do in class. Include TI-83 complete input and output screens; if appropriate, mention what went into which lists.

   Be sure to answer the question or decide on the claim that you stated in part A.

6. Reflections (if appropriate)
   You may have further comments or reactions to what you found. For example, think back to Chapter 10 for a possible follow-up if your hypothesis test doesn’t find a significant result.

   Not every project report will need this section — if you have nothing further to say, omit this section.

7. Original Data Sheets
   Attach your original data sheets, no matter how messy they might be. Never rewrite data sheets to make them look neater. If your survey was on odd-sized slips of paper, put them in an unsealed envelope.

8. Project Plan and Survey Approvals
   Attach your approved Project Plan, either the paper signed by Prof. Marvel or a printout of the email approval. For a survey outside TC3, also attach your approval from the Office of Institutional Research.

Project Talk

For extra credit, on project day you may make a five-minute presentation to the class.

See the last section of the Grading Rubric for the points you need to cover in your presentation. You may use any notes or visual aids that you want, but visual aids aren’t required.

A signup sheet will be available at the previous class meeting. You must sign up then if you want to present your project to the class. Presentations will be in random order, not signup order.

Only students who have done their project according to an approved Project Plan may receive extra credit.

Grading Rubric