Incorporating a Semester-Long Student Led Project into a Statistics Course

Dan Kernler

Elgin Community College

This presentation:

https://tinyurl.com/DKIMACC2023

Google Drive folder: https://tinyurl.com/StatsDK

Project Overview

• Project Part 1: Variables & Questions
• Project Part 2: Demographics
• Project Part 3: Initial Analysis
• Project Part 4: Confidence Intervals
• Project Part 5: Hypothesis Tests
• Final Project Report

Data Sources

• LGB Life and Health
• Health Behaviors in School Aged Children
• Children of Immigrants
• 2018-2019 Illinois School Data
• 2016 National Household Education Survey
• School Survey on Crime and Safety

Part 1: Variables & Questions

1. Pick a data set to analyze.
2. Once your data set is chosen, answer the following questions:
• What are the “individuals” in this data set? In other words, what does each row represent?
• Why did you choose this data set? What interests you about it?
3. Identify five questions you would like to answer. They should be “comparison” questions – comparing groups of students, types of households, different areas of the country, etc. You should include both quantitative and qualitative questions.
4. Identify two previously-published studies related to your questions in some way.

Sample Student Responses

I love the openness and nonchalance of Generation Z regarding sexuality. Having grown up with gay and bisexual family coming from the Silent and Boomer Generations, as well as many friends from my own Gen X group, I know their life experiences have been far different and frankly, quite heartbreaking at times. I want to know this younger generation’s demographics and see how it stacks up against those who came before them.

Sample Student Responses

My five questions:

1. What is the relationship between age and life satisfaction?
2. What is the relationship between smoking tobacco or marijuana and age?
3. What is the relationship between smoking tobacco or marijuana and gender?
4. Are life satisfaction and thoughts on body related?
5. What is the relationship between days physically active and overall health?

Part 2: Demographics

1. Describe your sample with the sample size in a complete sentence.
2. Pick 5 demographic variables that describe your sample.
3. For each variable chosen, calculate a measure of center. If the variable is quantitative, calculate a measure of spread as well.
4. For each variable chosen, create a relative frequency table. Edit your table to make it legible.
5. For each variable, create at least one graph. Be sure your graph is clearly titled and the axis labels are legible to a reader who may not be familiar with your data.

Sample Student Responses

Part 3: Initial Analysis

1. Find the five questions you included in Part 1 of your project, and copy them to this new document. You will focus on three of them to start to answer.
2. Give each question a separate section in your document. Each section should include the following:
1. A clearly stated question.
2. A hypothesis (educated guess) of what you think the answer will be.
3. Initial attempts to answer the question using at least two methods. Those methods can include statistics, tables, or graphs.
4. A summary of your results in paragraph form.

Sample Student Responses

When I was writing this question, I �wanted to discover if the urbanity of a school was related to whether or not �they had armed guards. I initially �thought the more urban a school was, �the more likely it would have armed guards. The bar graph and contingency table above confirmed my hypothesis. Schools in cities are more likely to have security guards than institutions in a suburb, town, or rural area. The percentage of rural schools that have armed guards and schools that do not are almost equal. This is quite different for schools in other areas.

Sample Student Responses

Originally my question was what the correlation between age coming out to family and levels of happiness was because I assumed that the younger the happier. My hypothesis was proven to be incorrect �because the average age for very happy seems to be a little bit older than those of pretty/not too happy. However, the age range was quite close to one another, yet it was shocking to me that it was 19-21 years old rather than a younger age because I �assumed that the younger, the more time to truly be themselves.

Part 4: Confidence Intervals

1. Find the three questions you attempted to answer in Part 3. You’ll focus on those same 3 for this part. Pick two of those questions that are either comparing means between groups or use a contingency table.
2. Give each question a separate section in your document. Each section should include the following:
1. A clearly stated question.
2. Your hypothesis (educated guess) of what you think the answer will be.
3. As check as to whether or the conditions are met in order to compute your specific confidence interval. State these conditions clearly in your summary.
4. Results for confidence intervals for each of your questions.
5. A summary of your results in a paragraph for each question.

Sample Student Responses

I feel like girls life satisfaction will be all around worse than guys life satisfaction because I feel guys can get away with more and aren’t held to such a high standard. I wouldn’t be surprised if guys' life satisfaction was significantly higher than girls.

Comparing the averages for life satisfaction by gender and creating a confidence interval for each gender we can see that the two intervals do not overlap. This shows that there is a statistical difference between the two genders on the basis of life satisfaction. Although the numbers don’t seem too far off, we have to realize that any statistical difference in the happiness of different groups of people is significant in life and tells us a lot about that group. Although the difference in life satisfaction isn’t as large as I thought it would be, this still backs up my hypothesis and shows that males' life satisfaction is higher than females.

Part 5: Hypothesis Tests

1. Find the two questions you attempted to answer in Part 4. You’ll focus on those same two for this part.
2. Give each question a separate section in your document. Each section should include the following:
1. A clearly stated question.
2. Your hypothesis (educated guess) of what you think the answer will be. This should be the same hypothesis you included for Part 4.
3. A check as to whether or not the conditions are met for your specific hypothesis test. Clearly state these conditions and the results in your summary.
4. Results for hypothesis tests for each of your questions.
5. A summary of your results in a paragraph for each question.

Sample Student Work

I wanted to figure out if there was a relationship between life satisfaction and whether the parent/guardian was loving. I thought if a parent/guardian is loving then the child's life satisfaction would sit higher than those whose parent/guardian wasn’t.

Children with an almost always loving parent/guardian are more likely to have higher life satisfaction. If children have a parent/guardian who sometimes or almost never is loving, their life satisfaction doesn’t drop that low. With a small p-value of < 0.0001, we can say that the mean life satisfaction is not the same for all levels and that there may be somewhat of a relationship.

Although, since the variances do not appear to be equal, we should be careful about drawing too many conclusions from these results.

Sample Student Work

Because we are comparing means among three groups, a One-Way ANOVA test would be the best vehicle to test the differences between the means of the three groups. However, only one demographic (18-26) passed the criteria of being normally distributed in order to conduct the test. As an alternative, a Two-Sample Independent Means hypothesis test comparing each group to one another via two at a time was conducted in its place.

Sample Student Work

At face value, knowing that younger generations realize their sexuality at a different age than their older counterparts can have many positive indications (are people more self-aware than they were 40 years ago, or has society’s views on sexuality evolved enough where people have the space to acknowledge a sexuality outside of hetero?).

Final Project Report

1. Title�Include the title of your project, your name, and your institution (i.e. Elgin Community College)
2. Population, sample details, project questions�• What population did you study? �• What was your sample size?�• What questions did you ask?
3. Charts and Tables�Include at least one chart and at least one table that describe the demographics of your sample.
4. Inferential Statistics (1 or 2 slides)�Summarize your inferential statistics from Parts 4 and 5 in a simple-to-read format.
5. Summary�This is the most important slide! What were the key results from your project? You should have 2-4 key points. Focus on clearly summarizing the statistical and practical significance of any results. 
6. Lessons Learned�What did you learn about statistics and/or your topic of choice as part of this project? Expand a bit here and share about your experience doing the project.

Sample Student Work

Let’s try one!

Part 1: Variables & Questions

1. Pick a data set to analyze.
2. Once your data set is chosen, answer the following questions:
• What are the “individuals” in this data set? In other words, what does each row represent?
• Why did you choose this data set? What interests you about it?
3. Identify five questions you would like to answer. They should be “comparison” questions – comparing groups of students, types of households, different areas of the country, etc. You should include both quantitative and qualitative questions.
4. Identify two previously-published studies related to your questions in some way.

Investigate the available data sources and their corresponding variables. When you’re ready, vote in the poll.

Poll:

Data:

https://tinyurl.com/IMACC23data

Part 1: Variables & Questions

Now we need 5 questions we’d like answered. �Remember, focus on relationships between variables.

1. What is the relationship between the % B/AA and % OSS?
2. $ Total per pupil and SAT Math Average
3. SAT Reading low-income vs non-low-income
4. Average class size and teacher retention rate.

If you’re following along in StatCrunch, you can join our group here:

https://www.statcrunch.com/groups/view?groupid=16404

Part 2: Demographics

1. Describe your sample with the sample size in a complete sentence.
2. Pick 5 demographic variables that describe your sample.
3. For each variable chosen, calculate a measure of center. If the variable is quantitative, calculate a measure of spread as well.
4. For each variable chosen, create a relative frequency table. Edit your table to make it legible.
5. For each variable, create at least one graph. Be sure your graph is clearly titled and the axis labels are legible to a reader who may not be familiar with your data.

Choose 5 demographic variables from our dataset:

1. Enrollment
2. Title I status
3. District Type

Table and Chart

Part 3: Initial Analysis

1. Find the five questions you included in Part 1 of your project, and copy them to this new document. You will focus on three of them to start to answer.
2. Give each question a separate section in your document. Each section should include the following:
1. A clearly stated question.
2. A hypothesis (educated guess) of what you think the answer will be.
3. Initial attempts to answer the question using at least two methods. Those methods can include statistics, tables, or graphs.
4. A summary of your results in paragraph form.

Choose One!

Pick one question to focus on from our initial�list of 5. Take a couple minutes and discuss how we could attempt to answer this with statistics, a table, or a graph.

Our question:

Initial Analysis

Part 4: Confidence Intervals

1. Find the three questions you attempted to answer in Part 3. You’ll focus on those same 3 for this part. Pick two of those questions that are either comparing means between groups or use a contingency table.
2. Give each question a separate section in your document. Each section should include the following:
1. A clearly stated question.
2. Your hypothesis (educated guess) of what you think the answer will be.
3. As check as to whether or the conditions are met in order to compute your specific confidence interval. State these conditions clearly in your summary.
4. Results for confidence intervals for each of your questions.
5. A summary of your results in a paragraph for each question.

Confidence Intervals

How could we use confidence intervals to answer that same question?

Our question:

Confidence Interval Results

Part 5: Hypothesis Tests

1. Find the two questions you attempted to answer in Part 4. You’ll focus on those same two for this part.
2. Give each question a separate section in your document. Each section should include the following:
1. A clearly stated question.
2. Your hypothesis (educated guess) of what you think the answer will be. This should be the same hypothesis you included for Part 4.
3. A check as to whether or not the conditions are met for your specific hypothesis test. Clearly state these conditions and the results in your summary.
4. Results for hypothesis tests for each of your questions.
5. A summary of your results in a paragraph for each question.

Hypothesis Tests

What hypothesis test would be most �applicable to answer this same question?

Our question:

Hypothesis Test Results

Interested in More?

This presentation:

https://tinyurl.com/DKIMACC2023

Google Drive folder: https://tinyurl.com/StatsDK