PSYC-1115 Applied Epidemiology and Statistics Formative Mini-Assessment, UK

PART 1. Variable Transformations

1. Create your health indicator variables. Currently, your dataset doesn’t include variables for diabetes status, BMI (or BMI categories), or a high blood pressure flag. Please create these three health outcome variables using the code in this week’s .do file. 

• Continuous BMI should be calculated as (weight in kg) / (height in meters * height in meters)
• Categorical BMI should follow the following cut-points:
•  Underweight if BMI is less than 18.5
•  Normal weight if BMI is greater than or equal to 18.5 AND less than 25
•  Overweight if BMI is greater than or equal to 25 AND less than 30
•  Obese if BMI is greater than or equal to 30 (but non-missing)

2. Descriptive Statistics: Table & Narrative Description 

Instructions: This is the babytown version of what you’ll do in the technical report. Create a traditional “Table 1” using the “table skeleton” below to guide you. Basically, figure out what counts and percentages go in each of the cells that are highlighted and fill in the right number. Maybe goes without saying but: DO NOT COPY TABLE DIRECTLY OUT OF STATA (not for this formative and definitely not in the summative assessment). 
After you fill in the numbers, USE YOUR WORDS TO DESCRIBE the three populations that you’ve identified (All Participants, People with Diabetes, People without Diabetes) based on their characteristics below, making special note of differences between the population with diabetes and without diabetes. A small paragraph should be enough. Don’t overdo it, just practice. 

3. Descriptive Statistics: Data Visualization & Narrative Description

Instructions: Create two plots. First: a double boxplot graph of BMI by gender (i.e., two boxplots side by side in the same graph). Second: a histogram of age with a normal plot overlay. After you create your plots, DESCRIBE the plots (2-3 sentences per plot). What do you see? What do you notice? You are allowed to copy the graphs that you make directly out of Stata. 

Add your plots (COPY THE PLOTS OUT OF STATA) and then write 1-2 sentences DESCRIBING what the plots are showing.

4. Inferential Statistics: Hypothesis Testing – Writing out your hypotheses

Instructions: I want you to test the following associations for statistical significance: 1) numerical, continuous BMI and gender, 2) numerical, continuous blood lead levels and age group, and 3) race and diabetes. That’s THREE hypothesis tests. Start by writing ONE sentence describing what variable TYPES you’re working with and then write out your hypotheses (H0 and Ha) for each test. HINT: DIFFERENT HYPOTHESIS TESTS HAVE DIFFERENT NULL HYPOTHESES!

Hypothesis Test 1: Examine the relationship between Gender and BMI

1) Write one sentence on what types of variables you’re dealing with and (if categorical) how many categories it has. 
2) Write one sentence explaining what statistical test you will use. 
3) Then write out a null hypothesis and the alternative hypothesis of the hypothesis test you will conduct to test the association between gender and BMI (in the empty space below). 
Our first hypothesis test explores the relationship between Gender , a ___________ variable (with ____ categories), and BMI, a ____________ variable. 
This relationship will be explored using a ___________________ test. This test was chosen based on _______________________________ . 

Inferential Statistics 1: Examine the relationship between Gender and BMI
Describe what hypothesis test you used, create your results table, summarize and INTERPRET your results. I’d like to see something about what you decided to do about the null hypothesis here.

Hypothesis Test 2: Examine the relationship between Age Group and Blood Lead Levels
Write one sentence on what types of variables you’re dealing with and (if categorical) how many categories it has.

Inferential Statistics 2: Examine the relationship between Age Group and Blood Lead Levels
Describe what hypothesis test you used, create your results table, summarize and INTERPRET your results. I’d like to see something about what you decided to do about the null hypothesis here. 

Hypothesis Test 3: Examine the relationship between Race and Cholesterol as a categorical variable
One sentence on what types of variables you’re dealing with and (if categorical) how many categories it has. 

Inferential Statistics 3: Examine the relationship between Race and Cholesterol as a categorical variable
Describe what hypothesis test you used, create your results table, summarize and INTERPRET your results. I’d like to see something about what you decided to do about the null hypothesis here.