Quantitative Project 1: Using a publicly available dataset, you will develop and test a causal research question on an exposure and outcome of your choosing. The project will include the following requirements Part 1:
A) Develop a testable causal research question (3 points)
B) Part 2: Conduct an analysis to a) compare exposure groups on potential confounders using statistical software and report relevant statistical tests (for example linear regression etc ), and summarize in a Table 1 (3 points);
C) test your hypothesis using a multivariable linear or logistic model using statistical software and summarize in Table 2 (3 points).
D) You will present your work to your classmates via a 4 to 6-slide PowerPoint ( title slide, an intro/background slide, a slide on your analysis process, a slide on your current findings , and lastly, a slide for any references). See parts a, b and d of this assignment regarding expected results you should add to your presentation.
More Information to consider:
For choosing a dataset to work with, we allow flexibility for those who have a particular dataset in mind; however, do not worry if you don’t have a clue. There are many datasets that ship with R and particularly with the “openintro” package
You can see the list of datasets within R through the command data(). Then learn more about what is in any particular dataset using the commands you learned in the initial labs. Once you activate a dataset, such as like data(nycflights), you can also see it listed along with other objects in the upper-right window in RStudio. Click on the dataset icon and it will open and look similar to a spreadsheet (note for large datasets, all the columns and/or rows may not be displayable but they are there).
NHANES is a particularly valuable dataset for public health research, which is also publicly available. R code has been provided in Blackboard within the Week 3 folder to help import data that has been downloaded from NHANES to your local computer. This is just one of many ways to import data into R, including directly downloading from a web source into R.
In any case, it is highly recommended that you first see what data is available to you before getting excited about a topic only to discover you cannot identify a relevant dataset to support your research.
When you submit part 1, please identify the following in your proposal:
-
What dataset you will analyze, ensuring you can open it in R?
-
What variables you will use and in what capacity (for example, is BMI the exposure or the outcome, etc) and list potential confounders, such as age and other demographic variables, confirming they are available in your dataset?
-
What is your hypothesis (what X variable “causes” change in what Y variable)?
-
Do you plan to use a linear or logistic regression and why (demonstrate that you understand how the outcome variable is formatted, such as continuous like “BMI” or binary like “obese or not”).