| Class 7 - Homework Assignment |
| Determine if obesity is a risk factor for hypertension. |
| Background: |
| A clinical study investigated the relationship between obesity and hypertension |
| The study enrolled 60 random subjects from a local population, using the natural sampling design. |
| (enrolling subjetcs without knowing their obesity or hypertension status in advance) |
| The scientists used the body mass index to classify the subjects as obese or non-obese. |
| The presence or absence of hypertension in the subjects was determined by measuring |
| the blood pressure under varying conditions. |
| Results: |
| The findings of the study were: |
| # of obese subjects with hypertension: | | | | | 14 |
| # of obese subjects without hypertension: | | | | | 7 |
| # of non-obese subjects with hypertension: | | | | | 12 |
| # of non-obese subjects without hypertension: | | | | | 27 |
| Total number of subjects: | | | | | 60 |
| Task 1: |
| In a new Excel file, create a 2 way frequency table for the results, including marginals. |
| (You don't need to create a summary table with the joint proportions) |
| (Tip: It's easier to do this in a new file, so you can keep this one side by side) |
| Task 2: |
| Using the chi-square test for association and Fisher's exact test in Minitab, determine |
| if the two variables (obesity and hypertension) are associated. |
| NOTE: For this task, use the screenshots from the posted Powerpoint file "Running A Chi-Square Test With Summarized Data In A 2 Way Table In Minitab" |
| Paste the Minitab results below the Excel table from task 1. |
| Write a few sentences drawing your conclusion. |
| Task 3 |
| NOTE: the calculations for this task are explained in the posted PowerPoint file "Calculating risk, odds, relative risk, and odds ratio.pptx" |
| Just stating whether the variables are associated doesn't seem a satisfactory result. |
| You also want to know by how much the risk for hypertension increases for an obese person. |
| Go over the Powerpoint file and learn how to calculate the following statistics |
| (all of the calculations are just divisions of numbers from your 2 way table. |
| Carry out the calculations under your 2-way table in your Excel file, using the respective formulas that reference the cells in the table. |
| Then you have a template for the future. You just need to change the numbers in the table and get all the other numbers. |
| | a) the risk for obese and for non-obese people to develop hypertension. |
| | As shown in the Powerpoint file, you can easily obtain this result from your 2 way table. |
| | Include the result underneath the table. |
| | b) the relative risk (or risk ratio) |
| | This is the ratio of the risk to develop hypertension for obese vs non-obese people. |
| | It tells you by how many fold the risk for hypertension increases in obese people, compared to non-obese people. |
| | The relative risk can also easily be determined from your 2 way table. |
| | c) the odds and the odds ratio |
| | As explained in the Powerpoint file, the odds to develop a particular disease are calculated slightly |
| | different than the probability to develop the disease. |
| | Calculate the odds for obese and non-obese people to develop hypertension. |
| | From there, calculate the odds ratio (in a similar manner than the relative risk). |
| | All of these calculations are also done using the 2 way table. |
| | Add the results from task 3 underneath your other results |
| | State your results in a sentence or two (do the odds and the relative risk increase…?…) |
| Task 4: Extra Credit (sounds complicated but is actually easier and faster than what you did up to now!) |
| Lastly, you want to know how well the relative risk and the odds ratio you calculated |
| from this study describe the whole population. |
| Here is where statistics comes in again. |
| Obtain the confidence interval for the relative risk and for the odds ratio. |
| These are complicated calculations, but fortunately there are programs that do them for you, |
| for example, the calculator on the website | | | | | http://statpages.info/ctab2x2.html |
| NOTE 1: Minitab and Excel don't include features or formulas for these calculations |
| NOTE 2: Statpages.info was created by the author of the book "Biostatistics for Dummies", an excellent resource |
| This online application calculates dozens of coefficients and statistics for 2 way tables, |
| including the confidence intervals for the relative risk and for the odds ratio that you need. |
| You enter your 4 numbers into the 2-way table at the top, select your confidence level and hit the compute button. |
| Don't be confused by some of the vocab and abbreviations in the table. It can be used for many different purposes. |
| Just enter the numbers exactLy how you have them in your 2-way table. |
| Unfortunately you can't chose ahead of time which calculations you want done. |
| The calculator performs all of the calculations all the time. |
| So you have to sift through the results and find what you need. |
| As a "side effect" you will again get the p value for the chi square test and Fisher's exact test that you calculated in task 2. |
| You can copy and paste these under the results of the minitab chi square and Fisher. They should match! |
| Tip: Don't bother trying to understand anything else from these results. |
| Just get your confidence intervals for the odds ratio and relative risk and move on. |
| These are very specialized statistics for particular situations (evaluating diagnostic kits for example or inter-rater reliability) |
| You would learn these when you are in a situation where this is needed. |
| FINISHING UP: |
| Add the confidence intervals from task 4 to your other results in your Excel file. |
| Express the results from task 4 in a sentence or two. |
| SUBMIT THE EXCEL FILE |
| As always, ask me if you have any questions or get stuck somewhere. |