 |
 |
National Center for Chronic Disease Prevention and Health Promotion
Behavioral Risk Factor Surveillance System
BRFSS Home | FAQs | Contact Us
|
||||||||
|
|
|
|
|
|
|
||
National Center for Chronic Disease Prevention and Health Promotion Behavioral Risk Factor Surveillance System BRFSS Home | FAQs | Contact Us |
|
|
||||||
|
|
Independent, Dependent, and Control variablesA dependent variable, also known as an outcome variable, is affected by one or more independent variables. Independent variables can include exposures or etiologic factors. For example, health is an outcome that may be affected by one or more independent variables, such as age, or diabetes. When setting up cross tabulations, analysts usually place the dependent variable (e.g., "General health") in rows. Independent variables (e.g., "Reported Age 18-64, 65+") are placed in columns. This generally makes interpretation of results easier. You may want to refine your analysis further by adding one or two control variables. Doing so is optional, but it produces additional tables with a single analysis. Control variables may be extraneous factors or confounds. They can help refine conclusions that you would otherwise draw from a given analysis. For example, you may be interested in learning how human health is affected by age, and whether the effect of aging is reduced by exercise. To answer this in the WEAT, you would select "General health" as the dependent variable, " Reported Age 18-64, 65+" as the independent variable, and "Exercised in past 30 days" as a control variable. You could even add a second control such as "Ever told have diabetes" to determine whether diabetes changes the effect of exercise on health among the elderly. When using the WEAT to perform logistic regression analyses, the dependent variables available to you all have dichotomous outcomes (e.g., at risk or not at risk). You can predict a given outcome by adding one or more dependent variables or predictors to the logistic model. The WEAT output will display an odds ratio for each predictor in the model. An odds ratio describes the relative risk associated with a particular level (e.g., male or female) for the given predictor (e.g., Gender). |
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
