These exercises consider data from the 2005-2006 NHANES. The objective of this set of exercises is to perform a typical imputation and analysis session. The logical steps include examination of missing data, imputation of missing data values using a multiple imputation software tool of choice, and analysis of the imputed data sets using a companion software tool of choice capable of handling multiply imputed data sets. Begin by downloading the following subset of data from the 2005-2006 NHANES: c11 _ exercises _ nhanes.dta (available from the book Web site). Note that this data set is limited to adults 18+ years of age and those that completed the NHANES medical examination (n = 5,534). The variables used in the imputation and analysis are gender (RIAGENDR), body mass index (BMXBMI), race/ethnicity (RIDRETH1), age (RIDAGEYR), and systolic blood pressure (BPXSY1). The data set also contains the NHANES complex design variables and probability weight (SDMVSTRA, SDMVPSU, WTMEC2YR).
a. Examine simple descriptive statistics for these variables (means, proportions, ranges, and counts of missing values) keeping in mind that the full n is 5,334. Use a software tool of choice for this step.
b. Pay close attention to the types of the variables with missing data (continuous, ordinal, binary, or nominal) and the amount of missing data; that is, what percent are missing on each variable? Prepare a table including the type of each variable in the data set along with the missing data rate for that variable.