The focus of this project is to download a data set and manipulate it to work around missing data.

a. First, download Data Set 3 “Body Temperatures” in Appendix B from www.TriolaStats.

b. Some statistical procedures, such as those involved with correlation and regression (discussed in later chapters) require data that consist of matched pairs of values, and those procedures ignore pairs in which at least one of the data values in a matched pair is missing. Assume that we want to conduct analyses for correlation and regression on the last two columns of data in Data Set 3: body temperatures measured at 8 AM on day 2 and again at 12 AM on day 2. For those last two columns, identify the rows with at least one missing value.

c. Here are two different strategies for reconfiguring the data set to work around the missing data in the last two columns (assuming that we need matched pairs of data with no missing values):

i. Manual Deletion Highlight rows with at least one missing value in the last two columns and then delete those rows. This can be tedious if there are many rows with missing data and those rows are interspersed throughout the worksheet instead of being adjacent rows.

ii. Sort Excel has a Sort feature that rearranges all rows using one particular column as the basis for sorting. See the procedure given under “Sorting Rows with Excel” in Section 1-4. The result is that all individual rows remain the same but they are in a different order. First use Excel’s Sort feature to rearrange all rows using the “8 AM day 2” column as the basis for sorting (so that all missing values in the “8 AM day 2” column are at the beginning), then highlight and delete all of those rows with missing values in the “8 AM day 2” column. Next, use Excel’s Sort feature to rearrange all rows using the “12 AM day 2” column as the basis for sorting (so that all missing values in the “12 AM day 2” column are at the beginning), then highlight and delete all of those rows with missing values in the “12 AM day 2” column. The remaining rows will include matched pairs of body temperatures, and those rows will be suitable for analyses such as correlation and regression. Print the resulting reconfigured data set.