BUS 308 Week 4 Lecture 3 Developing Relationships in Excel Expected Outcomes After reading this lecture, the student should be able to: 1. Calculate the t-value for a correlation coefficient 2. Calculate the minimum statistically significant correlation coefficient value. 3. Set-up and interpret a Linear Regression in Excel 4. Set-up and interpret a Multiple Regression in Excel Overview Setting up correlations and regressions in Excel is fairly straightforward and follows the approaches we have seen with our previous tools. This involves setting up the data input table, selecting the tools, and inputting information into the appropriate parts of the input window. Correlations Question 1 Data set-up for a correlation is perhaps the simplest of any we have seen. It involves simply copying and pasting the variables from the Data tab to the Week 4 worksheet. Again, paste them to the right of the question area. The screenshot below has the data for both the question 1 correlation and the question 2 multiple regression pasted them starting at column V. You can paste all the data at once or add the multiple regression variables later (as long as you do not sort the original data). Specifically, for Question 1, copy the salary data to column V (for example). Then copy the Midpoint thru Service columns and paste them next to salary. Finally copy the Raise column and paste it next to the service column. Notice that our data input range for this question now includes Salary in Column V and the other interval level variables found in Columns W thru AA. Question 1 asks for the correlation among the interval/ratio level variables with salary and says to exclude compa-ratio. For our example, we will correlation compa-ratio with the other interval/ratio level variables with the exclusion of salary. Since compa-ratio equals the salary divided by the midpoint, it does not seem reasonable to use salary in predicting compa- ratio or compa-ratio in predicting salary. Pearson correlations can be performed in two ways within Excel. If we have a single pair of variables we are interested in, for example compa-ratio and performance rating, we could use the fx (or Formulas) function CORREL(array1, array2) (note array means the same as range) to give us the correlation. However, if we have several variables we want to correlate at the same time, it is more effective to use the Correlation function found in the Analysis ToolPak in the Data Analysis tab. Set up of the input data for Correlation is simple. Just ensure that all of the variables to be correlated are listed together, and only include interval or ratio level data. For our data set, this would mean we cannot include gender or degree; even though they look like numerical data the 0 and 1 are merely labels as far as correlation is concerned. In the Correlation data input box shown below, list the entire data range, indicate if your dat ...