This week’s discussion topic allows you to continue with the second step in completing your Term Research Project for the course. In Week 6, you will complete the Quantitative Results, which should include your descriptive statistics as well as your correlation matrix from last week. Truly, if you did a really good job selecting your sample last week, this should not take that much time or effort. Once you have a suitable sample free of outliers and correlation issues, then you can test your hypothesized relationships using the same sample data to perform a multiple regression analysis.
The Week 6 video, along with your textbook assignments for this week should enable you to perform this calculation with ease. In the video, I go through all steps from downloading the data, through the descriptives, creating the correlation matrix, and testing the hypotheses. As such, the video is a bit repetitive. If you are confident about your work in weeks 4 and 5, then you should be able to skip ahead to the final 15 or 20 minutes to complete the regression.
For this assignment, complete the Week 6 Word Document. Complete all calculations using the Excel Data Analysis Toolpak, and save all your work on separate tabs as shown in the Week 6 Discussion Video. You may continue with the Week 5 documents and simply rename them so long as your last name is in the file name and all sections for this week are complete. You should also copy your Word document and paste it into the Canvas text box to make it easier for other students to review and comment. Be sure to upload attach both documents. If you have trouble uploading or inserting both documents, simply respond to your own post and attach them there.
First, if I notified you that you would need to make changes to your data selection, then select new variables accordingly. You should also clean up any mistakes you noticed in the written document if you have time. It will make the final week easier on you. Please keep all of your quantitative work on ONE excel book that you download from World Bank. I need to see the original data that you download, your cleaned data, your meta-data that was included in the download, and your calculations. This allows me to help you if you get off track. Next, document the total countries that were included for all series in your dataset. You do not have to list this number for each series since each may be different; just count the number of observations in your initial table. Then, filter out the missing data as I showed you in the videos. Once you have applied “listwise deletion” for all missing data, note how many were removed. Follow this up with your treatment for outliers. For simplicity, you might want to just address anything more than 3 standard deviations away from the mean. Note how many outliers here noted and how the sample was affected. Make sure that you retained at least 75 observations for your final sample size. If not, find replace variables with missing data with series that have higher response rates.
***Please note that you should not list every single country – I don’t want to read that any more than you want to write it. A simple statement like this one will be fine: “The measurement items were obtained from the Word Bank with 162 countries included in the dataset. Outliers and missing items were treated using listwise deletion. A total of 26 countries were purged due to missing data, reducing the sample size to 136. ***Make sure that you retained at least 75 observations for your final sample size. If not, find variables with a lot of missing data and replace with series that have higher response rates.
Finally, do the calculations for your correlation matrix in the Results section. You should write that first paragraph up and include the correlation matrix as a table for this week’s discussion. From here, you can easily go ahead and run your regression if you want, but you can wait until the final paper to format and write up the results and provide a brief discussion.
***You will want to do your calculation for descriptives and your correlation matrix before you start writing all this up. Make sure that you do not have to change variables due to low responses or multicollinearity. You don’t want to have to write all of this multiple times.