This week’s discussion topic allows you to begin the first of 3 steps in complet

This week’s discussion topic allows you to begin the first of 3 steps in completing your Term Research Project for the course. The assignment will build upon the assignment that you began in Week 4, and you may use the same variables and basic hypotheses you created last week, unless I specifically informed you to make a change. As such, it may be in your best interest to wait until you receive your grade for the Week 4 Discussion before you begin this one.
Note that I generally will not make you change anything unless your original assignment was not correct, was exactly like another student’s, or if I predict that your dataset will not be sufficient to meet the instructions given for the Term Project. To meet the final criteria, your dataset much contain at least 75 complete observations with no missing data or outliers, all observations should be from individual countries, the observations should represent the average value from 2006-2015, and the final dataset must not contain any correlation values above .7 (or below -.7).
In Week 5, you will complete the first two sections in the Word Document attached below (Introduction and Purpose, Definition of Variables). Much of this follows the same instruction as in Week 4, and you are not required to write a literature review or go into any great detail. You are not being graded on how well your hypotheses predictions are researched. I only need you to be clear and direct so that I may grade you on how well you are able to test your hypotheses in the final term project. Then insert a graphic to further describe your hypothesized relationships and regression design.
The new material for this week will begin in Section 3, where you will describe your sample and explain your final dataset. Using the same parameters as last week, download a dataset from the World Bank Millenium Development Goals database containing at least 5 continuous independent variables and 1 dependent variable (GNI per capita, Atlas method unless otherwise instructed). Set the parameters to ensure that no country is included more than once, and that each observation represents a 10-year average for the span of 2006-2015.
Millennium Development Goals | DataBankLinks to an external site. https://databank.worldbank.org/source/millennium-development-goals
Use the Data Analysis Toolpak in Microsoft Excel to create a table for your descriptive statistics. Be sure to format the table so that it is neat, easy to read, and may be easily inserted into a Word document. I recommend creating shorter variable names to make your charts, tables, and graphs easier to manage. I show you how to do this in the Week 4 discussion video, and I will revisit those instructions this week. Create a separate tab for your original data with shortened names as explained in the Week 4 and videos. Following the instructions in the Week 4 and Week 5 videos, create a descriptive statistics table using the Data Analysis Toolpak for your first entry into Section 3. Be sure that every variable has greater than 100 observations.
Next, identify all missing values and use listwise deletion as a treatment for missing data (purge any observation with missing values for any variable using the Filter Function in Excel). Copy your filtered data into a new tab labeled “No Missing” using the Paste Values command. Create descriptive statistics for the second entry into Section 3. Follow the directions on the Weeks 4 and 5 videos to ensure that your data are formatted correctly. Be sure to document how many countries were removed (if your sample size is below 75, you must replace variables with too many missing values and start over).
Next, you will treat outliers. For simplicity, your treatment for extreme outliers will be to remove all observations greater than 3 standard deviations from the mean for each variable, and you are only required to use one iteration. Use your No Missing tab to identify and filter outliers using the mean and standard deviations of each variable. I will show you how to do this relatively quickly in this week’s discussion video. Once you complete this step, copy your filtered data into a new tab labeled “No Outliers”. Ensure that your sample size is still greater than 75 and create a descriptive statistics table for your third entry into Section 3. If your sample size after purging missing data and outliers is less than 75, replace one or more variables, then start over. If your sample size is still greater than 75, and you have no extremely high correlation pairs in Week 6, then this will be the descriptive statistics table that you use for the final term project.
The final step for this week is to create a correlation table for the variables in your project and insert it into the end of the descriptive statistics section. You will not be graded on whether or not your correlations are in range. Your only objective in this step is to provide a neat and easy to read table to prove that your data is sufficient to use in a multiple regression equation, which you will learn to perform in Week 6. At this point, you only have to show that all possible correlations are between -.7 and .7. You will learn why next week. To do this, use the Data Analysis Toolpak to create a complete correlation table, format the data and numbers to look neat and easy to read (I recommend using short variable names as mentioned above and rounding all correlations to 3 decimal places), and identify any correlations with an absolute value greater than .7. For any variables too highly correlated, you will have to remove and replace at least one of them and start over at the beginning. This is why we are doing the correlation table early, then completing the written variable and sample descriptions a week later. I do not want you to have to complete the written requirements for these sections multiple times.