Final Portfolio (Due May 6th - Extension available below)

The final portfolio is your opportunity to integrate the activities we’ve workshopped all semester. This is your opportunity to “show off” the techniques you’ve learned in class. In doing so, you will begin to formulate, and test, a hypothesis! The portfolio is worth 20% of your final grade.

Components of Final Portfolio

  1. Select Variables (5 pts)

  2. Identify Hypothesis and “Real World Scenario” (5 pts)

  3. Do File (10 pts)

  4. Log File (10 pts)

  5. Word Doc of Analytical Techniques (40 pts)

  6. 1 Pg. Interpretive Memo (30 pts)

Let’s look at these components

in more detail

  1. Variables of Interest (5 points)

    In your word doc, list your variables of interest (10-15 will do). Be certain that you have a few interval/ratio variables to work with. In Week 7 you established a list of variables of interest. If the list changes as you go, that’s no problem at all.

  2. Formulate a Hypothesis (5 points)

    You’ll begin at the “end.” You’ve had enough time to play with your variables that you are ready to form a hypothesis. Not a full hypothesis, but you’ll hypothesize that a relationship of some kind exists between 2 (or more) of your variables.

    I’ll refer to this as your “umbrella hypothesis” in the activities. It will be very very vague… as vague even as “something is going on with income, sexual activity, and political affiliation.” You’ll get more specific about your hypothesis for each test.

    In your word doc for each technique, you should identify your hypothesis, identify whether your hypothesis is directional or non-directional, and name your “null hypothesis.”

  3. Start to consider in what “real world” situation, your question would be consequential (See step 6… you may want to consider this while formulating your hypothesis).

  4. Create a Do File (10 points)

    Use the directions provided to properly format a Do File. All of the commands you include in your portfolio should be listed (in order) with any notes you have. Do not include interpretations in your do file.

  5. Create a Log File (10 points)

    When you have completed your portfolio (your Do File is finished, your commands are working, and your outputs are interpreted), create a new log file, execute your Do File in its entirety, and close the log. You will submit both your Do File and your log file alongside your word doc.

  6. In a Word Doc, Spotlight your Techniques (40 points)

    Ok now the fun stuff. You have essentially learned 10 techniques in this class, and you’ll spotlight all of them in your portfolio in order to address your hypothesis. There is a brief description below of these commands and how I’d recommend you use them. But note 3 things:

    1. All 10 commands should all appear (in the proper form) in your Do File (no notes necessary)

    2. All 10 respective outputs should appear in your log file

    3. If any commands are not relevant or necessary to your hypothesis (for example, if you do not need to recode and generate a new variable), you will simply run that command on one of your variables, interpret it for credit, and note that it is not relevant to your “study.”

In the word doc, you will spotlight each of the following techniques:

  • tab & sum (4 points)

    • Include single variable and bivariate options for “tab”

    • Include one “if” option for either “tab” or “sum”

    • (You may want to do this for a few variables, the ones most relevant to your hypothesis)

  • recode, gen (4 points)

    • Recode at least 2 variables

    • Ideally you will recode a variable you need for an upcoming command

  • histogram/barchart (4 points)

    • You’ve learned many variations on these two functions, so they should be beautiful. I’d like you to use the histogram function to evaluate the normality of 2 variables. And I’d like you to use the barchart command to visualize differences between groups relevant to your hypothesis. Try using the “over” option here.

    • Be sure to screenshot the graphs into your word doc, just as you would the other outputs.

  • sum, det (4 points)

    • You’ve already used the sum command, so for at least your most important 2 variables, use this command (with the details options) to speak to skewness and kurtosis. 

  • corr (4 points)

    • This should be directly connected to your hypothesis. Is there a correlation between your variables of interest in your sample?

  • ci means (4 points)

    • By this point you should be testing the relationship between 2 or 3 variables. So you may choose what makes sense, based on your hypothesis here. 

  • ttest (4 points)

    • Run and interpret two, univariate ttests first for each of your 2 variables of interest (ensure that they are appropriate, you need one that is interval/ratio). You will have to create a more specific hypothesis for this step based on what you know so far, a threshold that you are testing. Revisit the Week 8 assignment to review).

  • ttest , by (4 points)

    • Then run and interpret a bivariate ttest, to test for the difference (on variable A) between two groups in variable B. You may have to recode a variable for this. Again, ensure that you have selected appropriate levels of measurement. And it should serve your hypothesis. Revisit Weeks 8 & 9 for review. Again, this test may require that you make your hypothesis and null hypothesis more specific. Include these specific hypotheses in your interpretation.

  • anova (4 points)

    • Then run a one-way analysis of variance test, and a two-way analysis of variance test. They can be related (as mine were in the Week 10 activity) or unrelated. You may have to recode a variable for this as well. Again, ensure that you have selected appropriate levels of measurement, and make sure your analysis serves your hypothesis. Again, this test may require that you make your hypothesis and null hypothesis more specific. Include these specific hypotheses in your interpretation.

  • reg (4 points)

    • Finally you will run an OLS regression analysis. Again, ensure that you have selected appropriate levels of measurement, and make sure your analysis serves your hypothesis. Again, this test may require that you make your hypothesis and null hypothesis more specific. Include these specific hypotheses in your interpretation.

You will organize each technique in the format below. In your word doc, list the command, the output, and your interpretation of the output in a brief paragraph.

example portfolio mntlhlth.png

6. 1 pg. Interpretive Write-up (30 points)

  • The last component of your portfolio invites you to bring all of your analysis together.

  • You will produce a 1 page visually-compelling memo. It should include select graphs/charts/visual aids from the above work, and a written interpretation of your analysis. Imagine you are circulating your work to a boardroom of colleagues planning to implement your findings in their business, political campaign, hospital, or nonprofit organization.

  • The planned changes (based on your analysis) are dramatic and expensive, and it’s your job to (1) present your hypothesis, (2) interpret your analysis, (3) describe how this might be relevant to “work on the ground,” and (4) present your findings in a brief, visually-appealing memo.

 

Do file, log file, and word document should be submitted in a single email to carrolct@bc.edu on the due date. Your 1 pg. interpretive writeup should be included as the first page of that word document.

Deadlines

(We will commit our entire Zoom class to Portfolio Q&A on April 21st & April 28th)

April 23rd

“Recommended” deadline to have a draft of your portfolio completed. You will not submit at this time

May 6th

Due date for portfolios. I recommend that you aim to submit by the 6th but (see below)

May 12th

Given the season we’re in, I will accept portfolios for full credit through noon EST on May 12th