Testing a hypothesis with data requires identifying a series of relations about what might be happening. In this paper, you will be using an existing dataset in this case from the Pew Research Center to test a hypothesis and write a journalistic piece about it. In practice, there are two parts to this project. You will need to test your hypothesis with data and you will need to write about your hypothesis test in a paper.
In this paper, it is OK to collaborate with others in the class when testing your hypothesis with the data (i.e., you can run the tests together), but you will need to write up your results entirely on your own.
To test your hypothesis with the data, you will need to do a few tasks:
– Choose a complex hypothesis (mediation, moderation, or spurious relation) that you are interested in testing with the data. Note that you will need to select three (3) variables to accomplish this.
– Determine what set of tests would be needed to identify whether your hypothesis is true.
– Run each of these tests and get a p-value for them using the approaches we learned in lectures and labs (Lab 11 and Lab 12).
After you have run your hypothesis test, you will be writing up a news-style article based on your results. Here, your goal is to not only report on whether the hypothesis appears to be true, but also to provide the necessary information to contextualize what was done to test the hypothesis. As in your other papers, make sure to put the most important results up front and to provide contextual information later. Your article should be between 600 and 1200 words (3-6 double-spaced pages).
Here are some things to make sure you mention in the piece:
– What did you find?
– What was the hypothesis tested?
– How can the results be generalized? (What does the sample tell you about that?)
– What were the actual measures used?
– How were the variables coded for analysis?
– What were the statistical tests that were used?
As in your other papers, make sure you:
Come up with a strong lede (opener)
Inform readers with the nut graf (key point that comes just after)
Make sure to be clear about what you find
Use the inverted pyramid format (take-away up front, then increasing details going down).
Be sure to include key information about:
Where the sample(s) come(s) from
What the important concepts are for the study, being clear about how they were conceptualized, and how they were operationalized
What operationalization was done (be specific about how things were measured)
What statistical tests were used
What was found
Who, if anyone, the results of the study could be generalizable to
Any major limitations readers should be aware of with respect to the survey
Your paper will be graded based on:
How well you write about what you find
Appropriate use of descriptive information
Your ability to identify and discuss key concepts and how things were measured and otherwise operationalized
Your ability to highlight some of what the study does and does not tell us
Your ability to conduct and write about the statistical tests you ran
No inter-item reliability test is required unless you are trying to build an index. No data visualization is required but it is a good way to present your findings.
I am also posting a few notes on common issues that may be helpful as well:
When using data with labels instead of numbers, you may need to use the “COUNTA” command instead of the “COUNT” command in pivot tables.
If you want to run a regression with labeled data, you will need to recode the labels to numbers. You can do this a few ways. I tend to just sort all my data by the column I want to relabel and make a new column by copying numbers down until each label change. The SUBSTITUTE command also works for this.
If you are using the data without the updates, you may find it helpful to use the “Filter” option in the pivot table to filter out any labels that shouldn’t be present.
Remember, we are not interested in whether you are able to do these tests perfectly, just that you demonstrate understanding of what you are getting (and discuss any limitations or other issues in your paper).
Last Completed Projects
| topic title | academic level | Writer | delivered |
|---|