Unit C Translation Activity

These activities are for practice only — there is nothing to turn in.

Every question so far has told you which model to fit. Research never does. This activity works the other direction: you are given a study in the words a researcher would use to describe it, and you have to translate it into a data frame, a formula, and a set of assumptions worth worrying about.

Work through each one in a small group if you can. There is usually more than one defensible answer — being able to say why you chose yours is the skill being practiced.

For each study below, answer:

  1. What might the data frame look like? Name the variables, give their types, and say what one row represents.
  2. What formula would you use to test the hypothesis with a linear model?
  3. Should you be worried about any of the linear model’s assumptions here?

Example 1 (Clinical / Health Psychology)

I read that doctors sometimes use the amount of beta-amyloid proteins in a patient’s cerebrospinal fluid (CSF) to help detect Alzheimer’s disease. I am interested to know if this amount is also related to patients’ cognitive functioning. To investigate this, I collect CSF from and measure beta-amyloid and working memory in 120 patients with different types of dementia. Assume that both measures are continuous.

Example 2 (Sports Science)

I want to help the tennis player I coach to improve her service return game. I suspect she struggles with faster serves and serves she tried to return backhand. To test this, I watch 70 of her recent matches and mark down the average speed of all the serves she received in that match, the percentage that she attempted to return backhand (as opposed to forehand), and the percentage that she successfully returned (as opposed to did not return).

Example 3 (Social Psychology)

I think that people at a bar will enjoy being approached by and talking with a stranger only if they find that stranger attractive, but my friend thinks that attractiveness has nothing to do with it. To test this, we observe people at a local bar and, after someone is approached by someone else, ask them if they knew that person (yes or no), how attractive they found that person (1 to 10), and how much they liked talking with that person (1 to 10). We end up with ratings from 20 participants being approached by 55 strangers and 30 friends.

Bring your own

Describe a study from your own area the same way — a short paragraph, no statistical vocabulary — and hand it to someone else to translate. Then compare what they wrote to the analysis you actually had in mind.