Research Scenario
Psychology researchers at LSU and Duke wondered if teaching people how to be more self-compassionate would affect their eating behavior (Adams & Leary, 2007). Researchers have observed that people on diets tend to over-eat significantly once they “mess up” (e.g. eat something they didn’t plan to eat or want to eat).
The researchers designed the following experiment:
Participants in the study were fed a donut.
Half of them completed a self-compassion exercise, while the other half did not.
Then all the participants were offered candy.
The researchers measured the amount of candy eaten in both groups.
The researchers believed that those who completed the self-compassion activity would not eat as much candy, because they wouldn’t be as susceptible to all-or-nothing thinking. Using the data below, test if those who completed the self-compassion activity ate the same or different amounts of candy compared to those who didn’t complete the self-compassion (SC) activity, using α=0.01.
Candy Eaten After Donut (in grams)
Control Group
(n1 = 28) SC Activity Group
(n2 = 28)
27 75 47 104 43 37 43 32
30 91 40 55 35 40 36 38
43 54 24 51 32 32 37 37
64 25 50 46 29 38 35 34
38 34 58 29 29 40 31 41
46 48 84 72 42 34 35 30
51 72 27 105 36 40 30 46

Initial impressions:

Do you agree with the researchers’ belief about self-compassion and eating behavior?

What do you observe about the data at first glance?

What kind of test is this? One or two sample?

State the null and alternative hypothesis in symbols and in words.
H0:
Ha:

Perform a hypothesis test.

Find your descriptive statistics (including means, standard deviations, standard error)

Find t-critical

Find t-calculated

What is your conclusion?

Do you retain or reject H0? How do you know?

What is the “so what” of this experiment? What did the researchers learn that might help people?