In the first class meeting of ANT192, "African Ethnographies," I asked the 34 students to freelist their responses to the frame "When I think of Africa, I think of ____________." All responses were anonymous, but I asked the students to list their year of graduation and gender. The average list was 7.8 items long and the longest had 28 items
I analyzed the data using ANTHROPAC, and then "cleaned" the data to combine substantively identical answers (e.g. "cultural diversity" and "diverse cultures" were edited to be counted as a single item).
original data
cleaned data
ANTHROPAC can sort results by Smith's S, a statistic that reflects the cognitive salience of items by taking into account both the frequency and the position of an item on the list--i.e. an item which is usually at the top of a list will have a higher Smith's S than an item which appears the same number of times but which is usually towards the bottom of a list. Here are the cleaned results sorted by Smith's S (RESP is the # of times the item appeared, PCT is the percentage of respondents listing the item, Avg. Rank is the average value of where the item appeared on the list).
Update (July 2007) - this data may also be viewed as tag cloud
In the following screenshot, I've color-coded the results based on a classification of the answers (the color-coding was based on my judgment, not on evaluations elicited from the students):
- bright red items focus on negative aspects of contemporary Africa
- light red items focus on negative aspects of Africa's past
- yellow items focus on aspects of the biophysical environment
- light green items focus on people in Africa in ways that are not necessarily negative or positive
- bright green items focus on positive aspects of people in Africa
I then broke out the data separately by first-year students (N=15) and upper-class students (N=17). One response didn't have a year of graduation and was discarded. The color-coding in the screenshots is only red for contemporary negative and orange for past negative.
First-year students:
Upper-class students:
The proportion of responses with negative connotations was roughly the same in the two subsamples (29% for first-year, 26% for upper-class, and 28% for the sample as a whole), although negative perceptions are more salient for upper-class students.
Return to the course home page