According to this study, you probably think you're good at recognizing people's faces. But are you really?
This report from NPR suggests that you might not be as skilled as you think. The report begins with stories of two individuals at each end of the spectrum: One person who is remarkably good at recognizing faces, and another person who is not.
Then we hear about a study of people's general ability:
"I think nobody really knew until the last few years just how bad we all are with unfamiliar faces," says Mike Burton, a professor of psychology at the University of York UK. Burton has run a number of facial recognition studies and has concluded that most people are remarkably bad at recognizing the faces of those they know only slightly. And to make matters worse, most people think they are good at this skill when they are not.
To understand more, Burton ran an experiment in which participants were given pairs of photos and asked whether the faces in those photos were of the same person – in essence, a matching game. Some of the faces were familiar to the participants and some were not. Sure enough, people tended to perform better at this task when the faces were familiar. They had a much harder time identifying pairs of unfamiliar faces.
a) In the simple experiment above, what is the independent variable? (What did Burton vary?) Was this IV manipulated as independent groups or within groups?
b) In the experiment above, what is the dependent variable?
c) Given your answers to a) and b), what kind of experiment is this: Posttest only? Pretest/posttest? Repeated measures? Concurrent measures?
d) Sketch a graph of the result that "people tended to perform better at this task when the faces were familiar"
The NPR report explains how the researchers dug deeper, to see if experience with faces (for example, as a passport officer) would make a difference in people's accuracy. He ran the same experiment, adding a participant variable to create a factorial design. As you read this description, look for the IV and PV.
"We did some work with passport officers last year where we showed that even passport officers find this a very difficult task and are often inaccurate."
Burton says experience does not seem to improve ability. He studied passport officers who had been on the job for 20 years. They were not necessarily any better than fresh recruits.
e) What was the IV in this design? What was the PV?
f) What is the DV?
g) Remembering that this is a factorial design, sketch a graph of the result that "passport officers who had been on the job for 20 years... were not necessarily any better than fresh recruits."
Suggested answers
a) The IV was whether people were familiar or unfamiliar; it was manipulated as within groups.
b) The dependent variable is the number of correct matches.
c) This was a repeated-measures design
d) The y-axis should read, 'Accuracy' and the x-axis should have "familiar and unfamiliar" categories. The bar graph should have the "unfamiliar" accuracy level to be very low.
e) The IV was (probably) whether people were familiar or unfamiliar, and the PV was level of experience (Fresh recruits vs. 20 years of experience)
f) The dependent variable was (probably) the number of correct matches.
g) Here's a suggested sketch. However these data are fabricated--you'd have to check the actual paper for the actual design and results.