Topic: Do Men and Women Differ in the Way They Remember Location Information?
Statistical Concepts Covered: In this applet you will learn more on how to interpret charts, especially misleading charts, and distinguish between significant and chance differences.
Introduction:
The study of cognitive differences between men and women has been a fertile ground for researchers over the past 20 years. In this chapter you read about the types of encoding processes people use to create enduring memories. Are gender-based differences in memory rare? In 1992, Silverman and Eals investigated gender differences in object location memory. Participants were asked to study an image containing several objects. Then, after a short delay, a new image was presented. In this image, some of the objects were in the same locations as in the original image, but others’ locations had been swapped with other objects. The goal of the task was to determine which objects had been moved. (If you would like to take the experiment yourself, you can find an online version of the task at the Online Psychology Laboratory.) Silverman and Eals’ results showed that women outperformed men at this task.
The applet for this unit will allow you to explore results from the online version of the experiment linked above. Experimental data from 100 male and 100 female college students will be presented for analysis. You will be able to analyze the data from multiple perspectives, some of which may suggest clear differences, and others may be harder to interpret. You might be surprised by how much you can influence interpretation of the data just by changing the layout of the graph. You will also investigate the role of practice on performance and try to determine just how large a difference between groups needs to be before you are comfortable saying that a true and reliable difference really exists.
1) Plot the data for the number of objects correctly marked as moved (CrMv) across all five trials. Which of the trials shows the largest difference between the genders for this measure? (Cycle through the various Trials in the dropdown, starting with “Trial 1”, and compare the scores for men and women.)
A. |
B. |
C. |
D. |
2) How would you characterize the men’s and women’s results as they move through the five different trials? (Use the graphs from question 1 or select “Men – Trial by Trial” and then “Women – Trial by Trial” from the groups to display.)
A. |
B. |
C. |
D. |
3) Now look at other measures of the participants’ performance on the memory task such as the number of items that were incorrectly marked as moved (InMv), or the number correctly marked as stationary (CrSt). What is the best explanation for the trends that you see across trials? (Select the “Trial by Trial” group to see the data collapsed across sex for each trial and response type.)
A. |
B. |
C. |
D. |
4) If you were designing a new memory experiment involving a task that most participants were unfamiliar with, which approach do you believe would generate the most accurate representation of their abilities?
A. |
B. |
C. |
D. |
5) Now instead of looking at all five trials separately, plot the average number of objects correctly marked as moved (CrMv). The new graph should have one bar for the men’s score and one for the women’s score and the y-axis should go from zero to twelve. How would you describe the object location memory differences between men and women? (Select the “Gender Differences - CrMv” group.)
A. |
B. |
C. |
D. |
6) Using that same data (CrMv) as question 5, we will change the y-axis so that the range only goes between 10 and 12. If you presented this graph to a classmate who had not seen the previous plot, how would he/she categorize the difference between men and women on this task? (Select the “Gender Differences – CrMv - Restricted” group.)
A. |
B. |
C. |
D. |
Statistical Lesson. If you play with the range of the y-axis, you can make the difference between the men’s scores and the women’s scores appear to be much larger or smaller, but that does not change the actual numerical difference between the data values. When creating graphs of your data, you should strive to make them as unbiased as possible, which includes selecting a y-axis range that does not artificially inflate the perceived differences. Similarly, whenever you are looking at someone else’s data, you should get into the habit of checking the range of the y-axis so that you do not fall prey to a misleading graph.
7) Now plot the trial-by-trial data for the number of objects correctly marked as stationary or not moving (CrSt) and the number of objects incorrectly marked as moved (InMv). There are separate bars for each trial so you can see the practice effects. Which of the following issues is a problem with this graph? (Select the “Trial by trial” group.)
A. |
B. |
C. |
D. |
Statistical Lesson. In many of these graphs there is only a small numerical difference between the women’s scores and the men’s scores. In situations like this it can be hard to look at the graph and determine if the difference is significant – if it is the result of an actual difference between the two groups – or if the difference was just a random occurrence based on the specific individuals in the sample. One way that researchers try to clarify this is by adding error bars to their graphs. Error bars indicate the range of values that may reasonably represent the actual population value based on the data from the sample. Therefore, if the error bars around one value overlap the error bars for the second value, there’s a reasonable chance that the populations are not truly different. But if the error bars do not overlap the comparison value, then that is an indication that there may be significant differences between the groups.
8) Using the graphs from questions 5 and 6, review the average number of objects correctly marked as moved for men and women. Change from using the full range in the one graph to the restricted range in the other graph so you have a clear view of the difference between the two groups, and easily see the range of the error bars in the graphs. Which range of values represent the likely actual population values for men based on the data? (Select the “Gender Differences – CrMv” group, then select the “Gender Differences – CrMv - Restricted” group.)
A. |
B. |
C. |
D. |
9) Look at the error bars in the graphs from question 8. What additional information do the error bars provide that the data values themselves do not?
A. |
B. |
C. |
D. |
10) Suppose that you were asked to create a plot to show the results of an experiment and you wanted the graph to be as unbiased and informative as possible. Which of the following statements is true?
A. |
B. |
C. |
D. |