Applications ExercisesResearch in Psych, 7e: Study Guide, Chapter 99-1

9.1. Correlations and the Nature/Nurture Issue

As a researcher, suppose you are interested in the extent to which biological factors are associated with shyness in females. Assume you have access to 20 adopted children, all females, and you give each a shyness test. The test yields scores from 0-20 and the higher the score, the greater the degree of shyness. You also test the biological mother and the adoptive mother for each child. Create some data and calculate the appropriate correlations that would lead you to conclude that (a) shyness is at least partially inherited, or (b) shyness seems to be primarily a function of the environment. The correlation can be calculated by following the guidelines in the online statistics routines or by using SPSS.

9.2. Assessing the Effect of Outliers

The following data were taken from a laboratory exercise I frequently use in my research methods course. Students complete one trial each of two different hand-eye coordination tests. The Purdue pegboard requires students to build a series of small objects using small washers and pegs. The Grooved Pegboard involves fitting pegs into a pegboard according to the certain orientation.

For the following data, calculate Pearson’s r and create a scatterplot, either by hand or via SPSS. Then remove the data for participant # 12 and repeat the calculations. What do you conclude?

N.B. Each of the numbers represents the time (in seconds) that it took one of my students to complete the task.

ParticipantPurdueGrooved

15434

28039

36540

47037

56033

65837

76238

85133

95937

107642

116635

129035

135333

147944

156940

165230

175027

Answers

9.1 Correlations and the Nature/Nurture Issue

If there is an inherited side to shyness, the correlation between shyness scores for the children and the biological mothers should be higher (+.40 perhaps) than the correlation for children and the adopted mothers (+.10 perhaps). If environment is the key, the opposite should be the case.

9.2 Assessing the Effect of Outliers

The exercise shows that one unusual score can have a noticeable effect on the size of a Pearson’s r, especially when there is a relatively small amount of data. When all of the data are included, Pearson’s r is +.66; when the person whose trouble with the Purdue pegboard is removed (the outlier), the Pearson’s r increases to +.84. So the effect of just one outlier reduced the correlation from +.84 to +.66, a substantial decrease.

- 1 -