The Psychology of Gender 4th Edition

Methods and History of Gender Research 35

sex astimulusortarget variable, meaning it

is the characteristic of something to which

people respond. Let’s take the research ques-

tion “Do we smile more at male or female

infants?” One way to answer that question

is to compare how often adults smile at male

and female infants. However, this would be

a correlational study; we would be correlat-

ing infant sex with smiling, and sex would

be a subject variable. We would not know if

infant sex caused the smiling or something

else about the infant caused the smiling; for

example, infant girls are more likely to wear

pink and perhaps pink causes smiling. A bet-

ter way to address this research question is by

conducting an experimental study in which

infant sex is a target variable. We could show

people pictures of the same child dressed in

gender neutral clothes and randomly tell one

group the infant is Sam and the other group

the infant is Samantha. Then we can look to

see if people smile more at infants they per-

ceive to be female compared to those they

perceive to be male. When sex is a target vari-

able, random assignment can take place and a

true experiment can be conducted.

There are advantages and disadvantages

of both correlational and experimental meth-

ods. The major ones are identified in Table 2.2.

The advantage of the experimental

method is that cause and effect can be de-

termined because all other variables in the

experiment are held constant except for the

independent variable (the cause). Thus, any

differences in the dependent variable (the ef-

fect) can be attributed to the independent

students who sit on the right side have a view

of the wall, so they might as well look at me.

Imagine you had asked participants to decide

whether they wanted to play with the puppy

or teach it tricks. If you let people choose their

condition, the people in the two conditions

would be different; nurturant people are likely

to choose to play with the puppy. Differences

in nurturant behavior between the two condi-

tions would be due to a selection bias because

people selected their own groups and were

not randomly assigned to condition.

In a true experiment, one must be able

to manipulate the independent variable to

study its effects. Notice that some of the inde-

pendent variables in Table 2.1 are changeable

and some are not; that is, one can manipulate

employment, testosterone, and status to study

their effects. Other independent variables are

not changeable, such as sex, race, and ethnic-

ity. When sex is a characteristic of a person, as

in the research question “Are men or women

smarter?” sex is referred to as asubject vari-

able. Studies in which sex is a subject variable

are not true experiments because someone

cannot be randomly assigned to be female or

male. The majority of research that compares

men and women—evaluating similarities

and differences between men’s and women’s

behavior—is not experimental for this rea-

son. We observe in the laboratory or in the

real world how women and men think, feel,

and behave. This research is correlational

because we cannot manipulate a person’s sex.

Is there any way we can use an experi-

ment to make sex comparisons? We can make

TABLE 2.2 EXPERIMENTALMETHODVERSUSCORRELATIONALMETHOD

Experimental Correlational

Strength Internal validity External validity

Weakness External validity Internal validity

M02_HELG0185_04_SE_C02.indd 35 6/21/11 12:19 PM