The Psychology of Gender 4th Edition

Methods and History of Gender Research 37

are more likely to be employed. A longitudinal

study may help to solve this problem. We could

measure both employment and health at one

time (Time 1) and then six months later (Time

2), as shown in Figure 2.3. If employment at

Time 1 is associated with a change in health

between Time 1 and Time 2 (depicted by line

a), employment is likely to have caused better

health. We can be even more confident of this

relation if health at Time 1 does not predict

change in employment between Time 1 and

Time 2 (depicted by line b).

Longitudinal studies help establish

causality and also help distinguishage

effectsfromcohort effects.Acohortrefers

to a group of people of similar age, such as a

generation. Let’s say that we conduct a cross-

sectional study of adult women in which we

find that age is negatively associated with

hours worked outside the home. Can we

conclude that women decrease the amount

of hours they spend in paid employment as

they get older? If so, this would be an age ef-

fect. Or, is it the case that older women work

fewer hours outside the home because they

have more traditional gender-role attitudes

than younger women? If so, this finding is a

cohort effect, an effect due to the generation

of the people. In a cross-sectional design, we

managers’ behavior that would be difficult to

control: the way the manager is treated by his

or her own boss, the nature of the manager’s

job (whether it involves working with oth-

ers or whether it involves technology), and

the number of employees a single manager

has. Would you be able to randomly assign

a manager to focus on technology with one

employee but focus on relationships with the

other employee? Because field experiments

do not have the same kind of controls over

behavior that laboratory experiments do,

they are more difficult to conduct and more

likely to pose threats to internal validity.

Cross-Sectional Versus Longitudinal Designs

Aside from conducting a field study, there is

another way to enhance the internal validity

of correlational studies. Recall that a correla-

tional study usually measures the relation be-

tween two variables at a single point in time.

This is not always the case. When a single

time point is used, we say the study iscross

sectional. However, we may measure the in-

dependent variable at one time and the de-

pendent variable later; this is alongitudinal

study. In a longitudinal study, there are mul-

tiple time points of study. Can we discern

cause and effect with a longitudinal study?

Remember, a key principle to establishing

causality is that the cause precedes the effect.

A longitudinal study helps establish causality

but does not ensure it. Let’s take an example.

We could survey a group of women from

the community to see if employment is re-

lated to health. If we conduct one survey at a

single point in time, we are conducting a cross-

sectional study. Let’s say we find a correlation:

Employment is associated with better health.

The problem is that we do not know if employ-

ment leads to better health or if healthier people

Health Health

Employment

Status ‘a’

‘b’

Time 1 Time 2

Employment

Status

FIGURE 2.3 Depiction of a longitudinal design

in which one can disentangle the causal relation

between employment and health.

M02_HELG0185_04_SE_C02.indd 37 6/21/11 12:19 PM