The Psychology of Gender 4th Edition

106 Chapter 4

only whether a significant difference is found

in a study but also the size of the difference,

or theeffect size. The effect size, calculated

in terms of the “d statistic,” is calculated by

taking the difference between the means

[M] of the two groups (in this case, women

and men), and dividing this difference by

the variability in the scores of the members

of these two groups (i.e., the standard devi-

ation [SD]), as shown in Figure 4.2. As the

size of the sample increases, the estimate of

the mean becomes more reliable. This means

the variability around the mean, the standard

deviation, becomes smaller in larger samples.

A small difference between the means of two

large groups will result in a larger effect size

than a small difference between the means of

two small groups. Hence a study that shows

men score 10 points higher than women on

the math SAT will result in a larger effect size

if there are 100 women and men in the study

than if there are 20 women and men in the

study. The rule of thumb used to interpret

thedstatistic is that .2 is a small effect, .5 is a

medium effect, and .8 is a large effect (Cohen,

1977). A .2 effect size means that sex accounts

for less than 1% of the variance in the outcome;

a .5 effect means that sex accounts for 6% of the

variance; a .8 effect means that sex accounts for

14% of the variance (Cohen, 1977).

If a large effect accounts for only 14% of

the variance, is a small effect even worth dis-

cussing? As you will discover in this chapter,

the three cognitive differences that Maccoby

and Jacklin noted did not appear until adoles-

cence. Adolescence is sometimes referred to as

a time ofgender intensification, a time when

girls and boys are concerned with adhering to

gender roles. Thus sex differences that arise as

a result of socialization pressures might not

appear until adolescence. Even sex differences

thought to be influenced by hormones might

not appear until puberty. When Block catego-

rized the studies into three age groups (under

4, between 5 and 12, and over 12), she found

that sex differences in many domains became

larger with increasing age.

In the end, Block agreed with the sex

differences that Maccoby and Jacklin found

but also found evidence of other sex differ-

ences. She concluded that boys, compared

to girls, were better on insight problems,

showed greater dominance, had a stronger

self-concept, were more active, and were

more impulsive. Girls, in comparison to

boys, expressed more fear, showed more

anxiety, had less confidence on tasks, main-

tained closer contact with friends, sought

more help, scored higher on social desirabil-

ity, and were more compliant with adults.

The conclusions of Maccoby and Jack-

lin (1974) and of Block (1976) were obvi-

ously not the same. Both, however, relied

on narrative reviews of the literature. In the

1980s, a new method was developed to re-

view the literature that led to more objec-

tive conclusions: meta-analysis. Much of the

recent literature on sex comparisons, which

is described in this chapter, has relied on

Meta-Analysis

Meta-Analysis

Meta-analysisis a statistical tool that quan-

tifies the results of a group of studies. In a

meta-analysis, we take into consideration not

2

SD^2 males 1 SD^2 females

Mmales 2 Mfemales

d 5

Note: SD 5 Standard Deviation

FIGURE 4.2 Thedstatistic,

as calculated by this formula, is

used to determine the size of a

sex difference.

M04_HELG0185_04_SE_C04.indd 106 6/21/11 8:02 AM