- Most mean differences between groups are small, rarely
exceeding one-half of a standard deviation, especially when
the groups being compared are similar in age and gender com-
position. This is to be compared with the large within-group
differences. - Many characteristics are correlated. For such characteristics,
prediction from one to another is easy and precise. - However, many characteristics are uncorrelated or only
slightly correlated; for these characteristics, prediction from
one to another is poor to nil. - Some characteristics are prized by society, especially “social
outcome” variables such as success in school or work. These
variables may be predicted from combinations of other
individual-differences variables. - Prediction of social outcome variables from person variables
is rarely high, often moderate to low, usually less than 0.50
correlation.
Researchers in the field of the psychology of individual differ-
ences have studied differences between groups intensively, espe-
cially differences between the genders, among age groups, among
racial or ethnic groups, and among socioeconomic groups. Most of
the differences reported between groups are mean differences—an
important detail often glossed over—and most of these mean dif-
ferences are small, less than half a standard deviation. A few re-
ported mean differences of up to one standard deviation have been
the object of much controversy (for example, in the IQ difference
between black and white groups, in gender differences in spatial
ability, favoring males, and in the ability to perceive detail, favor-
ing females). Because of sampling problems and, in many cases,
problematic data-gathering and recording techniques, most of these
reported group mean differences are unreliable (that is, they have
large error terms). The heated controversies of the day are based on
bad data!
456 CAREER CHOICE AND DEVELOPMENT