Statistical Methods for Psychology

probability in its common everyday usage, referring to the likelihood that some event will

occur. From this perspective it is logical to conclude that, because 67% of those with

HIV/AIDS contracted it from injected drug use, then if we were to randomly draw the

name of one person from a list of people with HIV/AIDS, the probability is .67 that the in-

dividual would have contracted the disease from drug use. To put this in slightly different

terms, if 67% of the area of the pie is allocated to IDU, then the probability that a person

would fall in that segment is .67.

This pie chart also allows us to explore the addition of areas. It should be clear that if

5% are classed as CSW, 7% are classed as CSW-cl, and 4% are classed as MSM, then

5 17 14 5 16% contracted the disease from sexual activity. (In that part of the world the

causes of HIV/AIDS are quite different from what we in the West have come to expect, and

prevention programs would need to be modified accordingly.) In other words, we can find

the percentage of individuals in one of several categories just by adding the percentages for

each category. The same thing holds in terms of areas, in the sense that we can find the per-

centage of sexually related infections by adding the areas devoted to CSW, CSW-cl, and

MSM. And finally, if we can find percentages by adding areas, we can also find probabili-

ties by adding areas. Thus the probability of contracting HIV/AIDS as a result of sexual

activity if you live in Eastern Europe or Central Asia is the probability of being in one

of the three segments associated with that source, which we can get by summing the areas

(or their associated probabilities).

There are other ways to present data besides pie charts. Two of the simplest are a

histogram (already discussed in Chapter 2) and its closely related cousin, the bar chart.

Figure 3.2 is a redrawing of Figure 3.1 in the form of a bar chart. Although this figure does

not contain any new information, it has two advantages over the pie chart. First, it is easier

to compare categories, because the only thing we need to look at is the height of the bar,

rather than trying to compare the lengths of two different arcs in different orientations. The

second advantage is that the bar chart is visually more like the common distributions we

will deal with, in that the various levels or categories are spread out along the horizontal

dimension, and the percentages (or frequencies) in each category are shown along the ver-

tical dimension. (However, in a bar chart the values on the Xaxis can form a nominal scale,

as they do here. This is not true in a histogram.) Here again, you can see that the various areas

of the distribution are related to probabilities. Further, you can see that we can meaningfully

Introduction 67

60.00

40.00

20.00

0.00

CSW CSW-cl IDU

Source

MSM Oth

Percentage

Figure 3.2 Bar chart showing percentage of HIV/AIDS cases attributed to different

sources

bar chart