Basically the same form of reasoning can be used to reach a universal conclusion. An example is the inductive inference discussed at the start of this chapter: All observed ravens are black, so all ravens are black. Again, we sample part of a population to draw a conclusion about the whole. Argu- ments of this form, whether the conclusion is universal or partial (as when it cites a particular percentage), are called statistical generalizations.
How do we assess such inferences? To begin to answer this question, we can consider a simple example of a statistical generalization. On various occasions, Harold has tried to use Canadian quarters in American payphones and found that they have not worked. From this he draws the conclusion that Canadian quarters do not work in American payphones. Harold’s inductive reasoning looks like this:
In the past, when I tried to use Canadian quarters in American payphones, they did not work.
Canadian quarters do not work in American payphones.
The force of the conclusion is that Canadian quarters never work in Ameri- can payphones.
In evaluating this argument, what questions should we ask? We can start with a question that we should ask of any argument.