Graphs can make information to understand, but they can also . One common trick is to start the vertical at a number other zero. A small , such as from 50 to 52, can then enormous. Another is to choose a time that supports a particular while ignoring that would tell a different .

One of the first lessons in any statistics course is that correlation does not imply causation. When two things tend to rise and fall together, it is tempting to assume that one causes the other. Often, however, the relationship has another explanation.

Consider a well-known example: ice cream sales and drowning deaths both increase in the summer. Ice cream does not cause drowning. Instead, a third factor, hot weather, leads both to more ice cream eating and to more swimming. Statisticians call such a hidden factor a confounding variable.

Sometimes the direction of cause is unclear. Students who use the library often may have higher grades, but it may be that motivated students both use the library and earn good grades, rather than library visits themselves raising grades.

Some correlations are simply coincidences. With enough data, researchers can find strange connections between things that have nothing to do with each other.

To establish causation, scientists usually rely on controlled experiments in which participants are randomly assigned to different conditions. Random assignment helps ensure that any differences between groups are caused by the treatment rather than by hidden factors.