How is a graph misleading?

A graph can be misleading in many ways. The first, and primary, mistake people make is to assume that correlation implies causation. If you see a graph with hours spent volunteering on the x axis and happiness on the y axis, with a positive line drawn on the graph, this could make people assume that volunteering makes you happy. The graph only shows, though, that happiness and volunteering are correlated. It could be that happy people are more likely to volunteer.
Another way graphs can be misleading is if the person reading them mentally extrapolates the line. If we have a graph of age (from 20 to 50) and the record running speeds for people of that age, the line would go down, and the person reading might conclude that the younger you get, the quicker you can run. This would mean that babies would be the world's best runners, which is obviously incorrect.
A third important way that graphs can be misleading is if the axes don't start their values at 0. For example, if we have a graph of cigarettes smoked on the x axis, and lifespan on the y axis, it would be sensible to start the y axis from 65 or something. Someone looking at this might see the line approaching the x axis and assume that this means 30 cigarettes a day means your lifespan is 0, and that even more can give you a negative lifespan, even though that's impossible.