I have a number of favorite interview tests, and this is yet another one. Here’s the game: I draw two different charts, just like what you see here,except without the labels on the axes. One chart has a very slow rise and the other has a very steep rise. Then, I ask my poor interviewee which chart they would like to represent the raises they will receive from year to year.
Inevitably, everyone chooses the second option which seems to suggest they will be receiving much bigger raises each year. Rarely am I challenged on the fact that the axes are unlabeled and that knowing where the zero lies on the chart is the only way to know which chart will mean mo’ money.
Different charting programs have different rules about what to do with the zero point. When your data hovers close to the zero point, it’s just common sense to include the zero. On the other hand, when your data lies far from the zero point, perhaps having values like 112, 99, and 103, you might actually have to make a decision about whether to include the zero. Your decisions rests firmly on one important criteria.
*Is your audience chart smart?*
If you know for sure that your audience will immediately evaluate the scales on the chart, then you might be able to get away without using the zero point. Keep in mind that those oh so creative slash marks in the axis do NOT mean that the problem is solved. Your chart smart audience might notice them but they still leave the scale tampered and open to misrepresentation.
If, however, your audience is unfamiliar with the concept that small sample sizes can cause massive fluctuations in trendlines, or if your audience is numerophobic, you’re going to have to include the zero.
No matter your decision, you can always learn a lesson from the folks at xkcd comics.
Related Articles
- The richness of nothingness (junkcharts.typepad.com)
The LoveStats Blog 
Any reason not to always set the axis to zero? I think a chart that doesn’t is exaggerating the rate of change.
When graphing percentages, I try to remember to set the axis from 0% to 100%. I think the magnitude of the emptiness at the top of the chart is illustrative.
I think there are some cases where not putting zero makes sense, definitely in the science space. For instance, numbers that perpetually hover far from zero and change by fractions of decimal places. In MR, there are far fewer cases where it really makes sense to ignore the zero.