Explain in your own words why a bar graph can be misleading if one or both of the scales begin at some value other than zero.

Answer 1

Good Morning, Kim,

As is the case with most issues in statistics, you will find a great deal of disagreement over several issues. In other words, the principle that all bar graphs need to begin at zero, is rooted in the theory of Darrell Huff, in his seminal work titled, "How To Lie In Statistics," but was countered quite well by another data visualization expert, Edward Tufte. However, at its core, the argument as advanced by Huff and several purist statisticians is that neglecting to begin at a fixed zero point in a bar graph leads to skew how the data is perceived.

Please note: there are several ways one can skew how data is perceived, with each of those ways rooted in truncating the y-axis. For example, if you showcase two sets of data, you really should convey that data by a bar graph with both comparisons beginning with zero, versus starting one graph with a y-axis beginning at zero, and and another with the bar graph beginning at some other value.

Finally, also note: Due to the debate which is fervent between Huff and Tufts, it is clear that what is actually effected is the perception the viewer of the data receives, and being misled. This is precisely the reason the statistician should always use the most effective manner in conveying the data, and doing so in an ethical manner.

I hope this answer provides you an adequate answer, and should you need further clarity, or simply wish to leave feedback, please do so directly underneath this solution. If you need further assistance in this particular area of statistics, please feel free to reach out to any of our more than qualified tutors. Best!