BUSI 3014 Week 7 Misleading Chart Critique Example

A Callback Chart That Makes One Technician Look Like a Failure: Critiquing a Truncated Bar Chart of Raw Counts With No Time Period, and Rebuilding It as Callbacks per 100 Treatments From Zero

Student Name

Walden University

BUSI 3014: Data Driven Decision Making Workplace Certificate

Instructor Name

Month Day, Year

A Callback Chart That Makes One Technician Look Like a Failure: Critiquing a Truncated Bar Chart of Raw Counts With No Time Period, and Rebuilding It as Callbacks per 100 Treatments From Zero

The False Message

The chart came from the monthly operations packet the Kestrel owner reads, part of the composite pest control case used all term. It is a bar chart titled "Callbacks by Technician," showing the number of callbacks, repeat visits made because a treatment did not hold, for each of the four technicians. An owner glancing at it before a performance review would conclude that Technician D is failing badly: D's bar towers over the other three, roughly seven times as tall as the shortest.

That conclusion is wrong, and the chart produces it through three separate distortions.

The Original Chart

Because the original cannot be reproduced exactly here, the table records what it showed and how it was drawn. Its vertical axis started at 40, not zero, and it carried no date range.

Table 1

What the Original Chart Displayed

Composite case data. The callbacks cover January to December 2027, a period the chart did not state.

Distortion One: The Truncated Axis

Starting the axis at 40 removes the first 40 callbacks from every bar. What remains exaggerates differences. Technician D's 71 callbacks are about 1.6 times Technician C's 44, but on the chart D's bar is nearly eight times as tall. Tufte (2001) proposed a "lie factor": how much a picture inflates a difference, relative to how big the difference really is. On the chart, D's bar is about 675 percent taller than C's; in the data, D's count is about 61 percent higher. The ratio of those two exaggerations, roughly 11, is the chart's lie factor. Pandey et al. (2015) found experimentally that truncated axes lead viewers to judge differences as larger than they are, even when the axis labels are visible. Bar charts are especially vulnerable, because readers judge bars by their length from the baseline.

Distortion Two: Counts Where Rates Belong

The chart shows raw callbacks, but the technicians did not do the same amount of work. In 2027, Technician D completed 2,020 treatments, because D runs the company's largest route and covers for others on leave. The other three completed between 1,090 and 1,310 each. More treatments mean more chances for a callback. Comparing raw counts treats a technician who did nearly twice the work as if the workloads were equal. This distortion supplies the core of the false message: it makes the busiest technician look like the worst.

Distortion Three: No Time Period

The chart has no dates. A reader cannot tell whether it covers a month, a quarter or a year, or whether it includes the summer, when ant and termite callbacks rise. Without a period, the reader cannot compare it with last year's chart or judge whether 71 is high. It also hides the fact that D's route is heavy with summer termite work, which has the highest callback rate in the company.

The Rebuilt Chart

The rebuilt chart corrects all three problems. It shows callbacks per 100 treatments, starts the axis at zero, and states the period in the title.

Figure 1

Callbacks per 100 Treatments by Technician, January to December 2027

Rates are callbacks divided by completed treatments, times 100. A: 52 of 1,310; B: 47 of 1,180; C: 44 of 1,090; D: 71 of 2,020.

What the Rebuilt Chart Says

Rebuilt, the chart tells the opposite story. Technician D has the lowest callback rate of the four, about 3.5 per 100 treatments against about 4.0 for the others. An owner reading the rebuilt chart would see four technicians performing similarly, with D slightly ahead despite the heaviest route.

The difference matters. The original chart was in the packet the week before annual reviews. Had the owner acted on it, the company's most productive technician would have been criticized for being the busiest. Experiments by Cleveland and McGill (1984) ranked reading position on a shared axis above other ways of encoding numbers, which is why a bar chart from zero, with bars on one scale, is the right form for this comparison once the values are rates.

What the Owner Should Do With the Rebuild

The rebuilt chart should replace the original in the monthly packet, with the period stated and the rate formula written under it. It also suggests a rule for every chart in the packet: compare people or routes by rates when their workloads differ, and start bar charts at zero.

A Note on What Even the Rebuild Does Not Show

The rebuilt chart still compares technicians who serve different routes and pest types. Some of the remaining difference may come from the mix of jobs rather than skill. A further version could compare callback rates within the same treatment type, such as general pest service only, to separate the technician from the work.

References

Cleveland, W. S., & McGill, R. (1984). Graphical perception: Theory, experimentation, and application to the development of graphical methods. Journal of the American Statistical Association, 79(387), 531-554. https://doi.org/10.1080/01621459.1984.10478080

Pandey, A. V., Rall, K., Satterthwaite, M. L., Nov, O., & Bertini, E. (2015). How deceptive are deceptive visualizations? An empirical analysis of common distortion techniques. In Proceedings of the 33rd Annual ACM Conference on Human Factors in Computing Systems (pp. 1469-1478). ACM. https://doi.org/10.1145/2702123.2702608

Tufte, E. R. (2001). The visual display of quantitative information (2nd ed.). Graphics Press.