3 ways we all (yes, even you) manipulate data

It's natural to want to paint yourself in the best light possible. Sometimes, this means crafting a story in a compelling way to make it more exciting, dramatic, inspiring, or impressive. My favorite example from the internet is how to add changing a light bulb to a resume to make it sound like much more than it really is.


Single-handedly managed the successful upgrade and deployment of new environmental illumination system with zero cost overruns and zero safety accidents.

We're all guilty of this in some way or another (although, probably not to this extreme). Many times, we're probably not even aware we're doing it. Unfortunately for you, I'm going to shine a light on this game of sleight of hand to make you think twice next time you try to pull a fast one.


Pay me now or pay me later


Sometimes the tricks we pull are comical. This first one is particularly entertaining to me. Let's whisk our minds away to corporate America for a minute and imagine a department manager (Fred) being asked to assess his performance for the prior year. Why use cold, hard facts to assess performance when you can use data?


Fred proudly reported a nearly 10% YoY reduction in annual department spend. How's that for progress!?


Here's what Fred presented to his boss

And here's Fred's boss' response:


Now here's what really happened.


Fred was sly and had been anticipating this question for several months. Because his annual raise and performance bonus hinged on convincing his boss he'd performed well for the prior year (and because he really wanted a new boat), he took matters into his own hands starting in November of 2020.


Here's what the data looked like 3 months into 2021 without Fred's creative spin:


So, Fred didn't reduce his department's spend by 10% YoY. Instead, he reduced his spend in November and December only to make up for it and then some in Q1 2021. Bad Fred.


Don't look behind the curtain


Data can sometimes look much more impressive when presented as raw numbers rather than percentages. Let's look at an example.


Mary is responsible for reducing the amount of waste from production her company sends to landfill each year. Mary has been diligently tracking this data for the last five years and was interested in being promoted to a data analyst position at her company.


Mary prepared the graph below to highlight how well she had done to reduce the weight of material sent to landfill under her leadership.


The hiring manager's response was not shocking:


Unfortunately, scratching a bit below the surface reveals a much different story. For the last five years, sales have been slipping and as a result the number of units produced has decreased year over year. Mary failed to include this in her analysis (not quite data analyst material, eh?).


Here's what a bit more thorough analysis revealed:

While the amount of waste going to landfill each year did decrease, the company ended up sending a higher % of waste to landfill in comparison to their total weight of production. Tsk tsk, Mary.


Don't look too closely


Graphs are easily to manipulate. You can do all sorts of fun things to make them tell a story the way you want it to be told.


Todd is a surfboard salesman with the Great Plains as his territory. He does well peddling his wares in this portion of the country, but he wants to make a splash with his superiors and highlight how much he's been able to increase sales each year.


Todd presented this gem of a graph:



Wow, how amazing! Look at that slope! Todd is absolutely crushing it! Prepare the corner office and parking spot!


This is exceptional trickery on Todd's part. Not including the vertical axis is a special kind of data manipulation. Here's what the graph really looks like when Todd isn't pulling the strings:


Even this version is attempting to hide the truth. The scale itself is so narrow it makes the year over year sales increase look dramatic. However, starting the vertical axis at $0 paints the true picture.


Sales. Are. Flat.


Takeaways


Hopefully you've learned examples of what NOT to do when presenting data. However, my fear is you've now learned additional ways to manipulate data and will begin using them to nefarious ends. Either way, tell your friends.




35 views0 comments