How to Understand What Data Really Means

Mike Labun

data, analysis, business, organizational, qualitative data, quantitative data, data analysis, data analyst

More and more, data is being used to summarize information about our complex world. Whether it’s epidemiologists reporting on COVID-19 or an HR assistant telling us which employees we should be disciplining, when someone processes the data – and does it well – it can be quite useful. But number crunching doesn’t automatically give us accurate or useful results.

As I describe four ways to better understand what the numbers are saying, I’ll illustrate with some COVID-19 statistics from my home country of Canada and worldwide. I’ll also discuss how we might think about the numbers when trying to determine if we should talk to an employee about their lateness.

Ask for the count and the percentage or rate.

On January 12th, the count of active COVID-19 cases in my birth province of Manitoba was 741. Another Canadian province, Quebec, reported a count of 8,737 active cases. It sounds like Manitoba’s active cases were a piddly 8% of Quebec’s, but their population is over six times larger than Manitoba’s. It turns out that the rate of active cases per 100,000 in Quebec was 103, while Manitoba’s was 54. Manitoba was doing better, but not as much as it initially seemed.

Numbers are meaningless without comparisons, but choosing the correct comparison can make all the difference.

Moving to employees, it seems like Sam has been late a lot, but you’re not sure if it’s a problem. You can think of at least 10 times in recent history when Sam was late – but that’s just a count. What you need to calculate is the percentage of days Sam is late. If, for example, Sam works many overtime days, the rate might be lower than you think. Or you may discover that Sam is rarely late overall and what you’re noticing is just a recent rash of tardiness.

Compare with a meaningful average.

Numbers are meaningless without comparisons, but choosing the correct comparison can make all the difference. Although Manitoba was doing better than Quebec, when we compare its rate of 54 active cases per 100,000 with other provinces, we see that the average rate across all of Canada was actually 26.7 on January 12th.

Similarly, it’s not fair to compare Sam with your best employee. Instead, calculate the average amount of tardiness in your company or department to compare with Sam.

Think about percentages in multiple ways.

Suppose you discover that, on average, your employees are late 1.5% of the time, while Sam is late 2% of the time. That makes Sam late 0.5% more than average – but it’s also 1/3 more than average. It’s up to you how you think of it, but to be fair to Sam, you should communicate it both ways if you decide to discuss their punctuality.

When in doubt, ask for a graph.

On December 5th of last year, according to Google, the world had 866,721 new COVID-19 cases in one day. A month later on January 8th, there were 809,260! That’s a drop of over 6%. However, the graph below instantly shows us how these numbers are deceiving.

While it’s true that January 8th, 2021, is a little lower than some of the high points in December 2020, the trend is clear: Despite daily fluctuations, active cases are on the increase worldwide. Indeed, using December 5th to compare with January 8th is a cheat. We can now clearly see that December 5th was a high point, which we have nearly reached at least once since that time, and could possibly reach again.

Frame your results so that the numbers will be easily understood without being misleading.

Similarly, if you want to know if Sam is becoming late more frequently, it would be wise to graph their arrival times to see if it’s true.

Consider how things have been measured.

There has been much discussion regarding how accurate COVID-19 tests are. Similarly, you should consider how accurate your measurement of lateness actually is, especially when making comparisons with other employees or employee averages.

For example, do all supervisors keep track of lateness? What constitutes lateness for them? Are they also tracking the reverse of lateness: early arrivals? What about staying late to make up for lost time? How would changing how we keep track of lateness possibly change our results? It would be wise to consider these things before approaching Sam so that we don’t end up reprimanding an employee who’s doing their best to be fair to the company.

Given that data is becoming increasingly important in the modern world, it’s important that we understand what the numbers we’re given do – and what they don’t do. If you’re the data cruncher, use these four tips to help you consider your data practices and frame your results so that the numbers will be easily understood without being misleading. If you’re the decision-maker, use the tips to discern when you can make a decision based on the data you have and when you need to ask for more information.

 

For more FREE RESOURCES on this topic and others, visit our free resources page.


Author: Mike Labun
Trainer, ACHIEVE Centre for Leadership
Mike is the co-author of ACHIEVE’s book, The Culture Question: How to Create a Workplace Where People Like to Work. The book is available on our website.To receive notification of a new blog posting, subscribe to our mailing list or follow us on Facebook, Twitter, and LinkedIn.© ACHIEVE Centre for Leadership
Interested in using the content of this blog? Learn more here.