Seven Tips for using Statistics on People Data
There are a number of myths and misconceptions associated with using statistics in HR that often prevent us from harnessing the value that they possess, such as “advanced statistics are better than basic statistics” or that “a negative correlation is bad”. Well we’re here to bust those myths and misconceptions and show you that actually using statistics in HR and your People Analytics projects can be quite powerful. In this practical online training course, Heather Whiteman, a lecturer at UC Berkeley helps you build a solid understanding of the power of basic and advanced statistics and how to apply them in an HR context to build actionable outcomes from your data analysis.
In this blog post and in the video above, Heather shares seven tips to help you build more confidence when using statistics and harnessing the power they possess, to produce excellent analysis and effectively communicate using your HR data.
Tip #1: Don’t be average
So many people stick to leveraging only basic averages when using statistics in their data analysis, when in fact there is so much more information available to help you understand what is really going on with your people data. You can learn so much by using other basic statistics like the median or mode. These are known as measures of central tendency, primarily because they measure the middle of data and can help build a more detailed picture allowing you to really understand what's going on with your people data and tell a compelling story when communicating to your stakeholders.
Tip #2: Correlation doesn’t equal causation
Tip number two is correlation does not always equal causation. While many people know that correlation doesn't equal causation, few people know the difference between a negative and a positive correlation. Negative and positive correlations just indicate the direction in which the data is heading. Too often people confuse a negative correlation with a bad correlation. When in reality all it is saying is that as one variable goes up, the other variable goes down. This is one of the more commonly confused areas that can prevent people from using statistics in their data analysis. It is also important to note that if you do have a correlation between two items or pieces of data, this does not automatically equal causation – there can be a number of explanations as to what is causing this correlation or the order that things happen in and it is important to bear this in mind.
Tip #3: Sometimes your goal in prediction analysis is to be wrong
Tip number three is sometimes your goal in prediction analysis is to be wrong. A prediction model is like a rear-view mirror in a car rather than a car windshield. We can see a bit about where we might be going based on where we've been in the past, but we are still yet to see the curves in the road ahead. When you build a prediction model the aim is not always to predict the future accurately, sometimes the aim is to prove your prediction model wrong. For example, you may develop a prediction model that helps predict attrition, rather than aiming to be correct, you may decide to make changes to your talent management activities that help drive employee retention, thus proving your prediction wrong.
Interested in getting certified in People Analytics? Take a look at our online People Analytics certifications on myHRfuture
Tip #4: Sometimes practical significance is even more important than statistical significance
Tip number four is that sometimes practical significance is even more important than statistical significance. While statistical significance relates to whether an effect exists, practical significance refers to the magnitude of the effect. Sometimes what you’re assesing, how it will be applied and how the outcomes will affect people matters more than what the number is on a statistical test. Statistical significance indicates only that you have sufficient evidence to conclude that an effect exists. It is a mathematical definition that does not know anything about the subject area and what constitutes an important effect.
Tip #5: Use multiple sources of data.
Tip number five is that you should use multiple sources of data. Your analysis should always take into account as much information as possible to really understand the full picture of what's going on. No statistical test can tell you whether the effect is large enough to be important in your field of study. Instead, you need to apply your subject area knowledge and expertise to determine whether the effect is big enough to be meaningful in the real world.
Tip #6: Be aware of biases
Tip number six is to be aware of biases. Biases are inherent in who we are as human beings, we have over 200 biases built into our DNA and they tend to shade how we see and interpret data. It is important that we are aware of potential biases and something we should consider and attempt to control as we’re working with people data.
Tip #7: People data is private data
Finally, tip number seven is that people data is private data! It is important to remember that when you’re using people data, be conscious and considerate of how you’re using the data and what you’re analysing. As Heather explains:
“when you're using it, just check yourself. Don't be creepy if you think what you're analysing is weird, others probably do too. And a tip for that is treat it like its private data. Treat it like it's your own data and don't be creepy.”
When using statistics to analyse your HR data and drive actionable outcomes it is important to remember these seven tips. Don’t be average – don’t just use basic averages, leverage median and mode to paint a more detailed picture. A negative correlation isn’t a bad correlation it’s simply saying that as one variable goes up, the other variable goes down. When you build a prediction model, the aim is not always to predict the future accurately sometimes the aim is to prove your prediction model wrong in the end and let’s not forget that practical significance is sometimes more important than statistical significance. Sometimes what you're assessing, how it will be applied, and how the outcomes will affect people, matters more than what the number is on a statistical test. Finally, be aware of biases and always remember that people data is private data!
Get certified in People Analytics and build the HR skills of the future!
The myHRfuture academy is a learning experience platform for HR professionals looking to invest in their careers. We have many online HR certifications that are designed for People Analytics professionals and HR Business Partners to help them build their knowledge and get a People Analytics certification or a certification in Digital HR and understand how to become more digital and data-driven. Our learning pathways and skill booster certifications comprise of a blend of HR training courses that help you build the HR skills of the future and expertly curated content that provides real life context to what you’ve learnt as well as an official digital badge that can be shared with your network on LinkedIn, Twitter and other social media platforms at the click of a button. All of our online HR certifications reflect the core HR skills you need to develop for the future, giving you a professional edge while providing evidence of the skills you’ve mastered, helping you get ahead and press PLAY on your career. Sign-up today for unlimited to access to myHRfuture for only £25 per month.