April 6, 2020
Understanding Regression Analysis in Workplace Safety
Written by: EHS Insight Resources
What would you say if we told you that safety data can be used to help reduce workplace incidents?
If you’re like most forward-thinking safety professionals, you’ve reached that conclusion already. The key is knowing how to leverage the data to reduce incidents.
Regression analysis can help. Here’s how regression analysis in workplace safety works and how it can help you turn raw data into actionable insights.
What Is Regression Analysis?
Let’s say you’re a safety manager. And let’s say you’re trying to predict next month’s safety numbers. The problem is that there are hundreds of factors that might affect the outcome. How do you know which factors should have more weight?
Regression is a statistical method that attempts to determine the strength and character of the relationship between one dependent variable (usually denoted as Y) and a series of independent variables.
Regression analysis is a type of predictive modeling technique investigating the relationship between a dependent variable (target) and an independent variable (predictor). It is used to find a cause and effect relationship between variables, as well as time series modeling and forecasting.
Types of Regression Analysis
There are several types of regression analysis, including:
- ElasticNet regression
- Lasso regression
- Linear regression
- Logistic regression
- Polynomial regression
- Ridge regression
- Stepwise regression
The most commonly used regression technique is linear regression, which is a regression model made up of linear variables. It uses a linear approach to model the relationship between the criterion (scalar response) and the predictors (explanatory variables).
In plain English, you take a dependent variable and an independent variable and model the relationship using a line. It’s fast and easy to model and equally fast and easy to understand, but it’s also highly sensitive to outliers.
Logistic regression is used to find the probability of event success and event failure. As such, it is used when the logistic value is binary (i.e. true or false).
As you can see, each type of regression analysis will offer you a different interpretation of the data - the key is to know what you’re trying to find, and thus what type of regression analysis to use.
Regression Analysis in Workplace Safety
If you’re trying to turn safety data into numbers and predictions, regression analysis is the tool you need for the job.
Let’s think back to our earlier example. You’re a safety manager with a wealth of data, and you’re trying to predict next month’s numbers. How do you know which variables to account for, which variables have the most weight, and how to make predictions accordingly?
Regression analysis allows you to determine what variables actually matter – and provide results to help support your conclusions. That way, you can stop guesstimating data and deliver predictions with confidence.
Putting Your Tools to Work for You
Regression analysis in workplace safety is a way for you to turn your data into actionable insights. That way, when you make your case to management, you can back it up with numbers and deliver real results.
If you’re looking for more ways to bring smart data into your EHS program, make sure to check out our blog for more great tips, like this post on how to become a data-driven EHS professional.
Related:
EHS Insight Resources
Since 2009, the team at EHS Insight have been on a mission to make the world a better place. Join us by subscribing to our Blog and receive updates on what’s new in the world of EHS, our software and other related topics.
