Research data for study "Making Regression Real: Investigating the Impact of 3D Augmented Reality Visualisation on Understanding Multiple Linear Regression".

Experimental response data for the study "Making Regression Real: Investigating the Impact of 3D Augmented Reality Visualisation on Understanding Multiple Linear Regression".

Abstract to study:

Many businesses wish to be ``data-driven", which requires decision-makers understand basic statistics so that they can make sense of the data they are using, and not be fooled by randomness and uncertainty. However, many people, including many decision-makers, find statistics abstract and difficult to understand. There is therefore a problem: a key skill required for businesses to become data-driven is likely often missing.

In particular businesses must deal with multivariate data, where many factors contribute to an outcome. Immersive Analytics, which uses Virtual and Augmented Reality to visualize data, may help here, as we can now display multivariate data in 3D scatterplots in true 3D space. In this novel study we therefore investigated whether Augmented Reality would help non-statistically trained people to understand the fundamental statistical method of Multiple Linear Regression.

Our results show that the accuracy of extracting information from 3D scatterplots displayed in true 3D space was about the same as on a 2D screen, but that participants using Augmented Reality reported higher levels of engagement, autonomy, and lower pressure/tension. This suggests that decision-makers may find it easier to learn the fundamental concepts of regression and other statistical methods more easily with immersive visualizations, and with less resistance than might otherwise be the case.

This is one of the first studies of its kind and so needs to be researched further, but anything that can make the task of producing statistically-literate decision-makers easier so that they can make good, ``data-driven" decisions should be a priority for investigation.

HOW TO USE:

The script to run the code to produce the results described in the study is included "CODE_FOR_PAPER_FINAL.R". Once the working directory is set to the folder with the Experience.csv and SLRandMLRresponses.csv the code should run to produce the results.