September 4, 2019 | Data

When to be Skeptical of Data: Summaries and Averages

Summaries and averages are powerful data tools, but need to be used with care. Learn how to get the full picture with companion reports.

3 min

Data is awesome, but it can also be tricky. Regardless of your skill level, we’re all prone to common mistakes when it comes to gathering, visualizing and interpreting data. Knowing when to be skeptical of your results, and how to dig deeper will help you make the best decisions with the data you get from LabStats.

Summaries and Averages

Summary reports are a great way to get started with data. They give you a high level view so you can get a feel for what’s happening in your computer labs. However, if you make decisions simply based on summary results, you may be misled. There’s two ways to ensure you’re getting balanced and accurate data with summary reports.

Cross Reference Reports

The Average Login History report shows a historical timeline of the average number of logins. Think of it as the “velocity” of student use, or how many individual sessions are being started within a time frame. For example, a kiosk area where students typically log in for 2 minute sessions, but many hundreds of students login a day, would have high student usage velocity.

Average Login History Report

This example shows the average number of logins. It looks like students use the Library much more than the Rocky Mountain lab. 

However, the number of logins only shows part of the picture. If you dive deeper into the login sessions, you can see that although fewer students are logging in to the Rocky Mountain lab each day, they’re using the computers for an average of 4 hours. Students in the Library, on the other hand are using the computers for an average of 2 minutes, which allows for higher login rates.

Rocky Mountain Stations (4 hour sessions)
Library Stations (2 minute sessions)

Most schools have a lab where students simply check email or print assignments. Those “quick hit” labs will show a disproportionately large number of logins when compared to another lab where students settle in to work on a project for several hours. 

Without checking the session times, you might be tempted to write off the Rocky Mountain lab as less of a priority and delay refreshes or technical support. This would be a mistake as the students are clearly relying on the Rocky Mountain lab to work on major assignments.

Although the data in the Average Login History report was accurate, it needed to be balanced with additional context from the Login Session reports. Using companion reports build a more complete picture so you can make better decisions.

Schedules and Holidays

When you just want to get a high level snapshot of usage, it’s easy to forget to set the parameters of reports. One criteria that can greatly affect the accuracy of your data (at any level) is setting schedules and holidays.

Labs that are open twenty-four hours a day can and should have a different definition of 100% utilization than labs that are only open for 3 hours a day. Class schedules, weekends, and regular maintenance hours should be considered when comparing usage of different labs.

Setting a schedule is especially important for the Usage History report, which shows the percentage of computer utilization based on open lab hours. 

In the same way, it’s easy to overlook holidays when analyzing reports that span a whole semester or year. Setting up holidays (and updating annually) before you start pulling reports can improve the accuracy of your summary and average reports.

Although there are some inherent dangers with relying on summaries and averages, they are still powerful tools to understanding computer usage better. Being aware of blind spots, proactively setting criteria and utilizing companion reports are the best ways to get the most accurate and reliable data.