Psychometrics 101: Know What Your Assessment Data is Telling You
Recently, Eric Ermie, Vice President of Sales, ExamSoft Worldwide, discussed how to put data into a useful context that will allow you to draw your own conclusions on what it means, how you should apply them, and why you should ignore rules that others may use for their specific situation.
The key takeaways from the presentation include:
It’s important to consider extraneous factors like cheating, conflicting content, and content delivery methods when analyzing psychometric exam item data points.
All exam statistics should be layered and considered together in order to get the clearest picture of how an item performed.
Setting general guidelines that outline desired psychometric statistical ranges will help educational programs to more effectively gauge exam item performance.
Keep it? Throw it out? Content/teaching issue? Bad question? Too easy? Too hard? What the heck? More than likely you have asked some or all of these questions at one point or another when trying to understand the performance of questions on an assessment. With differing opinions on how to interpret the statistics provided, how do you know what all this data is trying to tell you? Join us for a webinar on the fundamentals of item analysis, how the data is derived, and the different ways they can be interpreted. This presentation will cover how to put data into a useful context that will allow you to draw your own conclusions on what it means, how you should apply them, and why you should ignore rules that others may use for their specific situation.
“To get to the best conclusion on how an exam item performed, it's important that item analysis is always combined with faculty knowledge and expertise"