You can’t get rid of extraneous variables, but you can mitigate their effects. Measuring and controlling for users’ prior experiences and counterbalancing tasks and product order are two common ways to mitigate the effects of extraneous variables. measuringu.com/understanding…

In this article, we review two analytical methods used for two common #UXresearch activities: kappa (for classification studies) and any-2 agreement (for discovery studies). measuringu.com/assessing-int…

In our experiment comparing SUS scores from humans and digital twins, we found systematic differences in means and standard deviations. measuringu.com/how-well-do-s…

Picking the Right Methods to Improve Navigation. measuringu.com/methods-navig…

Do people use all available response options presented in agreement scales, even when there are eleven response options? measuringu.com/do-people-use…

In a study of rating behavior with radio button numeric scales and sliders, we found little difference in data quality, but respondents had faster completion times with and overwhelmingly preferred using radio button numeric scales. measuringu.com/time-and-pref…

We created two types of digital twins with different amounts of information about their matching humans. Which type would you expect to be more accurate—the twin with less information or the twin with more information? measuringu.com/how-well-do-s…

How do measures of Hedonic (Emotional) #Usability fit in with other #UX questionnaires? measuringu.com/three-branche…

12 Things to Know About Using the TAC-10 to Measure Tech Savviness. measuringu.com/how-to-use-th…

A guideline for estimating variance is to divide the range of the scores by 4, but some statisticians recommend dividing by 5 or 6. When conducting research with #UX rating scales and standardized questionnaires, which of these divisors works the best? measuringu.com/rating-scale-…

In our analysis of data from the 2024 UXPA salary survey, including the time needed to get the degree, it takes about 17 years for a master’s degree and 24 years for a PhD to catch up with the cumulative salary of those without an advanced degree. measuringu.com/does-an-advan…

As part of a comprehensive research effort to improve the 16-item SUPR-Qm, we developed an equivalent five-item version (the SUPR-Qm V2). measuringu.com/streamlining-…

Does AI Find Real UI Problems or Just Hallucinations? measuringu.com/does-ai-find-…

There are numerous ways to interpret SEQ scores, including percentiles, task success/time calibration, and an adjective scale. measuringu.com/evolution-of-…

AI Can't Replace Real Research in Empathy Mapping. (Nielsen Norman Group) nngroup.com/articles/ai-empa…

Extreme responses often predict behavior well. But three-point scales have no way of differentiating between extreme and tepid responses. measuringu.com/three-points/

We ran an experiment comparing quantitative results of one of our industry benchmark surveys with data generated by two types of digital twins. measuringu.com/how-well-do-s…

If you think the NPS is a poor measure, ask yourself why. Is it because you read a critical tweet from a prominent #UX personality, or is your opinion based on a comprehensive knowledge of the body of NPS research? measuringu.com/why-do-people…

The Single Ease Question (SEQ) is a popular measure of the perceived ease of task completion. We conducted three evaluations to better understand how variations in its format might or might not affect its measurement properties. measuringu.com/evaluation-of…

We recently conducted an experiment to compare measurement of key #UX attributes with numeric scales (5 response options) and with sliders (101 response options). After conversion to a 0–100-point scale, the difference between their means was just 0.2. measuringu.com/numbers-versu…