Part 1. Surveys in Educational Settings
In education, we tend to use Likert-scale surveys to measure just about everything.
In our quest to become more evidence-based in our planning and decision-making, we often resort to survey instruments, and in particular, Likert-scaled instruments, to measure key outcomes, get feedback from stakeholders, and evaluate programs and solutions. We use the Likert-scale instrument like a hammer that treats each measurement challenge like it is a nail. However, we have unfortunately reached a point in the arc of evidence-based practice where stakeholders are really starting to feel the effects of our use (and perhaps overuse) of “quick and easy surveys” in educational settings.
“We use the Likert-scale instrument like a hammer that treats each measurement challenge like it is a nail.”
– My new mantra
The Consequences of Our Common Practices
As a result of our overuse of Likert-scale surveys, many of the practices we employ within each phase of the survey administration process lead to unintended consequences, many of which I am sure you have also experienced:
Phase | Common Practices | Unintended Consequences |
Instrument selection | We use surveys to measure everything under the sun, such as engagement, attitudes, reactions to professional learning, and implementation of instructional practices. | We begin to measure low-priority constructs, seeking “feedback” about trivial operational decisions instead of the key strategic levers that affect organizational performance. |
Survey design | We design Likert-scaled instruments because they seem “easy” to write. | We fail to consider item designs that may better measure what matters in leading transformation. |
Survey administration | We gravitate toward self-service tools that grant affordable access to efficient data collection. | This luxury has led to survey ubiquity in our data-driven culture, which in some contexts, has minimized the value of self-report data. |
Survey analysis | We compute and report mean scores and percent agreement for most indicators. | We only skim the surface of the meaning and narrative that we desire from our data. |
Survey interpretation | We evaluate results based on which items “scored the lowest or the highest,” and use these scores to set goals or track improvement over time. | We fail to take into account the ways these items violate statistical assumptions, or meaningfully triangulate the results with other information and context. |
Over the last 10 plus years as a continuous improvement leader and evaluator, I know I have been guilty of these practices. In the past, Likert-scale instruments helped me obtain data where information was scarce, when outcomes seemed intangible, or when I desperately needed to verify the merit of a program I cared deeply about. But I have learned the hard way that the rush to collect such data led to downstream impacts I later came to regret.
While there are many factors that have contributed to the current decadence in survey management in schools, I believe that one of the key culprits that has had a trickle-down impact on all the other issues is our overreliance on the Likert-scale design.
The Diminishing Value of Self-Report Data
Of the myriad consequences of the perceived ease, affordability, and predictability of Likert-scale instruments, chief among them is the slow diminishing value of self-report data as a meaningful way to gather feedback about things that matter to leaders. Many educational leaders dismiss surveys in general as too “biased” or “unreliable” because they depend on self-report methods.
But frustrations related to our undisciplined survey practices should not undercut the importance of self-report data.
In a constructivist paradigm focused on improving the experiences of educators and learners, self-report measures still serve as the gold standard in evidence, and hold much unrealized potential as a key means to engage and empower stakeholders. Self-report instruments, when designed with intentionality and laser-like focus on the most important enablers of effectiveness, can provide a meaningful source of insight about practices and a great way to measure and track progress over time.
While we may continue to rely on the Likert-scale design in some instances, I believe we have an opportunity to rethink the Likert-scale and consider alternative item types and models for measuring and evaluating what matters. In part 2 of this blog series, we discuss why and how this might be achieved.