Civic discourse programs are often criticized as too "squishy" to measure. But the bigger question isn't whether these programs can be measured—it's whether campuses are measuring the right outcomes, and measuring them well. Openness to disagreement, intellectual humility, and willingness to engage across difference are all measurable outcomes. But getting useful data requires more than a post-program survey or a participation count. It requires knowing what you're trying to change, why you think your program will change it, and what evidence would actually show progress.

This post offers practical guidance for campus practitioners on how to measure civic discourse well.

Measure with Purpose

Measurement should begin with a clear goal.

Are you trying to improve a program? Decide what programs to purchase, renew, or scale? Demonstrate impact for stakeholders who expect evidence? Learn something generalizable about what works, for whom, and under what conditions? Each goal shapes what data you collect, how you measure and analyze it, and how you communicate results.

One of the most common mistakes in campus assessment is starting with the measurement metric instead of the measurement goal. "What metric should we report?" is the wrong first question. A better one is: "What are we trying to learn, and why?"

Clarify Your Constructs

Many of the outcomes that matter most in this work (such as intellectual humility and openness to disagreement) aren't visible or directly measurable. In measurement language, these are called constructs. That doesn't mean they're unmeasurable, but it means we have to be clear about what we mean before we decide how to measure them.

Take intellectual humility, which scholars define as the ability to recognize the limits of one’s knowledge and the possibility of being wrong. Based on that definition, you might then decide to administer a self-report scale, with questions such as "I accept that my beliefs and attitudes may be wrong." Or you could look at behavioral indicators: does a student acknowledge uncertainty, ask genuine questions, or revise a position? Or you might analyze qualitative evidence from self-reflections, interviews, or dialogue transcripts. Each is a different way of making the same construct into something observable. Clarify what you're trying to change before deciding how to measure it.

Good Measurement Starts with Program Design

Good measurement begins before evaluation. It begins in program design.

Strong civic discourse programs are designed around specific goals: reducing misperceptions, increasing perspective-taking, improving comfort with disagreement, or strengthening discussion skills. Those goals should shape what institutions choose to measure.

CDI's Perspectives program, for example, draws on research from social psychology, moral psychology, and conflict studies to address how people engage across differences. Because Perspectives is designed to reduce affective polarization, increase intellectual humility, and improve how students navigate conflict, those are outcomes worth measuring.

That alignment matters. Measurement is most useful when it reflects the actual aims of the program being evaluated, rather than relying on vague or generic indicators of success.

Good Measurement Requires More Than a Survey

Surveys are useful, but they shouldn't be the only measurement strategy. They capture what students report, which is genuinely informative but is only one kind of evidence. Combining surveys with behavioral and administrative data gives a fuller picture. As CDI has explored elsewhere, this is increasingly how the field measures constructive dialogue.

For example, participation data and learning analytics can show you the extent and depth of student engagement. Skill demonstration data, such as role plays, scenario responses, and actual conversations, lets you see what students can actually do rather than only what they say they can do. Qualitative data, like open-ended responses, reflections, focus groups, and dialogue transcripts, can help you understand the why behind the numbers and surface effects you didn't think to measure. Institutional indicators such as conduct referrals, bias incident trends, retention patterns, and campus climate data can signal whether relational dynamics are shifting over time.

No single measure is perfect on its own. The value comes from triangulation: looking across multiple sources and asking whether they point in the same direction.

Disaggregate Your Results

Don't stop at the average. A single overall number can mask a lot of variation in who participates, who completes, and who benefits.

The most useful analytics break results down by variables that reflect how the program was delivered and who experienced it: school year, major, course or section, completion status, baseline score, and political orientation if it's relevant to your goals. Look at participation, satisfaction, and outcomes across each cut. Where you see patterns, such as one group completing at lower rates, one group reporting a different experience, or one group showing different pre/post change, treat those as signals worth investigating rather than facts to simply report.

Disaggregation also helps you answer the question most campuses actually care about: not "did it work?" but "what worked, for whom, and under what conditions?" That framing turns measurement into something you can act on rather than a single grade for the program.

Distinguish Outcomes from Impact

Measurement gives you change in outcomes over time. Causation is a separate claim.

In a typical campus implementation, there is no control group, no random assignment, and only a single cohort going through a program. You can say with confidence that students' scores moved from one point to another. What you can't say, on the basis of that data alone, is that the program caused all of the change you observed. Maturation, parallel events on campus, self-selection into the program, national events, and natural drift in attitudes can all produce movement that has nothing to do with the program. Strong causal claims require strong designs, and most campus assessments aren't built to produce them.

That's not a reason to avoid measurement. It is a reason to write up findings carefully. Use language that describes what you observed, such as "students improved over the course of the program" or "completion was higher in courses that embedded the program in regular class time," and reserve causal language like "the program produced" or "the program caused" for studies that earned it. The strongest assessment reports are the ones that combine clear data with thoughtful interpretation, and that name what the data can and cannot tell you.

What Better Measurement Looks Like

Civic discourse is measurable. Doing it well takes:

• Clarity of purpose: measure for a reason, not by reflex.

• Construct before instrument: define what you're changing before you choose how to measure it.

• Design that drives measurement: let what the program is built to do shape what you assess.

• Multiple forms of evidence: triangulate across surveys, behavior, qualitative data, and institutional indicators.

• Honest reasoning about cause: outcomes are measured, but impact is inferred.

The goal isn't to make measurement overly technical. It's to make it useful: to help campuses improve programs, make better decisions, and build cultures where students engage across difference more constructively.