Originally published on May 10, 2019, on LindaSuskie.com
A recent question posted to the ASSESS listserv led to a lively discussion of direct vs. indirect evidence of student learning, including what they are and the merits of each.
I really hate jargon, and “direct” and “indirect” is right at the top of my list of jargon I hate. A few years ago I did a little poking around to try to figure out who came up with these terms. The earliest reference I could find was in a government regulation. That makes sense—governments are great at coming up with obtuse jargon!
I suspect the terms came from the legal world, which uses the concepts of direct and circumstantial evidence. Direct evidence in the legal world is evidence that supports an assertion without the need for additional evidence. Witness knowledge or direct recollection are examples of direct evidence. Circumstantial evidence is evidence from which reasonable inferences may be drawn.
In the legal world, both direct and circumstantial evidence are acceptable and each alone may be sufficient to make a legal decision. Here’s an often-cited example: If you got up in the middle of the night and saw that it’s snowing, that’s direct evidence that it snowed overnight. If you got up in the morning and saw snow on the ground, that’s circumstantial evidence that it snowed overnight. Obviously both are sufficient evidence that it snowed overnight.
But let’s say you got up in the morning and saw that the roads were wet. That’s circumstantial evidence that it rained overnight. But the evidence is not as compelling, because there might be other reasons the roads were wet. It might have snowed and the snow melted by dawn. It might have been foggy. Or street cleaners may have come through overnight. In this example, this circumstantial evidence would be more compelling if it were accompanied by corroborating evidence, such as a report from a local weather station or someone living a mile away who did get up in the middle of the night and saw rain.
So, in the legal world, direct evidence is observed and circumstantial evidence is inferred. Maybe “observed” and “inferred” might be better terms for direct and indirect evidence of student learning. Direct evidence can be observed through student products and performances. Indirect evidence must be inferred through what students tell us, through things like surveys and interviews, or what faculty tell us through things like grades, or some student behaviors such as graduation or job placement.
But the problem with using “observable” and “inferred” is that all student learning is inferred to some extent. If a crime is recorded on video, that’s clearly direct, observable evidence. But if a student writes a research paper or makes a presentation or takes a test, we’re only observing a sample of what they’ve learned, and maybe it’s not a good sample. Maybe the test happened to focus heavily on the concepts the student didn’t learn. Maybe the student was ill the day of the presentation. When we assess student learning, we’re trying to see into a student’s mind. It’s like looking into a black box fitted with lenses that are all somewhat blurry or distorted. We may need to look through several lenses, from several angles, to infer reasonably accurately what’s inside.
In the ASSESS listserv discussion, John Hathcoat and Jeremy Penn both suggested that direct and indirect evidence fall on a continuum. This is why. Some lenses are clearer than others. Some direct evidence is more compelling or convincing than others. If we see a nursing student intubate a patient successfully, we can be pretty confident that the student can perform this procedure correctly. But if we assess a student essay, we can’t be as confident about the student’s writing skill, because the skill level displayed can depend on factors such as the essay’s topic, the time and circumstances under which the student completes the assignment, and the clarity of the prompt (instructions).
So I define direct evidence as not only observable but sufficiently convincing that a critic would be persuaded. Imagine someone prominent in your community who thinks your college, your program, or your courses are a joke—students learn nothing worthwhile in them. Direct evidence is the kind that the critic wouldn’t challenge. Grades, student self-ratings, and surveys wouldn’t convince that critic. But rubric results, accompanied by a few samples of student work, would be harder for the critic to refute.
So should faculty be asked or required to provide direct and indirect evidence of student learning? If your accreditor requires direct and indirect evidence, obviously yes. Otherwise, the need for direct evidence depends on how it will be used. Direct evidence should be used, for example, when deciding whether students will progress or graduate or whether to fund or terminate a program. The need for direct evidence also depends on the likelihood that the evidence will be challenged. For relatively minor uses, such as evaluating a brief co-curricular experience, indirect evidence may be just as useful as direct evidence, if not even more insightful.
One last note on direct/observable evidence: learning goals for attitudes, values, and dispositions can be difficult if not impossible to observe. That’s because, as hard as it is to see into the mind (with that black box analogy), it’s even harder to see into the soul. One of the questions on the ASSESS listserv was what constitutes direct evidence that a dancer dances with confidence. Suppose you’re observing two dancers performing. One has enormous confidence and the other has none. Would you be able to tell them apart from their performances? If so, how? What would you see in one performance that you wouldn’t see in the other? If you can observe a difference, you can collect direct evidence. But if the difference is only in their soul—not observable—you’ll need to rely on indirect evidence to assess this learning goal.