I have a complicated relationship with passing scores. I’ve seen too many of them treated as proof that someone knows something, understands something, or can actually do something.
Take a fairly ordinary corporate training example. An employee completes mandatory compliance training, takes the final assessment, and scores 90%. The LMS records the result. Training completed. Assessment passed. All good. Until the employee faces the kind of situation the training was supposed to prepare them for and has no idea what to do. There may be absolutely nothing wrong with that 90%. I still want to know what exactly we measured.
Was the assessment delivered and recorded as intended? Did it require the employee to demonstrate the knowledge or skill we care about? And do we have enough confidence that the result belongs to that employee and was achieved under the conditions we set?
These questions belong to different parts of the assessment process. In this article, I want to look at three of them (SCORM, interactivity, and integrity) and how they work together in an online assessment.
What SCORM does, and what it doesn’t need to do Our employee’s 90% has already done quite a bit of traveling. The assessment produced the score, passed it to the LMS, and the LMS recorded it against the employee’s training history. That part of the story matters too.
SCORM gives learning content and an LMS a standard way to exchange this kind of information. Depending on the version and implementation, that can include completion and pass/fail status, scores, and other data about the learner’s progress.
In our example, everything could be working exactly as intended. The employee completed the course. The assessment returned 90%. The LMS received 90%. The training record shows 90%. So far, we have established that 90% was recorded correctly.
How much you can learn about the activity behind a result depends partly on what your learning ecosystem tracks. SCORM and xAPI , for example, differ considerably in the learning activities and data they can capture. That becomes important when you need a richer picture of what learners do.
Our original problem starts one step earlier, though. Before deciding how much information to record about an assessment, we need an assessment worth recording.
SCORM has no opinion on whether a multiple-choice question was a good way to test the employee’s knowledge. It will happily report the result of a brilliant assessment and an utterly useless one.
I find that a useful boundary to keep in mind. A technically successful SCORM assessment can give us a reliable record of what happened in the course. The meaning of that record depends on what the learner had to do to produce it.
Interactivity: make learners demonstrate, not just recognize Back to our employee. We’ve already talked about them quite a lot, and we’re going to keep talking about them, so I think they deserve a name. Let’s call them Alex.
Alex scored 90% on the final assessment and still struggled to apply the training in practice. Suppose the assessment consisted mostly of questions asking Alex to recognize the correct policy, definition, or procedure from a list of options. They may well have answered every question correctly. The difficulty appeared later, when they had to recognize a situation in real life, decide which rule applied, and act on it. Those are different tasks.
This is why I tend to think about interactive assessment in terms of what the learner actually has to demonstrate. If the learning objective is remembering a policy, a straightforward multiple-choice question may be perfectly adequate. If Alex needs to apply that policy, make a judgment, troubleshoot a problem, or choose what to do next, the assessment should give them an opportunity to do exactly that.
That could mean a scenario with several possible responses, a sequence of decisions, matching actions to situations, putting steps in the correct order, identifying a problem in an image, or constructing an answer rather than selecting one. The 14 question types available in iSpring QuizMaker give us a lot more to work with than the good old radio buttons.
More interaction doesn’t automatically make an assessment better. Dragging things around the screen is still just dragging things around the screen if the activity has little to do with the skill we want to assess. Then that 90% begins to tell us something much more interesting.
Design for a result that’s harder to game Now let’s give Alex some colleagues. A few of them took the same compliance training last month. They know which questions appear in the final assessment. Maybe they even remember the answers. By the time Alex takes it, some very useful intelligence may already be circulating in Slack.
This is another reason I’m wary of treating a passing score as a self-explanatory result. Alex may know the material. Alex may also know that the answer to question 7 is B. Good assessment design can make that second route considerably less useful.
Question pools are an obvious place to start. Instead of giving every learner the same ten questions, you can draw a different set from a larger bank. Randomizing question order and shuffling answer choices add another layer. Depending on the purpose of the assessment, you can also limit attempts, set a time limit, and control whether learners can see correct answers after completing the test.
The questions themselves matter just as much. A policy definition is easy to pass around. A scenario that asks Alex to decide what to do with that policy is harder to reduce to “question 7 = B,” especially when the details vary across questions in the pool.
None of these controls will stop colleagues from talking to each other (great for team building, less great for online assessment integrity), taking screenshots, searching for answers, or using tools outside the assessment. They do make the assessment itself less predictable and reduce how much someone can gain simply by knowing what previous learners saw.
Not every assessment needs the same level of integrity Alex is taking mandatory compliance training. Imagine there’s a short knowledge check halfway through the course and a final assessment at the end.
The questions may cover much of the same material, but the results have different jobs. The knowledge check helps Alex see whether they understood the topic and where they might need another look. The final assessment may become part of the formal training record and serve as evidence that Alex met a requirement. That difference matters when deciding how much assessment integrity you need.
The same applies beyond Alex’s training. Think of a practice quiz, an internal skills check, a compliance assessment, or a professional certification exam. Each can benefit from thoughtful design, while the consequences attached to the result can be very different.
I like to think about integrity controls in proportion to those consequences. The more you are going to rely on a result — to document compliance, grant a certification, confirm a qualification, or make another consequential decision — the more confidence you will probably want in how that result was produced.
For Alex’s mid-course knowledge check, good questions and sensible assessment settings may do the job perfectly well. Once their 90% becomes formal evidence of something, we may want to know a little more about where it came from.
When assessment design isn’t enough Once we want more confidence in Alex’s 90%, I start asking questions that the assessment itself cannot answer.
Was Alex actually the person who completed it? Did someone help them? Were they looking up answers elsewhere? Did they use AI when the rules said they shouldn’t? And if something unusual happened during the assessment, would anyone know?
This is the territory of a proctored assessment.
Online proctoring can add another layer of evidence around an assessment: identity verification, controls over the testing environment, detection of events that may need attention, and records that can be reviewed afterward. Depending on the setup, that review can involve automated analysis, a human reviewer, or both.
The important word for me here is evidence. A flag that Alex switched tabs, another person appeared on camera, or a prohibited application was detected tells us what happened during the session. Someone can then look at that information in context instead of asking the software to decide whether Alex cheated.
This is also where SCORM, assessment design, and integrity finally meet in our example. The assessment asks Alex to demonstrate what they learned. The LMS records the result. And, when the stakes justify it, proctoring provides additional information about how that result was achieved.
In Alex’s case, we can keep the whole journey inside iSpring LMS . At OctoProctor, where I work, we’ve integrated proctoring directly with iSpring LMS , so Alex can move from the course to the assessment without being sent off into a completely separate process.
Their organization gets the additional integrity controls and evidence it needs around the assessment, while the learning experience remains connected to the LMS.
Putting it together in iSpring LMS For the person building the training, I’d work backward from what the assessment result needs to tell you:
Define what the learner needs to demonstrate. Start with the behavior, knowledge, or decision you actually care about.Choose interactions that let them demonstrate it. Give Alex something meaningful to do with what they learned.Set the assessment rules. Decide on question pools, randomization, attempts, timing, feedback, and other controls that make sense for this assessment.Publish and track it in the LMS. Make sure the completion and assessment data you need are recorded.Add an integrity layer where the stakes call for one. If you need stronger confidence in who completed the assessment and how, add the appropriate proctoring controls. In iSpring LMS, proctoring can be set up directly for the assessment , with the settings applied before Alex starts.From Alex’s perspective, most of this should simply feel like one learning experience. They take the course and move on to the assessment in iSpring LMS . The complexity belongs behind the scenes.
Before we finally let Alex go There’s a temptation with online assessments to keep adding things. More question types. More tracking. More controls. More data.
How much of that you actually need comes down to what you plan to do with the result. A quick knowledge check can stay a quick knowledge check. A result used as evidence of compliance, competence, certification, or qualification deserves more scrutiny.
For me, good secure online assessment design is largely about getting that proportion right. Use what helps the assessment do its job, and leave the rest out.
And Alex can finally have a break. We’ve made them work quite hard for that 90%.
for your mission-critical project
9 min
9 min
14 min
