Rethinking Evidence of Learning

Three days after publishing “The Illusion of Learning: How Do You Know You Are Learning?”, I realized something surprising. The framework itself needed revision. This was not because the original question had failed. It was because further reading, discussion, and reflection changed how I interpreted the evidence.

Version 1.0 proposed six observable behaviors that might support the inference that learning occurred. The signals were presented from weaker to stronger evidence for explanatory purposes. They were not intended to describe a chronological learning sequence.

• Observation
• Imitate
• Perform Independently
• Transfer
• Teach
• Creation

While continuing to read about cognition, psychology, and learning, I began questioning one assumption in the original framework: Does observation itself provide any observable evidence that learning occurred? Observation may create an opportunity for learning, but opportunity is not evidence that a lasting change occurred.

Imagine a child watching a football game for twenty-four hours. Has the child learned? The observation alone does not allow us to answer. The child may have been attentive, distracted, confused, asleep, or already familiar with the game. The child may have learned the rules, remembered a player, noticed nothing, or simply remained exposed to the event.

• Observation confirms exposure
• Observation does not demonstrate retention
• Observation does not demonstrate understanding
• Observation does not demonstrate application
• Observation does not identify what changed

The first version treated Observation as the weakest evidence of learning. The revision clarified that observation is not merely weak evidence. By itself, it may provide little or no behavioral evidence that learning occurred. It tells us that information was available to the learner. It does not show that the information became retrievable, usable, transferable, or integrated.

The purpose of EoL is to organize observable behaviors that support an inference about learning. Because Observation did not demonstrate any observable change in knowledge, understanding, or ability, it no longer fit the purpose of the framework. Observation may remain part of the learning environment, but it should not automatically be treated as evidence of learning.

• Necessary conditions are not always evidence
• Exposure may occur without meaningful processing
• Presence does not demonstrate internal change
• Attention itself remains difficult to infer
• The framework should begin with a demonstrated consequence

Recall is the conscious retrieval of previously encountered information, ideas, or experiences. When a learner recalls something, there is at least some evidence that information remained accessible beyond the original exposure. Recall therefore provides more direct behavioral evidence than observation alone.

Recall demonstrates that something from the previous experience can be retrieved. However, it remains relatively low-weight evidence. A learner may recall inaccurate information, reproduce memorized words without understanding, combine unrelated memories, or depend heavily on cues. Recall increases confidence that some learning occurred, but it does not establish the depth, accuracy, or transferability of that learning.

• Information remains retrievable
• Retrieval extends beyond direct exposure
• Recall may be accurate or inaccurate
• Memorization does not guarantee understanding
• Delayed recall may provide stronger evidence than immediate recall

Reflection generally builds upon recalled information. A person retrieves a previous experience, decision, idea, or piece of knowledge and then analyzes, reorganizes, compares, or reinterprets it through current understanding. In simple terms, reflection can be understood as recall combined with analysis.

Imagine someone recalling a past decision. Recall brings the event back into awareness. Reflection begins when the person asks what might have happened under another decision, identifies a weak assumption, compares the past model with new evidence, or changes how the event is understood. Reflection provides evidence that prior information is not only accessible but available for evaluation and model revision.

• Retrieve a prior event or idea
• Analyze the original decision
• Compare alternatives
• Identify mistakes or assumptions
• Reinterpret the experience
• Update the internal model
• Recognize remaining uncertainty

Reflection may involve previously acquired knowledge, newly acquired information, or the relationship between the two. A learner may revisit an old experience after encountering a new concept and reinterpret what happened. This suggests that learning is not only the accumulation of information. It can also involve reorganizing the meaning of previous information.

Version 1.1 removed Observation and introduced Recall and Reflection. The revised Evidence of Learning framework contains seven independent behavioral signals.

• Recall
• Reflection
• Imitate
• Perform Independently
• Transfer
• Teach
• Creation

The Evidence of Learning framework does not describe the chronological stages through which learning must progress. A learner may demonstrate one, several, or all of the signals depending on the task, domain, objective, environment, prior knowledge, and available support. The framework organizes evidence. It does not prescribe one universal path of learning.

• Recall does not always occur first
• Reflection can appear at several moments
• Imitation may be irrelevant to some learning tasks
• Transfer may occur without teaching
• Creation may reveal gaps rather than mastery
• The relevance of each signal depends on context

The revision changed the starting point of the framework. Version 1.0 began with exposure. Version 1.1 begins with an observable consequence of exposure.

• Removed — Observation
• Added — Recall
• Added — Reflection
• Preserved — Imitate
• Preserved — Perform Independently
• Preserved — Transfer
• Preserved — Teach
• Preserved — Creation

Revising the framework does not mean the first version had no value. Version 1.0 preserved the first structured interpretation of the problem. Version 1.1 records how that interpretation changed after further evidence and reasoning. Keeping both versions makes the intellectual development visible.

• The original assumption remains documented
• The cause of the revision is explicit
• The framework has a visible history
• Readers can compare both models
• Revision becomes evidence of changing understanding

The most important change did not come from adding more detail to the later parts of the framework. It came from returning to its beginning and questioning the assumption that Observation belonged there. Instead of polishing Version 1.0 around an unstable foundation, I reconsidered the original premise.

• The first assumption shaped the entire framework
• A necessary process was confused with observable evidence
• The problem appeared at the starting boundary
• Questioning Version 0 produced the major revision

The question that began this journey was: “How do you know you are learning?” Ironically, revising the framework may itself provide evidence that my own learning occurred. I recalled the original model, reflected on its assumptions, compared it with new information, revised the structure, explained the revision, and created a new version.

• Recall — retrieve the original framework
• Reflection — analyze the Observation assumption
• Transfer — apply new learning to the existing model
• Teach — publicly explain the revision
• Creation — produce EoL v1.1

I hope the framework continues to evolve. Future changes should not occur simply to make the framework appear new. They should occur when stronger evidence, better definitions, deeper reflection, or more useful distinctions reveal a limitation in the current version. A framework should not remain unchanged merely because it has already been published.

Version 1.1 improves the interpretation of the earliest evidence signals, but it remains a conceptual framework. The signals do not yet have universal definitions, validated weights, or task-specific operational criteria.

• Recall may vary in accuracy and durability
• Reflection may be difficult to observe externally
• The signals may overlap
• Evidence strength depends on the learning objective
• Creation does not guarantee correctness
• No empirical validation has established a universal ordering

This essay was originally published on Medium on June 22, 2026.

• https://medium.com/@nirajtechx/rethinking-evidence-of-learning-bff94529abd7
• It documents the transition from EoL v1.0 to EoL v1.1.

The original Evidence of Learning framework appeared in “The Illusion of Learning: How Do You Know You Are Learning?” Original article: https://nirajtechx.medium.com/the-illusion-of-learning-how-do-you-know-you-are-learning-bb8796b3ea35 Reading both versions preserves the full development of the idea rather than presenting v1.1 as though it appeared fully formed.