✳ IskraAI

Find the Gap: How to Set Up a Task to Identify Misunderstandings

The "Find the Gap" tool helps teachers dissect student answers and see exactly where misunderstanding lies, rather than just a wrong result. It accepts two text fields: "Source Material" and "Goal or Desired Result". The result is output as a single text in a separate panel, which can be copied or downloaded as TXT. This is not homework checking "for a grade" and not a course generator: the tool analyzes what you insert and suggests what to pay attention to and what tasks to give next. The quality of the analysis depends almost entirely on how accurately you describe the context, goal, format, and limitations. Below is how to do it in practice.

Open tool: Knowledge gap finder ↗

What the Tool Does and What to Expect

"Find the Gap" receives text with student answers and a learning goal, then forms an analysis: which fragments indicate persistent misunderstanding, which errors look random, and which exercises can close the detected gap. This is a text tool: on output you get connected text, not an interactive trainer and not automatic scoring with points.

Understanding the limits is important. The tool does not know your program, does not remember previous lessons, and does not check facts for you. It relies only on the inserted material and the goal formulation. If the source text contains no examples of student reasoning, the analysis will be superficial: from one wrong answer you cannot reliably distinguish misunderstanding from a typo.

The result may contain errors, and availability may vary. Therefore, the analysis should be read as a draft of hypotheses: the teacher themselves decides which conclusions to accept, and which to verify with additional questions at the lesson.

Example formulation: "Below are 7th grade student answers on the topic of percentages. Goal: teach choosing the correct operation with successive discounts. Need: indicate 2-3 probable misunderstandings with quotes and suggest one short exercise for each."

The "Source Material" Field: What to Insert

In this field, place the student's actual answers, not their summary. It is useful to add minimal context: class or level, topic, formulations of problems the student solved. Then the tool sees which question the answer was given to, and can link the error to a specific reasoning step.

Better to insert verbatim fragments, including intermediate calculations and explanations. These exactly show the thought process. If a student wrote "20% + 10% = 30%", this is more valuable for analysis than a mark "incorrect answer". At the same time, remove the superfluous: other people's work, grades, personal data that are not needed for analysis.

The field is limited in volume, so select the most informative. If there is a lot of material, divide it into several runs by topics or types of tasks - so the analysis will be more specific than trying to fit an entire term into one query.

Source material fragment: "Problem: item 1000 rubles, discount 20%, then another 10%. Student answer: 1000 - 30% = 700 rubles. Explanation: percentages are added, because both discounts are on one item."

The "Goal": How to Formulate the Result

The goal is not the lesson topic, but what you want to get from the analysis. It is useful to indicate three things: which skill is being checked, in what form the answer is needed, and what limitations to consider. For example: "determine if the student confuses percentage of a number and percentage change; provide a list of three exercises; no theory, only tasks."

The more specific the goal, the fewer general advice. Formulation "help with math" will give vague text. Formulation "find at which step the meaning is lost when working with successive discounts, and suggest exercises distinguishing additive and multiplicative approach" directs the analysis.

Indicate limitations: student level, permissible volume, language of answer, whether it is necessary to explain the error cause or only give practice. If you are preparing a 20-minute lesson, write it - otherwise suggested exercises may turn out too long.

Poor goal: "analyze errors". Good goal: "identify no more than two key misunderstandings on the percentages topic, confirm each with a quote from the answers and give two short exercises for each, 7th grade level, no formulas more complex than multiplication."

Example of Bad and Good Goal Setting Aloud

Bad option: in the source material field one line "student did not understand percentages", in the goal field - "give exercises". The tool will not see the thought process or level, so it will output general exercises, which may not match the real gap. Such a launch will almost certainly have to be repeated.

A good option contains context, verbatim answers, goal and limitations. For example, in the source material they insert three tasks with student answers and explanations, and in the goal they write: "Determine exactly where the logic breaks: in translating percentages to a fraction, in choosing the base or in adding results. For each case give a quote and one correction exercise. Answer - bulleted list, no more than 400 words."

The difference is not in the politeness of the formulation, but in the volume of information for analysis. The tool works with what was given: the clearer the gap you are looking for is described, the more useful the suggested practice will be.

Scenario (fictional example): a teacher inserts student answers on the "percentages" topic, where the latter adds successive discounts, and asks to find the misunderstanding and give exercises distinguishing the base of calculation from simple carelessness.

How to Check the Analysis Result

First, check whether every conclusion relies on a quote from your material. If the tool claims the student confuses concepts but does not show where this is visible, the conclusion is worth rechecking with a question at the lesson. It is useful to prepare two-three clarifying questions in advance for controversial spots.

Then evaluate the suggested practice: does it correspond to the level, is it solvable in reasonable time, does it really train the found gap, not a neighboring topic. Exercises are worth adapting to your class and removing those that duplicate already mastered material.

Finally, compare the analysis with your own observation. The tool sees only text, the teacher - student reaction, pace, questions. If conclusions diverge, trust the combined data, not one source.

Checking question for the student: "Discount 20% and discount 10% - from which sum is the second discount calculated?" The answer will show whether the matter really is in choosing the base.

Limitations and Typical Mistakes in Setting the Task

The tool does not replace diagnostic conversation and does not guarantee that the found gap is the only one. It analyzes the provided text and may miss the cause that manifests only in oral work or in another topic. Do not perceive its output as final diagnosis.

A frequent mistake - inserting too heterogeneous material. Then the analysis disperses, and suggested exercises turn out unrelated to each other. Another mistake - formulating the goal as "make it nice" or "check everything": without criteria the result is hard to use.

Also do not expect the tool to invent a month-long program or evaluate work by points. Its task is to help see misunderstanding and point to targeted practice. The decision about what and how to give further remains with the teacher.

If after the analysis doubts remain, run the tool again with refined goal: "check the hypothesis that the student chooses the base of calculation incorrectly, and suggest an exercise that distinguishes this error from simple carelessness."

An exact task for "Find the Gap" is built from two parts: verbatim student answers with minimal context and a clear goal with format and limitations. The more specifically you describe which exact gap you are looking for and in what form the result is needed, the more useful the analysis and suggested practice will be. Check conclusions by quotes, adapt exercises to your class, and remember that the tool gives hypotheses, not final diagnosis.

Open tool: Knowledge gap finder ↗

Interface translations are machine-assisted and may need refinement.