Learning analytics basics

Welcome — this topic is a compact, practical guide for teachers who want to use classroom data wisely. No dashboards that suck your time. Just simple, meaningful measures you can collect, interpret and act on — in the spirit of the Top Teacher approach: student-centered, formative, respectful of learners’ prior knowledge, motivation and self‑esteem.

Think of this as "low-cost analytics": small, regular evidence that helps you tune teaching, strengthen feedback, and support individual learners.

---

Why bother with learning analytics?

Briefly:

• It helps you find what students already know (and what they don’t) — Ausubel and Piaget insist: learning must anchor to prior knowledge.
• It makes formative assessment practical and focused — feedback that actually improves learning.
• It reveals patterns (who’s confident, who’s shaky, who’s slipping) so you can intervene before rapport or self‑esteem suffers.
• You can use simple numbers (average, spread, gain) as conversation-starters with students and parents — not as labels.

Goal: collect a few reliable signals that guide your next lesson — not measure everything.

---

Keep it simple: choose up to 3 indicators

Limit yourself. Pick three indicators that match your lesson goals. Example sets:

Option A — Foundation check

• Pre‑test average (quick 5 Q)
• Exit‑ticket correctness (3 Q)
• Student confidence (self-rated 1–4)

Option B — Skills & transfer

• Mastery grid (skill checklist: yes/partly/no)
• Ratio of students scoring ≥ target
• SD (spread) of test scores

Option C — Engagement & process

• Attendance/participation
• Number of formative attempts (e.g., quiz retakes)
• Quality of self‑reflection journal entries (rubric 0–3)

Why 3? Because anything more becomes hard to interpret and act upon without extra time or support.

---

What to collect (easy, low friction)

• Micro‑pretests (3–7 items) to check prior knowledge
• 3‑question exit tickets:
  1. One factual check (correct/incorrect)
  2. One connection to previous learning (short answer)
  3. One confidence self-rating (I’m confident / somewhat / unsure / lost)
• Short polls (single multiple‑choice) during class
• Observation checklist (2–5 items) when circulating (e.g., “on task”, “collaborating”, “asks questions”)
• Simple rubrics for one main skill (0–3)
• Quick peer/self‑assessments (e.g., “I can explain this to a classmate: yes/no/maybe”)
• LMS quiz logs or Google Form responses (automated capture)
• Attendance + activity completion (homework or digital task)

Tools: paper slips, sticky notes, Google Forms, Kahoot/Quizzes, your LMS gradebook, a simple spreadsheet.

---

How to collect without overload

• Automate where possible: Google Forms → spreadsheet; LMS quizzes → reports.
• Timebox: collect 3 exit tickets in 5 minutes at lesson end.
• Rotate deep checks: not every lesson needs a pretest. Use sampling (e.g., one group per week).
• Reuse: same 3 exit‑ticket questions format saves marking time.
• Delegate: students can collect peer‑feedback, or use self‑assessment to generate class summary.
• Visual cues only: use traffic lights or sticky notes for quick confidence checks.

---

Quick analyses you can do (no statistics degree needed)

1. Class average — snapshot of how the group did.
2. Standard deviation (SD) — how spread out scores are. Interpretation:
   • Small SD + high average = most learned it.
   • Small SD + low average = everyone struggled → reteach/adjust.
   • Large SD = mixed learning: differentiate; some students need extra support.
     (Petri Lounaskorpi: large dispersion often means teaching didn’t reach all learners.)
3. Item difficulty — percent correct per question. Helps spot misunderstood content.
4. Pre/post gain — % correct on pretest vs. posttest; shows learning progress.
5. Mastery grid counts — how many have “yes/partly/no” per skill.
6. Confidence vs. correctness cross‑tab — reveal over/underconfidence.

You can compute average and SD in Google Sheets:

• AVERAGE(range)
• STDEV.S(range)

Item difficulty: =COUNTIF(range,"correct")/COUNTA(range)

---

Interpreting results — rules of thumb and actions

Use the data to answer these teacher‑friendly questions:

1. Did most students reach the lesson objective?

   • Yes → move on or deepen.
   • No → reteach, change task, provide worked examples.

2. Is there a big spread (high SD)?

   • Target the lower group with scaffolds; offer extension tasks to high performers.
   • Consider whether the assessment measured application (transfer) or rote memory.

3. Are students confident but wrong?

   • That’s a metacognition gap — build reflective discussion, confidence‑calibration exercises.

4. Are students unsure but mostly correct?

   • Boost self‑esteem with positive feedback and low‑stakes practice — they need reassurance.

5. Item(s) with low correct %

   • Reteach that specific concept, provide different content/context (Vygotsky: scaffolding helps).

6. Whole class low average

   • Check alignment: was the test harder than instruction? (Lounaskorpi: large dispersion or mismatch could reflect test fault.)

Action menu (pick 1–2):

• Short reteach using a different modality (visual, hands‑on, analogy).
• Small group instruction for the lower band.
• Pair high + low students for structured peer teaching (social constructivism).
• Give targeted homework with immediate feedback.
• Reassess with a 3‑question follow‑up next lesson.

---

Quick workflows you can use (weekly micro‑cycle)

Micro‑cycle (for each lesson)

1. Before: 3‑question pretest (2–3 min) or quick diagnostic for new topic.
2. During: one poll/checkpoint; observe 2–3 learners with a short checklist.
3. After: 3‑question exit ticket (5 min).
4. Within 24 hours: glance at spreadsheet, compute average and SD, look at item difficulty (10–15 min).
5. Plan next lesson: one targeted action (reteach, small groups, enrich).

Monthly deep cycle

• Pre/post for a unit, mastery grid across all skills, attendance trends, and a short student survey about motivation and self‑esteem.
• Share results with students as class goals, and with colleagues for pedagogy adjustments.

---

Templates you can copy (use immediately)

Exit ticket (3 Q)

1. One graded question (MCQ or short answer) — 1 point.
2. "How does this connect to something we learned before?" (1–2 sentence)
3. Confidence: circle one — confident / somewhat / unsure / lost

Observation checklist (on clipboard, tick when circulating)

• On task (Y/N)
• Asking or answering questions (Y/N)
• Collaborating appropriately (Y/N)
• One note (1 sentence) about a student’s struggle or success

Simple spreadsheet columns

• Student | Pre% | Post% | Q1 correct (1/0) | Q2 correct | Q3 correct | Confidence avg | Notes
• Add conditional formatting: red/amber/green for quick visual.

Mastery grid (skills vs. students)

• Rows = students; Columns = skills; cells = Yes / Partly / No → use counts at top for class snapshot

---

How to present feedback so it protects self‑esteem

• Always start with what the student can do — build from strengths.
• Offer one clear next step, not a laundry list.
• Use formative grades as descriptors, not judgments (e.g., “Skill A: Partly — next: review strategy X”).
• Celebrate small gains publicly (group level) and give private, targeted feedback when needed.
• Encourage metacognition: ask “What helped you learn this?” and “What will you try next?”

This follows Lounaskorpi’s emphasis: strengthen self‑esteem first, then motivation and learning follow.

---

Avoiding common pitfalls (overload + misinterpretation)

• Don’t track everything. Choose 3 meaningful indicators and stick with them.
• Don’t confuse activity with learning. Completion ≠ understanding.
• Watch for assessment drift: “harder test than instruction” gives false signals.
• Don’t punish low confidence — use it as diagnostic information to scaffold.
• Respect privacy: anonymize data where possible, be transparent with students and parents.

---

Ethics & communication

• Tell students why you collect data and how it helps them.
• Ask for permission if you share identifiable data outside school.
• Use data to support learning, not label children.
• Keep parents informed about patterns and next steps — focus on growth.

---

Quick example: item analysis in 10 minutes

1. Export quiz responses to Google Sheets.
2. Create a column per question with 1 = correct, 0 = incorrect.
3. Add a row at top: item difficulty = =AVERAGE(column)
4. Add class average = AVERAGE(total_score_range)
5. Add SD = STDEV.S(total_score_range)
6. Interpret: Item difficulty under 0.6 → teach again in different way; SD > threshold (e.g., > 2 on a 10‑pt test) → mixed learning.

(If you want, I can give you an exact sample spreadsheet that does this automatically.)

---

Final checklist for low‑overhead learning analytics

• [ ] Pick 3 indicators for now.
• [ ] Use a simple, repeatable exit ticket every lesson.
• [ ] Automate capture (Forms/LMS) when possible.
• [ ] Analyze weekly: average, SD, item difficulty.
• [ ] Choose 1 targeted instructional move each week.
• [ ] Give feedback that builds competence and self‑esteem.
• [ ] Keep parents/students informed about growth.

---

Small task to try in your next lesson (5–30 minutes)

1. Create a 3‑question exit ticket (use the template above).
2. Collect exit tickets at the end of class.
3. In Google Sheets, compute class average, confidence average, and one item difficulty.
4. Decide one concrete action for the next lesson (e.g., group reteach for 4 students, or a mini‑lesson addressing the most missed item).
5. Reflect: did the students’ confidence match correctness? Note one thing you’ll change.

If you want, paste your exit ticket questions here and I’ll help turn them into a Google Form + spreadsheet template.

---

Want templates (exit ticket form + spreadsheet) or a one‑page printable mastery grid? I can create them for you — tell me which grade/subject and I’ll make a ready‑to‑use file.