Skip to content

Scientific Method Workflow

Overview

The scientific method is the systematic process that scientists use to investigate phenomena, acquire new knowledge, and correct or integrate existing knowledge. This interactive flowchart visualizes each step of the scientific method as used in physics and all scientific disciplines.

Interactive Diagram

View Fullscreen

<iframe src="main.html" width="100%" height="2140px" scrolling="no"></iframe>

Process Steps

  1. Observe Phenomenon or Ask Question - Notice something interesting about the physical world—like why objects fall, why the sky is blue, or how magnets work
  2. Background Research - Review existing scientific literature and previous experiments to understand what's already known
  3. Formulate Hypothesis - Create a testable statement predicting the relationship between variables (e.g., "If mass increases, then falling time decreases")
  4. Design Experiment - Plan controlled procedures, identify independent and dependent variables, consider controls and constants
  5. Conduct Experiment & Collect Data - Carefully follow procedures, make accurate measurements, record all observations systematically
  6. Analyze Data - Create graphs, calculate statistics, look for patterns and relationships in the measurements
  7. Does Data Support Hypothesis? - Compare experimental results to predictions—do they match within experimental error?
  8. Accept Hypothesis (if Yes) - Hypothesis is supported by evidence, but remains open to future testing and refinement
  9. Revise or Reject Hypothesis (if No) - Modify hypothesis based on findings or develop an entirely new hypothesis
  10. Communicate Results - Share findings through lab reports, presentations, or scientific papers using standard formats
  11. New Questions Raised? - Scientific investigation often leads to new questions and avenues for research

Understanding the Process

The scientific method is not a rigid linear process but rather a flexible, iterative approach to investigation. The flowchart illustrates several key features:

Core Steps

  1. Observation & Question - Scientific inquiry begins with curiosity about the natural world. In physics, this might involve noticing patterns in motion, energy, or forces.

  2. Background Research - Before conducting experiments, scientists review existing knowledge to understand what has already been discovered and to refine their questions.

  3. Hypothesis Formation - A hypothesis is a testable prediction about the relationship between variables. In physics, hypotheses often take the form of mathematical relationships (e.g., "acceleration is inversely proportional to mass").

  4. Experimental Design - Careful planning identifies independent variables (what you change), dependent variables (what you measure), and controlled variables (what you keep constant).

  5. Data Collection - Systematic observation and measurement following established procedures. Accuracy, precision, and reproducibility are essential.

  6. Data Analysis - Raw data is processed using graphs, statistical analysis, and mathematical models to identify patterns and relationships.

Decision Points

The diagram includes two critical decision points:

  • Does Data Support Hypothesis? - If experimental results match predictions within acceptable error margins, the hypothesis is supported. If not, it must be revised or rejected.

  • New Questions Raised? - Scientific investigation typically generates new questions, leading to additional cycles of inquiry.

The Iterative Nature

Notice the feedback loops in the diagram: - Failed hypotheses loop back to reformulation based on experimental evidence - Successful investigations often reveal new questions, restarting the cycle - Each iteration builds on previous knowledge, advancing scientific understanding

Key Concepts

  • Controlled Experiment: An investigation where one variable is changed while others are held constant
  • Independent Variable: The factor deliberately manipulated by the experimenter
  • Dependent Variable: The factor measured in response to changes in the independent variable
  • Hypothesis: A testable prediction about the relationship between variables
  • Reproducibility: The ability of an experiment to yield consistent results when repeated
  • Peer Review: The process of scientists evaluating each other's work before publication

Connection to Physics Learning

Throughout this course, you will apply the scientific method to explore:

  • Kinematics: Investigating relationships between position, velocity, and acceleration
  • Dynamics: Testing Newton's laws through force and motion experiments
  • Energy: Analyzing energy transformations and conservation principles
  • Waves & Optics: Examining wave behaviors and light phenomena

Each physics concept in this textbook was discovered and refined through countless iterations of this process.

These links do not exist! - [Experimental Error and Uncertainty](../../chapters/01/index.md) - Understanding measurement limitations - [Graphing and Data Analysis](../../chapters/01/index.md) - Visualizing experimental results - [Significant Figures](../../chapters/01/index.md) - Reporting measurements accurately

Further Exploration

Try applying the scientific method to these physics questions:

  1. Does the angle of a ramp affect how fast an object slides down?
  2. How does the length of a pendulum affect its period?
  3. What factors influence the range of a projectile?

For each question, work through the flowchart: formulate a hypothesis, design an experiment, and predict what data analysis might show.

What This Diagram Records — and What It Cannot

This MicroSim emits xAPI statements to the panel below the diagram (nothing is sent to a server). It conforms to the xAPI Producer Contract v1, and it is the reference non-p5 emitter — every other instrumented sim here is p5.js, so this one is what proves the contract is not tied to a single rendering library.

It cannot tell you whether a student understands the scientific method

This is worth stating plainly, because it is the thing people most want from a diagram like this and it is the thing hover data cannot give.

Hovering is not knowing. Only the answered verb carries result.success, and result.success is the sole input to attempts/successes and therefore to Bayesian Knowledge Tracing. Every statement this diagram emits is interacted or experienced, so it contributes statements_compressed at attempts = 0 — permanently, by design. A student can pause on all twelve steps and understand nothing; a student who already knows the method may pause on none.

What it does give is engagement evidence, which is genuinely useful and is not the same claim:

  • Coverage — which steps were studied, and which were never opened.
  • Dwell — how long each step held attention.
  • Order — the sequence steps were explored in.
  • Intensity — a hover is passing attention; a pinned click is deliberate study.

To measure understanding, this diagram needs questions. metadata.json has no pedagogical.keyQuestions yet; adding them, and wiring them the way the chapter quiz is wired, is what would produce answered statements and a real mastery signal.

Prior exposure: the same diagram lives in more than one textbook

This workflow appears in physics and chemistry courses too. That has a consequence worth understanding before reading any number this produces.

A statement's object.id is the activity's canonical published URL (contract §1) — so it is the same IRI no matter which textbook embeds the diagram. Only grouping[0], the textbook version IRI (§4), differs between books.

Now look at how the data is aggregated. Every rollup in clickhouse.sql is keyed (district_id, student_key, concept_id) or (district_id, student_key, object_id) — none of them key on textbook_id. So a student's physics exposure and their chemistry exposure merge into a single vertex, with dwell summed across both. The rollup cannot tell you the two visits happened in different courses.

That is not a bug. It is the two-store split working as designed: lrs.statements keeps textbook_id and version_id on every statement, so the question "had this student already met this diagram elsewhere?" is answerable from the immutable log. It is simply not answerable from ConceptMastery, because the graph deliberately holds compressed summaries rather than events.

The interpretation trap: a student who moves through this quickly because they met the scientific method in physics looks identical to a student who could not be bothered. Low engagement here is not evidence of low mastery — it may be evidence of prior mastery. The two are distinguishable only by asking the log whether earlier statements for these concepts exist under a different textbook_id.

How each interaction maps

You do this It emits Why
Sweep the mouse down the diagram nothing A crossing is not evidence. A hover must exceed 0.6s to count — otherwise a single mouse movement down a tall workflow would fabricate twelve statements.
Pause on a step (>0.6s) one interacted, engagement-mode: hover Transient attention, with the step's dwell time.
Click a step to pin it one interacted, engagement-mode: pinned Unambiguous intent — pinning is a decision, so it counts regardless of dwell.
Leave the page / hide the tab one experienced with result.duration Page-level dwell (contract §7). This diagram has no Start/Pause control, so the interval is simply time on the page.

Full vs. Compact, Simulate Done, and View Formatted JSON

The xAPI panel under the diagram has the same controls as the Bouncing Ball and Sine Wave sims:

  • xAPI events: Full / Compact. Full (the default) is the stream in the table above: one statement per step studied, sent as it happens, the way a full LRS ingests it. Compact is LRS-Lite's stream: the same hovers and pins are folded into one session, and when the diagram loses focus it emits one experienced summary. Its controls extension lists each step studied with its dwell (ms), how it was studied (modes: hover or pinned), and its concept. statements_represented counts the Full statements it stands for.
  • Simulate Done does what the host page does when the reader moves on (switches tabs, leaves, scrolls away, or goes idle). In Compact it ends the session and emits the summary. In Full it closes the page-level dwell interval and emits that experienced statement, exactly as hiding the tab would. This is the difference from Sine Wave: a diagram has no Start/Pause, but it does have time on the page, and in Full mode that is the one thing not already sent.
  • View Formatted JSON ↗ opens the most recent statement in a new tab, pretty-printed, with the compact-only fields highlighted and a table explaining each field. Clicking any line in the log opens that statement instead.

Switching modes never loses or double-counts dwell. Each mode keeps its own record of time on the page: Full mode keeps the page interval, and Compact mode keeps the session summary. Switching from Full to Compact emits Full mode's page interval so far (run-ended-by: "mode-switch"). Switching back emits the compact summary (end_reason: "mode-switch"), and a fresh page interval starts from then. make test-sims checks both modes, both switches, Simulate Done, and the formatted view.

Where the mode comes from

Two config files decide the mode, and an agent can change either one:

  • The textbook's docs/js/lrs-config.js sets the default for every MicroSim in the book. Here it is xapi: { compact: true, teaching: false }: compact, and silent.
  • This sim's metadata.json xapi block overrides it for this one sim:
"xapi": { "compact": false, "teaching": true, "idleMs": 90000, "offscreenMs": 10000, "blurMs": 30000 }

teaching: true is what shows the statement log, the Full/Compact switch, Simulate Done, and View Formatted JSON. They exist only to teach what xAPI events are. A production MicroSim leaves teaching off: it emits the same statements, silently, in whichever mode its config says. This diagram starts on Full, like the other two teaching sims.

Why steps are Control and the page is MicroSim: the page-level statement's object is the page IRI with no fragment, so it becomes exactly one PageEngagement vertex. Each step's object is fragment-qualified (…/scientific-method/#hypothesis) and typed Control, which keeps it out of mv_student_page_rollup — that view groups by object_id, so MicroSim-typed steps would produce twelve PageEngagement vertices for a page visited once.

Twelve steps, nine concepts. Several steps share a concept — Does Data Support Hypothesis?, Accept Hypothesis, and Revise or Reject Hypothesis all carry hypothesis-testing. They compress into one ConceptMastery vertex, which is precisely what the (student, concept) grain exists to do.