Regression Reality Lesson Plan Regression Reality
Science & Math Interdisciplinary Lesson Plan
Subject: Algebra II / Physics
Duration: 50-60 Minutes
Grade Level: 9-12
Learning Objectives
Identify linear vs. exponential trends in physical experimental data.
Perform regression analysis using GeoGebra tools.
Interpret the R-squared (\(R^2\)) value as a measure of "goodness of fit."
Explain why real-world data deviates from theoretical models.
Materials Needed
Student laptops/tablets with internet access.
GeoGebra Graphing Calculator (geogebra.org/calculator).
Lab data (Pendulum period or Ball bounce height).
"The Messy Lab" Worksheet & "GeoGebra Wizard" Guide.
Instructional Sequence
05 min
The Context: Real Labs are Messy
Recap a recent science experiment (e.g., pendulum swing or ball bounce height). Ask: "If we graphed every single trial, would it make a perfectly smooth line? Why or why not?" Introduce the concept that math models are "ideals" while lab data is "real."
05 min
Video: The Power of Regression
Watch "Approximating Functions for Data" (focus on 5:47–6:37). Discuss the \(R^2\) scale (0 to 1). Define 1.0 as a perfect fit and 0 as no relationship.
30 min
Lab Analysis & GeoGebra
Students input their lab data into GeoGebra. They must run both `FitLine` and `FitGrowth` commands and use `RSquare` to determine which model is mathematically superior for their specific dataset.
05 min
Discussion: Why isn't it 1.0?
Class discussion on experimental error. Source identification: Human reaction time (stopwatch), air resistance, friction, parallax error. Connect these physics realities to the mathematical "imperfection" of the fit.
10 min
Extension: Beyond Linear
For gravity experiments (height vs. time), introduce the `FitPoly(L1, 2)` command for quadratic regression. Compare this to the kinematics equation \(h = \frac{1}{2}gt^2\).
Essential Vocabulary
Regression
The statistical process of estimating the relationship between variables to find a "line of best fit."
R-Squared (\(R^2\))
The "Coefficient of Determination." Measures how much of the data variance is explained by the model.
Residual
The vertical distance between a data point and the regression line (the "error" for that point).
Teaching Tips & Differentiation
Scaffolding: Provide the "GeoGebra Wizard" guide to all students. For those struggling with syntax, pre-create a GeoGebra file with the list `L1 = {}` already defined.
Common Misconception: Students often think a "high" \(R^2\) (like 0.8) means the model is correct. Emphasize that context matters—in physics, we expect >0.95.
Extension: Challenge advanced students to calculate one residual by hand: \(y_{observed} - y_{predicted}\).
Video Note: Remind students at 3:38 to use `FitGrowth`, not `FitExp`, as GeoGebra's `FitExp` defaults to base \(e\).
Fit Finder Slides Regression Reality
Bridging the Gap Between Math & Science
Algebra II
Physics
The "Messy" Reality
Theoretical World
Equations are perfect. Every dot lands exactly on the line. No friction. No wind. No mistakes.
Experimental World
Data is "approximate." Points are scattered. Human error and physics realities make things messy.
"In real life, data usually doesn't fit perfectly on a function."
Video Spotlight
Embedded media
Focus: R-Squared Analysis (5:47 – 6:37)
Measuring "Goodness of Fit"
R²
Coefficient of Determination
1.0 = Perfect Fit (The model hits every point)
0.9+ = Strong Fit (Common in Science labs)
0.0 = No Fit (The model is useless for this data)
"The higher the R², the better the model explains the data."
GeoGebra Command Center
FitLine(L1)
Creates the best linear model for your list.
FitGrowth(L1)
Creates the best exponential growth model.
RSquare(L1, f)
Calculates the R² for function f using list L1.
WARNING: Do not use FitExp—it uses base e!
Why isn't R² = 1.0?
If the physics formula says the relationship is perfect, why does your calculator say it's 0.985?
Human Reaction Time
Air Resistance
Instrument Precision
Extension: Gravity & Curves
If you drop a ball, the height vs. time isn't linear or exponential—it's Quadratic.
GeoGebra Command:
FitPoly(L1, 2)
The Physics Connection:
\[ h = \frac{1}{2}gt^2 \]
Does your R² value for a quadratic fit beat your linear and exponential values for falling data?
Messy Lab Worksheet The Messy Lab
Regression Analysis & Experimental Fit
Name:
Date:
1
Raw Lab Data
Record your measurements from the lab below. Identify your independent (x) and dependent (y) variables.
Independent Variable (x):
Dependent Variable (y):
Trial # x-value (units: ) y-value (units: ) 1 2 3 4 5 6 7 8
2
GeoGebra Analysis
Linear Model (FitLine)
Equation:
R-Squared (\(R^2\)):
Exponential Model (FitGrowth)
Equation:
R-Squared (\(R^2\)):
3
Conclusion & Discussion
1. Which model mathematically fits your data better? How do you know? Refer to specific \(R^2\) values.
2. In physics, we often expect a "perfect" fit (\(R^2 = 1.0\)). Why did your experiment fall short of this? List 2 specific sources of "Experimental Error."
Extension: Quadratic Check
If your data represents falling objects or gravity, run `FitPoly(L1, 2)`.
Quadratic \(R^2\):
GeoGebra Wizard Guide GeoGebra Wizard
v2.0 // Regression Command Reference
01
Build Your List
You must group your data points into a single "list" variable before running regression commands.
Input: L1 = {(0, 10), (1, 15), (2, 22)}
* Use curly brackets { } and separate points with commas.
02
The Regression Commands
Enter these into the input bar to create your model. GeoGebra will automatically name the functions f(x), g(x), etc.
Linear Model
FitLine(L1)
Exponential Model
FitGrowth(L1)
DO NOT USE FitExp(L1).
03
Check the Quality (R²)
To see which model is better, calculate the R-Squared for each function.
RSquare(L1, f)
RSquare(L1, g)
Pro-Tips for Data Detectives
Window: Click the gear to adjust x-min and y-max if you can't see your points.
Rounding: Don't round decimals in your head! GeoGebra keeps them precise for a reason.
Extending: For gravity labs, try FitPoly(L1, 2) to find the quadratic fit.
Reference Document // Science-Math-Interdisciplinary GeoGebra Technical Specs // 2026.01.19
