Residual Research Teacher Guide Residual Research
Teacher Facilitation Guide | Tier 2 Intervention
Standard CO HS.S-ID.B.6.b
Instructional Focus
This small-group intervention helps students move beyond "eyeballing" a line of best fit. Students will learn that a residual is the vertical distance between a data point and the model. By analyzing a residual plot, students will identify whether a linear model is appropriate or if a non-linear model should be explored.
Materials Needed
Graphing Tech (Desmos/TI-84)
Fit Factor Worksheets
Straightedges & Pencils
Small Group Scaffolding
Common Misconceptions
The "Pattern" Paradox: Students often think a pattern in residuals is "good." Emphasize that a pattern means the model missed something predictable.Calculation Errors: Residual = Observed - Predicted. Students often flip these. Use the "y - y-hat" mnemonic.
Discussion Prompts
"If our model was 'perfect,' where would all the dots on the residual plot be?"
"I see a U-shape in our residuals. What does that tell us about our straight line?"
"Why is 'random scatter' our goal?"
Lesson Flow (30-40 mins)
5 min
Connect: Quick review of scatter plots and lines of best fit. Introduce the "Residual" as the "Error" or "Leftover" distance.
10 min
Explore (Tech): Students use Desmos/TI-84 to plot a linear regression for a "Curvy" dataset. Ask: "Does the line look okay? Now look at the residuals."
15 min
Analyze: Complete the Fit Factor Worksheet . Focus on identifying patterns (U-shapes vs. Random scatter).
5 min
Reflect & Assess: Group summary of findings and individual Exit Ticket.
Progress Monitoring Tracker
Record observations during the small group session to inform future Tier 2 or Tier 3 decisions.
| Student Name | Defines Residual
Qualitative Notes / Next Steps
Analysis Key for Teacher:
Random Scatter: The linear model is a good fit. Data is "honestly" varied.
U-Shape / Curve: The linear model is a poor fit. A non-linear model (quadratic/exponential) is needed.
Residual Research Slides Residual Research
Is your model actually a good fit?
Stats Lab: Tier 2
Beyond the Eyeball
A line might look like it goes through the middle of the dots...
"But how do we know if it's the right TYPE of model?"
Scatter Plot
What is a Residual?
\( y - \hat{y} \)
Observed
The actual data point recorded in the real world.
Predicted
Where the line (model) says the point should be.
Residual = The "Error" or vertical distance to the line.
The Secret Signal
Random Scatter
No clear pattern. This means your linear model is a good fit .
U-Shape / Curve
A clear pattern exists. This means your linear model is a poor fit .
If you see a pattern in the errors, there is a pattern the model missed!
GOOD
BAD
Tech Mission
Your Workflow:
1 Input your data points into Table (L1, L2) .
2 Calculate the Linear Regression (Line of Best Fit).
3 Select Plot Residuals to see the error map.
Launch Desmos or TI-84
Refer to your worksheet for the "Curvy Dataset" coordinates.
Lab Discussion
Prompt 1:
"Why is it possible for a line to look 'okay' on a scatter plot, but 'terrible' on a residual plot?"
Prompt 2:
"If you see a U-shape, what should you do next as a researcher?"
Fit Factor Worksheet Fit Factor Lab
Student Research Log
Researcher Name
Date
1 The Science of Residuals
A residual is the vertical distance between the data point and the line.
Residual = Observed ( \( y \) ) \( - \) Predicted ( \( \hat{y} \) )
2 Dataset: The "Mystery Curve"
Input these coordinates into your tech tool (L1 and L2) to find the line of best fit.
x y (Observed) Predicted (\( \hat{y} \)) Residual 1 2 2 5 3 10 4 17
Hint: After you enter the data in your calculator, find the regression line. Then, use that line to calculate the predicted values for each x!
My Regression Equation (\( y = mx + b \)):
3 Residual Visualization
Sketch your residual plot here:
X-AXIS (Values) RESIDUALS
Analyzing the Shape
What do you see in the dots above?
Random Scatter (Points everywhere)
Clear Pattern (U-Shape or Curve)
4 The Researcher's Conclusion
1. Based on your residual plot, is a linear function a good fit for this data? Explain how you know.
2. If you see a clear pattern in a residual plot, what does that tell you about the relationship between the x and y variables?
Lab Summary
A linear model is only a "good fit" if the residuals are randomly scattered around the horizontal axis. If there is a pattern (like a curve), it means the linear model is not appropriate, and you should try a different type of function!
Residual Research Answer Key Answer Key
Fit Factor Lab | Teacher Reference
Material ID: residual-fit-answer-key
1. Definitions
A residual is the vertical distance between the observed data point and the predicted line.
2. Dataset: The "Mystery Curve"
Regression Equation: \( \hat{y} = 5x - 4 \)
x y (Observed) Predicted (\( \hat{y} \)) Residual 1 2 1 1 2 5 6 -1 3 10 11 -1 4 17 16 1
3. Analysis & Conclusions
Shape Analysis:
Students should identify a Clear Pattern (U-Shape or Curve). The residuals go from positive to negative back to positive.
Conclusion Q1: Is linear a good fit?
"No. A linear function is not a good fit because the residual plot shows a distinct U-shape pattern rather than random scatter. This means the actual data is curved while the model is straight."
Conclusion Q2: What does a pattern tell you?
"It tells me that there is a predictable relationship in the data that the current model is missing. A curved pattern suggests that a non-linear function (like a quadratic) would be a better choice."
