01 / 21

Interpreting
Linear Models

Learn to read \( y = mx + b \) as a real-world model, understand what slope and y-intercept mean in context, and recognize when extrapolation goes too far.

Pre-Algebra

What Is a Linear Model?

A linear model is an equation of the form \( y = mx + b \) that describes a straight-line relationship between two real-world quantities. The equation is built from data, and each part tells a story about the situation it represents.

\( m \) : Slope
The rate of change: how much \( y \) increases (or decreases) for every 1-unit increase in \( x \).
\( b \) : Y-Intercept
The starting value: the value of \( y \) when \( x = 0 \).

Seeing It on a Graph

Hours Studying (x) Test Score (y) 0 1 2 3 4 50 60 70 80 90 \( b = 54 \) +1 hr +8 pts

Interpreting Slope in Context

The slope \( m \) is more than just a number: it describes a rate of change in real-world units. Always translate it into a sentence:

Positive Slope
As \( x \) goes up, \( y \) goes up. Example: more hours practiced → higher score.
Negative Slope
As \( x \) goes up, \( y \) goes down. Example: more miles driven → less gas in tank.

Interpreting the Y-Intercept

The y-intercept \( b \) is the predicted value of \( y \) when \( x = 0 \). In context, it answers the question: "What do we start with?"

1
Substitute \( x = 0 \) into \( y = mx + b \) to get \( y = b \).
2
Translate to the scenario: "When no [x-quantity] has occurred, the model predicts [y-quantity] is \( b \)."
3
Check if it makes sense. Does \( x = 0 \) actually occur in the real situation? If not, the y-intercept may not have practical meaning.

The Core Rule: Reading \( y = mx + b \)

\( y = m \cdot x + b \)
 Rate of change (slope) Starting value (y-intercept)

Interpolation vs. Extrapolation

Data Range Extrapolation ? \(x\) \(y\)
Interpolation
Predicting within the range of your data. Generally reliable because you have nearby evidence.
Extrapolation
Predicting beyond the range of your data. Risky: the real pattern might change outside what was observed.

When Extrapolation Goes Wrong

A linear model that fits your data perfectly today can produce absurd predictions if you push \( x \) far outside the observed range. Here is why:

Age (years) Height (cm) 0 5 10 15 20 50 100 150 200 250 229 cm?! (7 ft 6 in)

Reasonable vs. Unreasonable Predictions

Reasonable Unreasonable
Where? Within or near the data range Far beyond the data range
Why? Supported by observed evidence No data to confirm the pattern continues
Physical sense? Result is realistic May give negative values, impossibly large values, etc.
Example Predicting score after 3 hrs study (data: 0-5 hrs) Predicting score after 100 hrs study

How to Interpret a Linear Model

1
Identify the variables. What real-world quantity does \( x \) represent? What does \( y \) represent?
2
Interpret the slope \( m \). Write a sentence: "For every 1 [x-unit] increase, [y-quantity] changes by \( m \) [y-units]."
3
Interpret the y-intercept \( b \). "When [x-quantity] is 0, the model predicts [y-quantity] is \( b \)." Check: does \( x = 0 \) make sense?
4
Make predictions. Substitute a value for \( x \) and solve for \( y \).
5
Check for extrapolation. Is the \( x \)-value inside the data range? If not, flag the prediction as potentially unreliable.

Example 1: Lemonade Stand

A student tracks daily temperature (\( x \), in °F) and cups of lemonade sold (\( y \)). Data was collected for days between 65°F and 95°F. The best-fit line is:

Slope: \( m = 1.2 \)
For every 1°F increase in temperature, the stand sells about 1.2 more cups.
Y-int: \( b = -50 \)
At 0°F, the model predicts \( -50 \) cups: impossible! The y-intercept has no practical meaning here because 0°F is far outside the data range.

Example 1: Making Predictions

Using \( y = 1.2x - 50 \) with data from 65°F to 95°F:

✓ Interpolation
At \( x = 80 \)°F: \( y = 1.2(80) - 50 = 46 \) cups. This is within the data range, so this prediction is reasonable.
✕ Extrapolation
At \( x = 120 \)°F: \( y = 1.2(120) - 50 = 94 \) cups. This is far outside the data: temperatures that high were never observed, so we cannot trust this.

Example 2: Plant Growth

A scientist measures a plant's height (\( y \), cm) over weeks (\( x \)) from week 1 to week 8. The model is:

Weeks (x) Height in cm (y) 0 1 2 3 4 5 6 7 8 0 10 20 30 \( b = 2 \)

Slope: The plant grows 3.5 cm per week. Y-intercept: At week 0, predicted height is 2 cm (the seedling's starting height): this makes sense!

Example 3: Negative Slope

A car starts with 15 gallons of gas. After driving \( x \) miles (data from 0 to 300 miles), the remaining gas (\( y \)) follows:

m
Slope = \( -0.04 \): For every 1 mile driven, the car uses 0.04 gallons of gas. (Or: every 25 miles uses 1 gallon.)
b
Y-intercept = 15: At 0 miles driven, the tank has 15 gallons: the full tank.

Common Mistakes

✕ Confusing Slope and Intercept
Saying the slope is the "starting value" or the intercept is the "rate." Remember: \( m \) = rate, \( b \) = start.
✕ Ignoring Context
Writing "\( m = 3 \)" without units. Always say: "3 what per what?" (e.g., 3 cm per week).
✕ Trusting Extrapolation Blindly
Using the model far outside the data range and treating the answer as reliable. Always check where the data stops.

Summary Table

Concept What It Means Watch Out For
Slope \( m \) Rate of change of \( y \) per unit \( x \) Always include units
Y-intercept \( b \) Value of \( y \) when \( x = 0 \) May not have real-world meaning
Interpolation Predictions within data range Still an estimate, not exact
Extrapolation Predictions outside data range Can produce impossible values
Domain limits Realistic bounds on \( x \) Negative time, negative quantities

Question 1 of 5

A model for a gym membership is \( y = 25x + 50 \), where \( x \) is the number of months and \( y \) is total cost in dollars. What does the slope mean?

Question 2 of 5

In the model \( y = -2x + 100 \), where \( x \) is hours spent watching TV and \( y \) is a quiz score, what does the y-intercept represent?

Question 3 of 5

A model \( y = 4x + 10 \) was built from data where \( x \) ranged from 2 to 12. A student uses it to predict \( y \) when \( x = 50 \). This is an example of:

Question 4 of 5

A candle's height is modeled by \( y = -0.5x + 12 \), where \( x \) is hours burning and \( y \) is height in inches. Data was collected from \( x = 0 \) to \( x = 20 \). What prediction is unreasonable?

Question 5 of 5

A model \( y = 0.6x + 20 \) describes daily ice cream sales (\( y \), in scoops) based on temperature (\( x \), in °F). Data was from 50°F to 100°F. The y-intercept \( b = 20 \) means: