Eating Out vs.

Eating at Home

While eating out has many fun perks, there are also the associated costs to consider. Is it true
that eating out is always more expensive than cooking the same meal at home? If so, has this
always been the case? Today we will look at the ratio of restaurant to grocery store prices for
the same meals over time.

1. Look at the data showing the restaurant to grocery retail price ratio in the years since 1960. What
do you notice? What do you wonder? I notice that the ratio is always
than 1 the ratio tripled
greaterwonder . I notice
meals
has almost since
what kinds of
1900 . I are
being compared
wonder if fast food is included as a .
restaurant
2. Go to Stapplet.com and click on the “Two Quantitative Variables” option under Data Analysis.
a. Which variable is the Explanatory Variable?
years since 1960
b. Which variable is the Response Variable?
Restaurant to Grocery Retail Price Ratio
c. Copy and paste the data from Column B and Column C in the appropriate spots. (You can
paste a whole column of data at once, entries will be automatically separated by spaces.)

3. Look at the scatterplot. Describe how the restaurant to

grocery retail price ratio is changing over time. Then

............
sketch the scatterplot below.

The ratio has increased over time

in recent years.
and more rapidly

4. Suppose we wanted to find an equation that would help us model the price ratio based on the
number of years since 1960. A least squares regression line fits a line to the data. Click
“Calculate least-squares regression line.”
a. Write the equation of the line below. # of years since 1960

predicted >
- Y =
1 0735
. + 0 .
0414x/
rest : grocery
value
b. What does this model assume about how the price ratio is changing each year?
It assumes the ratio increases by the same amount
each year
.

c. Do you think this is a good model? Why or why not?

No , the pattern does not look linear
. Growth
very
is
speeding up
.
5. A residual plot shows the difference between the actual value and the value predicted by the
regression model. Sketch a picture of the residual plot for the least-squares regression line.
over estimates
consistently
...... consistently underestimates
-

Actual- -..... and

predicted .......... for certain time

intervals
.
 6. In which years does the regression model give an underestimate of the true ratio?
For values of X less than 12 and greater than 45

1960-1972 and 2005 -

2014
actual -

predicted > O
7. In which years does the regression model give an overestimate of the true ratio?
actual s predicted
O
1972 -
2005
actual -

predicted <

actual predicted
8. What would you expect the residual plot to look like if the regression model was a good fit for
the data?
Notern The residual plot should look random with about

Presidual
lot
!
half the dots + half
above the dots

9. Perhaps a linear model is not the right fit. Under “Calculate other regression model”, you can
below
.

P
see what a quadratic or exponential model would look like.
a. Describe what a quadratic model indicates about how the restaurant to grocery ratio is
changing over time.
linearly.
The change in the ratio is growing
(The rate of change of the rate of change is constant).
b. Write the equation of the quadratic regression model below.
predicted
value
- Y =
1 .
59693 -
0 .
0178X + 0 . 00110x2
c. Describe what an exponential model indicates about how the restaurant to grocery ratio
is changing over time.
ratic is changing by
a
The rest. to grocery .
constant from one
year to the next
proportion
d. Write the equation of the exponential regression model below.
X

Y = 1 .

2854x1 .
0182

e. Looking at both residual plots, which function do you think is the better fit? Explain.
the residual
The quadratic model because
is more random and not
plot u-shaped like
the
exponential .
one
10. Using your chosen model, predict the restaurant to grocery retail price ratio for 2023.
↓
1. 59093 -
0 0178
. (b3) +0 .
00110 (63)2 63 after
years
1966

#4143
 Lesson 4.8 – Using Regression Models
QuickNotes
is function that describes the
A regression model
a

relationship between an independent + dependent

, usually
2 more variable data set
a or
variable given
found with technology. least squares regression model
↑
The type of regression model (linear , quadratic exponential
,

is chosen the user

.
by
The model is considered appropriate if the
graph of
the residuals appears without pattern
.

Residual value is the difference between the actual

value + the value predicted the model .
by
Check Your Understanding
Data for a random sample of high school students is shown in the scatterplot.

-
1. Describe the relationship between the
number of absences and the student’s GPA.

Asme otabsences
a
incs
s

2. Does the data seem to suggest a linear,

quadratic, or exponential model? How do you
the
know? Linear because points
fall into a column and there doesn't
seem to be a .
curve

3. A least squares regression model for this data is given by !" = 3.53496 − 0.0146-, where - is a
student’s number of absences and ! is the student’s GPA.
a. Use the residual plot to determine if a linear regression model is appropriate.

Yes ,
there is no
pattern in

residual and
ine plot
The residuals look

randomly distributed
.

b. Interpret the slope of the regression model.

student is absent there is
For each day that a

a
predicted decrease in GPA of 0 .
0146
.
c. Use the regression model to predict the GPA of a student who has 23 absences.
"

Y
=
3 53496
.
- 0 .
0146(23) = 3 .
19914