Scribd
Upload
92 views4 pages

HW3 Solutions - Stats 500: Problem 1

This document contains solutions to statistical modeling problems. For problem 1, the variables sex and income are found to be significant predictors of gambling in a linear regression model. For problem 2, the variable "takers" is found to improve model fit when added. For problem 3, confidence intervals are calculated for predicted gambling expenditures under various scenarios, demonstrating how intervals widen with extrapolation beyond available data.

Uploaded by

Souleymane Coulibaly
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
92 views4 pages

HW3 Solutions - Stats 500: Problem 1

This document contains solutions to statistical modeling problems. For problem 1, the variables sex and income are found to be significant predictors of gambling in a linear regression model. For problem 2, the variable "takers" is found to improve model fit when added. For problem 3, confidence intervals are calculated for predicted gambling expenditures under various scenarios, demonstrating how intervals widen with extrapolation beyond available data.

Uploaded by

Souleymane Coulibaly
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

HW3 Solutions - Stats 500

Problem 1.

library(faraway)
data(teengamb)
gambmod = lm(gamble ~ ., data = teengamb)
RSS = sum(residuals(gambmod)**2)
DF = df.residual(gambmod)
summary(gambmod)

##
## Call:
## lm(formula = gamble ~ ., data = teengamb)
##
## Residuals:
## Min 1Q Median 3Q Max
## -51.082 -11.320 -1.451 9.452 94.252
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 22.55565 17.19680 1.312 0.1968
## sex -22.11833 8.21111 -2.694 0.0101 *
## status 0.05223 0.28111 0.186 0.8535
## income 4.96198 1.02539 4.839 1.79e-05 ***
## verbal -2.95949 2.17215 -1.362 0.1803
## ---
## Signif. codes: 0 *** 0.001 ** 0.01 * 0.05 . 0.1 1
##
## Residual standard error: 22.69 on 42 degrees of freedom
## Multiple R-squared: 0.5267, Adjusted R-squared: 0.4816
## F-statistic: 11.69 on 4 and 42 DF, p-value: 1.815e-06

(a).

The variables sex and income are significant, with p-values equal to 0.0101 and 1.79e-05, respectively.

(b).

Assuming all other variables are held constant, a female teenager is expected to spend $22.12 less on gambling
than a male teenager.

(c).

gambmod2 = lm(gamble ~ income, teengamb)


rss = sum(residuals(gambmod2)**2)
df = df.residual(gambmod2)
F = ((rss - RSS) /(df - DF)) / (RSS / DF)

1
pvalue = pf(F, df - DF, DF, lower.tail = FALSE)
pvalue

## [1] 0.01177211
We conclude that at the 5% level, including the other predicotrs significantly improves the fit of the model.

Problem 2.

(a).

data(sat)
satmod = lm(total ~ expend + ratio + salary, data = sat)
summary(satmod)

##
## Call:
## lm(formula = total ~ expend + ratio + salary, data = sat)
##
## Residuals:
## Min 1Q Median 3Q Max
## -140.911 -46.740 -7.535 47.966 123.329
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 1069.234 110.925 9.639 1.29e-12 ***
## expend 16.469 22.050 0.747 0.4589
## ratio 6.330 6.542 0.968 0.3383
## salary -8.823 4.697 -1.878 0.0667 .
## ---
## Signif. codes: 0 *** 0.001 ** 0.01 * 0.05 . 0.1 1
##
## Residual standard error: 68.65 on 46 degrees of freedom
## Multiple R-squared: 0.2096, Adjusted R-squared: 0.1581
## F-statistic: 4.066 on 3 and 46 DF, p-value: 0.01209
Based on the summary, the p-value for H0 : —salary = 0 is 0.0667, so at the 5% level we fail to reject —salary = 0.
On the other hand, the p-value for the F-test with null hypothesis: {—expend = —ratio = —salary = 0} is
0.01209 so we can say that at least one of these variables is significant.

(b).

satmod2 = lm(total ~ expend + ratio + salary + takers, data = sat)


summary(satmod2)$coef[ takers ,]

## Estimate Std. Error t value Pr(>|t|)


## -2.904481e+00 2.312600e-01 -1.255937e+01 2.606559e-16
The t-test suggests that ‘takers’ is significant. (p-value = 2.606559e ≠ 16 ¥ 0.) Now we compute the
F-statistic:

2
RSS = sum(satmod2$residuals**2); DF = df.residual(satmod2)
rss = sum(satmod$residuals**2); df = df.residual(satmod)
F = ((rss - RSS) /(df - DF)) / (RSS / DF)
pvalue = pf(F, df - DF, DF, lower.tail = FALSE)
pvalue

## [1] 2.606559e-16
As expected, the F-test yields the same result.

Problem 3:

(a).

x0 = data.frame(t(c(0, apply(teengamb[2:4], 2, mean))))


colnames(x0)[1] <- "sex"
predict(gambmod, new = x0, interval = prediction )

## fit lwr upr


## 1 28.24252 -18.51536 75.00039

(b).

x1 = data.frame(t(c(0, apply(teengamb[2:4], 2, max))))


colnames(x1)[1] <- "sex"
predict(gambmod, new = x1, interval = prediction )

## fit lwr upr


## 1 71.30794 17.06588 125.55
This confidence interval is wider because of extrapolation.

(c).

transformed.model = lm(sqrt(gamble) ~ ., data = teengamb)


transformed.prediction = predict(transformed.model, new = x0, interval = prediction )
transformed.prediction

## fit lwr upr


## 1 4.049523 -0.245035 8.344082
Ô 
Note that we always have gamble Ø 0, so the negarive part of the interval is irrelevant, i.e. YX0 œ

[0, 8.344082] with 95% confidnence. Now if we are 95% confident that YX0 œ [0, 8.344082], then we’re also
equally confident that YX0 œ [0, 8.34412 ] = [0, 69.624]. Hence a 95% (prediction) C.I for for an average male’s
gambling expenditure is [0, 69.624].

(d).

3
x2 = x0
x2[1,] <- c(1, 20, 1, 10)
predict(transformed.model, new = x2, interval = prediction )

## fit lwr upr


## 1 -2.08648 -6.908863 2.735903

A negative fit for YX2 is meaningless, and as before, the negative part of the interval is not informative. So

we choose 0 as our fit (point estimator) and [0, 2.736] as the 95% C.I. for YX2 . Equivalently, YX2 œ [0, 7.49]
with 95% confidence.
We now consider why the predicted value is negative. According to the scatter plot, we can see that female
verbal is between 4 and 8. It cannot cover 10 which is a new observation value. Therefore, the fitted model
may not be valid for the new observation, or it generates a huge error in terms of prediction. Note also that
the lowest income for a female is 1.5.
Income = teengamb$income[which(teengamb$sex == 1)]
Verbal = teengamb$verbal[which(teengamb$sex == 1)]
plot(Income, Verbal, main = Verbal vs. Income for Females )

Verbal vs. Income for Females


8
7
Verbal

6
5
4

2 4 6 8 10

Income

You might also like