Confidence Intervals for the Regression Slope 12.1b Target Goal: I can perform a significance test...

Upload
joannarichard 
Category
Documents

view
217 
download
0
Embed Size (px)
Transcript of Confidence Intervals for the Regression Slope 12.1b Target Goal: I can perform a significance test...
Confidence Intervals for Confidence Intervals for the Regression Slopethe Regression Slope
12.1b12.1bTarget Goal: Target Goal: I can perform a I can perform a significance test about the slope β of a significance test about the slope β of a population (true) regression line.population (true) regression line.
h.w:h.w:Pg 761: Pg 761: 13  19 odd 13  19 odd
Review Linear RegressionReview Linear Regression
60 62 64 66 68
Height
Wei
ght
60 62 64 66 68
How much would an adult male weigh if
he were 5 feet 4 inches tall?
Weights of men will vary – in other words,
there is a distribution of weights for adult males who are 5 feet 4 inches
tall.
This distribution is normally
distributed.
60 62 64 66 68
60 62 64 66 68
We want the standard
deviations of all these normal
distributions to be the same.
Let’s look at the heights and weights of a Let’s look at the heights and weights of a population of adult men.population of adult men.
Are some of these weights more likely
than others?What would this distribution look
like?
What would you expect for other heights?
Where would you expect the population
regression line to be?
70 72
The Sampling Distribution of The Sampling Distribution of b : Old Faithful b : Old Faithful cont.cont.
Infe
ren
ce fo
r Linear R
eg
ressio
nIn
fere
nce
for Lin
ear R
eg
ressio
n
6.159: 1.30
1 1.083 20 1 b
xs n
Spread
Let’s return to our earlier exploration of Old Faithful eruptions. For all 222 eruptions in a single month, the population regression line for predicting the interval of time until the
next eruption y from the duration of the previous eruption x is µy = 33.97 + 10.36x.
The standard deviation of responses about this line is given by σ = 6.159. If we take all possible SRSs of 20 eruptions from the population, we get the actual sampling distribution of b.
Shape:Center : (b is an unbiased estimator of β)
In practice, we don’t know σ for the population regression line. So we estimate it with the standard deviation of the residuals, s. Then we estimate the spread of the sampling distribution of b with the standard error of the slope:
1b
x
sSE
s n
Approx. Normalµb = β = 10.36
Performing a Significance Test for the Performing a Significance Test for the SlopeSlope
Infe
ren
ce fo
r Linear R
eg
ressio
nIn
fere
nce
for Lin
ear R
eg
ressio
n
Suppose the conditions for inference are met. To test the hypothesis H0 : β = hypothesized value, compute the test statistic
Find the Pvalue by calculating the probability of getting a t statistic this large or larger in the direction specified by the alternative hypothesis Ha. Use the t distribution with df = n  2.
Suppose the conditions for inference are met. To test the hypothesis H0 : β = hypothesized value, compute the test statistic
Find the Pvalue by calculating the probability of getting a t statistic this large or larger in the direction specified by the alternative hypothesis Ha. Use the t distribution with df = n  2.
0
b
bt
SE
Typically = 0
Example: Crying and IQExample: Crying and IQIn
fere
nce
for Lin
ear R
eg
ressio
nIn
fere
nce
for Lin
ear R
eg
ressio
n
Infants who cry easily may be more easily stimulated than others. This may be a sign of higher IQ. Child development researchers explored the relationship between the crying of infants 4 to 10 days old and their later IQ test scores. A snap of a rubber band on the sole of the foot caused the infants to cry. The researchers recorded the crying and measured its intensity by the number of peaks in the most active 20 seconds. They later measured the children’s IQ at age three years using the StanfordBinet IQ test. A scatterplot and Minitab output for the data from a random sample of 38 infants is below.
Do these data provide convincing evidence that there is a positive linear relationship between crying counts and IQ in the population of infants?
Example: Crying and IQExample: Crying and IQIn
fere
nce
for Lin
ear R
eg
ressio
nIn
fere
nce
for Lin
ear R
eg
ressio
n
State: We want to perform a test ofH0 : β = 0 Ha : β > 0
where β is the true slope of the population regression line relating crying count to IQ score. No significance level was given, so we’ll use α = 0.05.
Do these data provide convincing evidence that there is a positive linear relationship between crying counts and IQ in the population of infants?
Example: Crying and IQExample: Crying and IQ Infe
ren
ce fo
r Linear
Infe
ren
ce fo
r Linear
Reg
ressio
nR
eg
ressio
n
Plan: If the conditions are met, we will perform a t test for the slope β.
• Linear: The scatterplot suggests a moderately weak positive linear relationship between crying peaks and IQ.
The residual plot shows a random scatter of points about the residual = 0 line.
• Independent:
Later IQ scores of individual infants should be independent.
Due to sampling without replacement, there have to be at least 10(38) = 380 infants in the population from which these children were selected.
Example: Crying and IQExample: Crying and IQIn
fere
nce
for Lin
ear
Infe
ren
ce fo
r Linear
Reg
ressio
nR
eg
ressio
n
• Normal: The Normal probability plot of the residuals shows a slight curvature, which suggests that the responses may not be Normally distributed about the line at each xvalue. With such a large sample size (n = 38), however, the t procedures are robust against departures from Normality.
• Equal variance: The residual plot shows a fairly equal amount of scatter around the horizontal line at 0 for all xvalues.
• Random: We are told that these 38 infants were randomly selected.
Example: Crying and IQExample: Crying and IQIn
fere
nce
for Lin
ear R
eg
ressio
nIn
fere
nce
for Lin
ear R
eg
ressio
n
Do: With no obvious violations of the conditions, we proceed to inference. The test statistic and Pvalue can be found in the Minitab output.
Conclude: The Pvalue, 0.002, is less than our α = 0.05 significance level, so we have enough evidence to reject H0 and conclude that there is a positive linear relationship between intensity of crying and IQ score in the population of infants.
0 1.4929 0
0.3.
487007
b
bt
SE
The Minitab output gives P = 0.004 as the Pvalue for a twosided test. The Pvalue for the onesided test is half of this,P = 0.002.
Testing the Hypothesis of No Testing the Hypothesis of No Linear RelationshipLinear Relationship
To test whether or not there is a To test whether or not there is a correlation correlation between between two two quantitative variablesquantitative variables, consider the , consider the slope of the regression line.slope of the regression line.
If there is If there is no correlationno correlation, the , the slope slope would be would be zerozero. .
HH00: : ββ = 0 = 0(The mean of y does not change at all as x (The mean of y does not change at all as x
changes.)changes.)
To test this hypothesis, compute To test this hypothesis, compute the the tt statistic and Pvalue. statistic and Pvalue.
Note:Note: Regression Regression outputoutput from statistical from statistical
software usually software usually gives gives tt and its and its twotwosidedsided Pvalue. Pvalue.
For a onesided testFor a onesided test,, divide the Pdivide the Pvalue in the output by 2.value in the output by 2.
b
bt
SE
Ex: Beer and Blood AlcoholEx: Beer and Blood Alcohol
How well does the number of beers a How well does the number of beers a student drinks predict his or her blood student drinks predict his or her blood alcohol content? alcohol content?
Sixteen Sixteen of age of age college student college student volunteers drank a randomly assigned volunteers drank a randomly assigned number of cans of beer. number of cans of beer. Thirty minutes Thirty minutes later, a police officer measured their later, a police officer measured their blood alcohol content (BAC).blood alcohol content (BAC).
Data revealed:Data revealed:
They noticed a variation in the data They noticed a variation in the data and didn’t believe that the number of and didn’t believe that the number of drinks predicted the BAC well. Here is drinks predicted the BAC well. Here is the scatter plot.the scatter plot.
The solid line is the LSRL.The solid line is the LSRL. The scatterplot shows a The scatterplot shows a clear linear clear linear
relationship.relationship.
Minitab output: Minitab output:
Because r2 = 0.8000, the number of drinks accounts for 80% of the observed variation in BAC.(The students are wrong).
To test hypothesis that the To test hypothesis that the number of beers has no effect on number of beers has no effect on
BAC;BAC;HHoo: : ββ = 0 = 0 There is no correlation between There is no correlation between the number of beers consumed and BAC.the number of beers consumed and BAC.
HHaa: : ββ > 0 > 0 There is a positive correlation There is a positive correlation between the number of beers consumed between the number of beers consumed greater the BAC.greater the BAC.
Examine the PvalueExamine the Pvalue From the output: From the output: Pvalue = 0.0000Pvalue = 0.0000, , The The one sided Pvalueone sided Pvalue is is half half of this so of this so
it is also close to 0.it is also close to 0.
Thus, Thus, we reject Hwe reject Hoo and conclude and conclude that the number of beers that the number of beers does have does have an effect on BAC.an effect on BAC.
CalculatorCalculator (page 804): (page 804): LinRegTestLinRegTest
(Do for “crying vs. IQ.)(Do for “crying vs. IQ.)