10 Regression and Correlation

 

The previous chapter looked at comparing populations to see if there is a dif- ference between the two. That involved two random variables that are similar measures. This chapter will look at two random variables that do not need to be similar measures, and see if there is a relationship between the two or more variables. To do this, you look at regression, which finds the linear relationship, and correlation, which measures the strength of a linear relationship.

Please note: there are many other types of relationships besides linear that can be found for the data. This book will only explore linear, but realize that there are other relationships that can be used to describe data.

 

Regression

When comparing different variables, two questions come to mind: “Is there a relationship between two variables?” and “How strong is that relationship?” These questions can be answered using regression and correlation. Regres- sion answers whether there is a relationship (again this book will explore linear only) and correlation answers how strong the linear relationship is. The vari- able that are used to explain the change in the other variable is called the ex- panatory variable while the variable that is changing is called the response variable. Other variables that help to explain the changes are known as co- variates. To introduce the concepts of regression and correlation it is easier to look at a set of data.

 

Example: Determining If There Is a Relationship

Is there a relationship between the alcohol content and the number of calories in 12-ounce beer? To determine if there is one a sample of beer’s alcohol content and calories (Find Out How Many Calories in Beer?, 2019), is in table #10.1.1.

 

333

 

 

Table #10.1.1: Alcohol and Calorie Content in Beer

Beer <- read.csv( “https://krkozak.github.io/MAT160/beer_data.csv”)head(Beer)

##beer brewery location ## 1 American Amber Lager Straub Brewery domestic ## 2American Lager Straub Brewery domestic

## 3American Light Straub Brewery domestic

## 4Anchor Porter Anchor domestic

## 5Anchor Steam Anchor domestic ## 6 Anheuser Busch Natural Light Anheuser Busch domestic ## alcohol calories carbs

##	1	0.04	136 10.5
##	2	0.04	132 10.5
##	3	0.03	967.6
##	4	0.06	209NA
##	5	0.05	153 16.0
##	6	0.04	953.2

Code book for data frame Beer

Description Collection of the most popular beers from large breweries. The data is of the calories, carbs and alcohol of a specific beer. The data is shown for a 12 ounce serving. The collection includes both domestic and import beer. For the imported beers the information is per 12 oz. serving even though many imports come in pints.

This data frame contains the following columns:

beer: The name of the beer.

brewery: the brewery that brews the beer.

location: whether the beer is brewed in the U.S. (domestic) or brewed in another country (import).

alcohol: the alcohol content of the beer. calories: the number of calories in the beer.

carbs: the amount of carbohydrates in the beer (g).

Source Find Out How Many Calories in Beer? (n.d.). Retrieved July 21, 2019, from https://www.beer100.com/beer-calories/

References (Find Out How Many Calories in Beer?, 2019)

Solution:

To aid in figuring out if there is a relationship, it helps to draw a scatter plot of the data. First it is helpful to state the random variables, and since in an

 

 

algebra class the variables are represented as x and y, those labels will be used here. It helps to state which variable is x and which is y.

State random variables

x = alcohol content in the beer

y = calories in 12 ounce beer

gf_point(calories~alcohol, data=Beer)

 

 

 

 

 

 

300

 

 

 

 

 

 

 

calories
200

 

 

 

 

 

 

 

100

0.0000.0250.0500.0750.1000.12

alcohol

Figure 10.1: Scatter Plot of Calories in Beer versus Alcohol Content This scatter plot looks fairly linear.

gf_point(alcohol~calories, data=Beer)%>% gf_lm()
To find the equation for the linear relationship, the process of regression is used to find the line that best fits the data (sometimes called the best fitting line). The process is to draw the line through the data and then find the vertical distance from a point to the line. These are called the residuals. The regression line is the line that makes the square of the residuals as small as possible, so the regression line is also sometimes called the least squares line. The regression line on the scatter plot is displayed in (Figure ??).

The find the regression equation (also known as best fitting line or least squares line)

 

 

 

0.125

 

 

 

 

0.100

 

 

 

 

0.075

 

 

 

 

alcohol
0.050

 

 

 

 

0.025

 

 

 

 

0.000

100200300

calories

 

Figure 10.2: Scatter Plot of Beer Data with Regression Line

 

Given a collection of paired sample data, the regression equation is 𝑦̂ = 𝑚𝑥 + 𝑏

where the slope = 𝑚 and y-intercept = (0, 𝑏)

To find the linear model using R Studio, use the command:

lm(response variable~explanatory variable, data=Data Frame)

The residuals are the difference between the actual values and the estimated values.

The independent variable, also called the explanatory variable or pre- dictor variable, is the x-value in the equation. The independent variable is the one that you use to predict what the other variable is. The dependent variable depends on what independent value you pick. It also responds to the explanatory variable and is usually called the response variable. In the alco- hol content and calorie example, it makes slightly more sense to say that you would use the alcohol content on a beer to predict the number of calories in the beer.

Assumptions of the regression line when creating a regression line from a sample:

• The set of ordered pairs is a random sample from the population of all such possible pairs.
• For each fixed value of x, the y-values have a normal distribution. All of

 

the y distributions have the same variance, and for a given x-value, the distribution of y-values has a mean that lies on the least squares line. You also assume that for a fixed y, each x has its own normal distribution. This is difficult to figure out, so you can use the following to determine if you have a normal distribution.

• Look to see if the scatter plot has a linear pattern.
• Examine the residuals to see if there is randomness in the residuals. If there is a pattern to the residuals, then there is an issue in the data.

Example #10.1.2: Find the Equation of the Regression Line

• Is there a relationship between the alcohol content and the number of calories in 12-ounce beer? To determine if there is one a sample of beer’s alcohol content and calories (Find Out How Many Calories in Beer?, 2019), is in table #10.1.1.

Solution:

State random variables

x = alcohol content in the beer

y = calories in 12 ounce beer

• Find the regression equation: To find the regression equation using R Studio, the command is

lm(response variable ~ explanatory variable, data=Data Frame)

For this example, the command would be

lm(calories~alcohol, data=Beer)

##

## Call:

## lm(formula = calories ~ alcohol, data = Beer) ##

## Coefficients:

## (Intercept)alcohol

##14.532672.36

From this you can see that the y-intercept is 14.53 and the slope is 2672.36. So the regression equation is 𝑦̂ = 2672𝑥 + 14.3.

Remember, this is an estimate for the true regression. A different sample would produce a different estimate.

Assumptions check:

• A random sample of alcohol content and calories was taken. Check: There is no guarantee that this was a random sample. The data was collected off of a website, and the website does not say how the data was obtained.

 

However, it is a collection of most popular beers from large breweries, so it may be alright that it isn’t a random sample.

• The distribution for each calorie value is normally distributed for every value of alcohol content in the beer. Check:
• From Example #10.2.1, the scatter plot looks fairly linear.
• 𝑦̂ − 𝑦
  To graph the residuals, first the residuals need to be calculated. This means that for every x value, you need to calculate the y value that can be found from the regression equation. Then subtract the y value that was

measured from this calculated y value. This is. will calculate these for you. The command is:

Luckily, R Studio

 

Table #10.1.2: Residuals for Alcohol and Calorie Content in Beer

 

lm.out<-lm(calories~alcohol, data=Beer) residuals(lm.out)
##	1	2	3	4	5
##	14.5785382	10.5785382	1.3021259	34.1313626	4.8549504
##	6	7	8	9	10
## -26.4214618 -17.8686374 -27.4214618				6.8549504	20.1313626
##	11	12	13	14	15
##	28.8549504	20.1313626	14.8549504	5.8549504	14.8549504
##	16	17	18	19	20
##	4.8549504	22.8549504	21.1313626	18.1313626 -18.1450496
##	21	22	23	24	25
## -51.8686374 -33.1450496 -11.4214618				-5.4214618 -37.8686374
##	26	27	28	29	30
##	-3.1450496	-28.8686374	-22.4214618	-12.9742863	-15.1450496
##	31	32	33	34	35
##	-5.8686374	-26.4214618	16.5785382	-0.8686374	0.8549504
##	36	37	38	39	40
##	-17.4214618	13.8549504	5.1313626	-11.5922251	3.8549504
##	41	42	43	44	45
##	-6.3158129	9.8549504	-9.8686374	-22.2101640	-24.8686374
##	46	47	48	49	50
##	15.9605993	32.2370115	-10.1450496	-4.5922251	9.6841871
##	51	52	53	54	55
##	-20.8686374	-7.3158129	-11.8686374	-13.5922251	-19.1450496
##	56	57	58	59	60
##	6.2370115	-17.1450496	-0.1450496	13.8549504	-18.8686374
##	61	62	63	64	65
##	-0.1450496	13.8549504	-42.8686374	-25.8686374	-18.4214618
##	66	67	68	69	70
##	-4.1450496	-4.1450496	-11.4214618	-10.4214618	-18.4214618
##	71	72	73	74	75

 

##-32.86863740.854950414.85495043.854950421.8549504##7677787980##17.854950416.8549504-16.4214618-11.421461811.8549504##8182838485##11.8549504-17.8686374-3.145049617.85495046.8549504##8687888990##3.8549504-11.421461826.8549504-8.4214618-26.4214618##9192939495##-17.8686374-17.8686374-11.4214618-5.1450496-11.4214618##96979899100##-30.6978741-5.1450496-11.4214618-25.42146186.5785382##101102103104105##-11.4214618-11.4214618-11.42146186.5785382-23.4214618##106107108109110##-30.868637430.1313626-1.5922251-1.5922251-8.1450496

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

##	111	112113114	115
##	11.8549504	6.8549504 -13.3158129 -3.1450496	55.4728892
##	116	117118119	120
##	-7.4214618	-2.1450496 -14.8686374 -26.3158129	-8.3158129
##	121	122123124	125
##	-2.1450496	-4.1450496 -27.6978741 25.8549504	-1.1450496
##	126	127128129	130
##	4.1313626 -13.5922251 -6.1450496 -28.1450496		11.8549504
##	131	132133134	135
##	11.8549504	17.8549504 46.85495040.1313626	-2.4214618
##	136	137138139	140
##	-3.1450496	-6.1450496 -2.1450496 -11.4214618	10.1313626
##	141	142143144	145
##	15.1313626	48.2370115 12.40777498.8549504	15.1313626
##	146	147148149	150
##	13.4077749	29.40777490.1313626 16.4077749	19.1313626
##	151	152153154	155
##	50.1313626	9.8549504 31.57853824.8549504 -25.1450496
##	156	157158159	160
##	-6.3158129	-6.3158129 -14.86863740.8549504	-8.4214618
##	161	162163164	165
##	20.1313626	-1.1450496 14.85495047.8549504	1.8549504
##	166	167168169	170
##	13.5785382	2.8549504 12.8549504 -1.1450496	-3.1450496
##	171	172173174	175
##	1.8549504	13.5785382 -23.4214618 13.5785382 -22.4214618
##	176	177178179	180
##	-2.1450496 -14.8686374 -2.1450496 -5.1450496		-3.9742863
##	181	182183184	185
##	-2.1450496	6.5785382 24.57853827.8549504	-0.1450496

##186187188189190

 

 

##	0.8549504	7.8549504	18.4077749	7.8549504	16.5785382
##	191	192	193	194	195
##	11.8549504	-6.1450496	-3.1450496	31.5785382	-6.1450496
##	196	197	198	199	200
##	1.8549504	21.8549504	-22.4214618	-13.5922251	4.8549504
##	201	202	203	204	205
##	14.8549504	4.8549504	13.2370115	4.8549504	-4.8686374
##	206	207	208	209	210
##	14.0257137	4.8549504	-38.1450496	-6.1450496	38.5785382
##	211	212	213	214	215
##	4.8549504	9.8549504	25.8549504	9.5785382	-16.5922251
##	216	217	218	219	220
##	-0.1450496	1.8549504	-10.1450496	31.5785382	6.5785382
##	221	222	223	224	225
##	-41.5922251	3.4077749	-2.1450496	4.8549504	20.5785382
##	226	227
##	11.8549504	4.8549504

lm.out saves the linear model into a variable called lm.out

resduals(lm.out) finds the residuals for each x value based on the linear model, lm.out, and displays them.

Now graph the residuals to see if there is any pattern to the residuals. The command in R Studio is:

gf_point(residuals(lm.out)~alcohol, data=Beer)

The residual versus the x-values plot looks fairly random.

It appears that the distribution for calories is a normal distribution.

b. Use the regression equation to find the number of calories when the alcohol content is 7.00%.

Solution:

𝑦̂ = 2672 ∗ 0.07 + 14.3 = 201.34
If you are drinking a beer that is 7.00% alcohol content, then it is probably close to 200 calories.

The mean number of calories is 154.5 calories. This value of 201 seems like a better estimate than the mean when looking at the data since all the beers with 7% alcohol have between 160 and 231 calories. The regression equation is a better estimate than just the mean.

c. Use the regression equation to find the number of calories when the alcohol content is 14%.

Solution: 𝑦̂ = 2672 ∗ 0.14 + 14.3 = 388.38

If you are drinking a beer that is 14% alcohol content, then it has probably close to 389 calories. Since 12% alcohol beer has 330 calories you might think

 

 

 

60

 

 

 

 

 

30

 

 

 

 

 

residuals(lm.out)
0

 

 

 

 

 

−30

0.0000.0250.0500.0750.1000.12

alcohol

 

Figure 10.3: Residual Plot of Beer Data

 

that 14% would have more calories than this. This estimate is what is called extrapolation. It is not a good idea to predict values that are far outside the range of the original data. This is because you can never be sure that the regression equation is valid for data outside the original data.

Notice, that the 7.00% value falls into the range of the original x-values. The processes of predicting values using an x within the range of original x-values is called interpolating. The 14.00% value is outside the range of original x-values. Using an x-value that is outside the range of the original x-values is called extrapolating. When predicting values using interpolation, you can usually feel pretty confident that that value will be close to the true value. When you extrapolate, you are not really sure that the predicted value is close to the true value. This is because when you interpolate, you know the equation that predicts, but when you extrapolate, you are not really sure that your relationship is still valid. The relationship could in fact change for different x-values.

An example of this is when you use regression to come up with an equation to predict the growth of a city, like Flagstaff, AZ. Based on analysis it was determined that the population of Flagstaff would be well over 50,000 by 1995. However, when a census was undertaken in 1995, the population was less than 50,000. This is because they extrapolated and the growth factor they were using had obviously changed from the early 1990’s. Growth factors can change for many reasons, such as employment growth, employment stagnation, disease,

 

 

articles saying great place to live, etc. Realize that when you extrapolate, your predicted value may not be anywhere close to the actual value that you observe.

What does the slope mean in the context of this problem?

The calories increase 26.72 calories for every 1% increase in alcohol content.

The y-intercept in many cases is meaningless. In this case, it means that if a drink has 0 alcohol content, then it would have 14.3 calories. This may be reasonable, but remember this value is an extrapolation so it may be wrong.

Consider the residuals again. According to the data, a beer with 7.0% alcohol has between 160 and 231 calories. The predicted value is 201 calories. This vari- ation means that the actual value was between 40 calories below and 30 calories above the predicted value. That isn’t that far off. Some of the actual values differ by a large amount from the predicted value. This is due to variability in the response variable. The larger the residuals the less the model explains the variability in the response variable. There needs to be a way to calculate how well the model explains the variability in the response variable. This will be explored in the next section.

On last thing to look at here, is that you may wonder if import beer has a different amount of calories than domestic beer. The location is a co-variate, a third variable that affects the calories, and you can graph a scatter plot that separates based on the co-variate. Here is how to do this in R Studio.

gf_point(calories~alcohol, data=Beer, color=~location)%>% gf_lm()

Looking at the scatter plot, there doesn’t appear to be an affect from domestic or import. This is what is nice about scatter plots, You can visually see a possible relationships.

 

Homework

** For each problem, state the random variables. The Data Frame in this section are used in the homework for sections 10.2 and 10.3 also. **

• When an anthropologist finds skeletal remains, they need to figure out the height of the person. The height of a person (in cm) and the length of their metacarpal bone 1 (in cm) were collected and are in table #10.1.3 (”Prediction of height,” 2013). Create a scatter plot and find a regression equation between the height of a person and the length of their metacarpal. Then use the regression equation to find the height of a person for a metacarpal length of 44 cm and for a metacarpal length of 55 cm. Which height that you calculated do you think is closer to the true height of the person? Why?

Table #10.1.3: Data Frame of Metacarpal versus Height

 

 

 

 

 

 

 

 

300

 

 

 

 

calories
200location

domestic import

 

 

100

 

 

 

 

 

0

0.0000.0250.0500.0750.1000.125

alcohol

Figure 10.4: Calories versus Alcohol Content Separated by Location of Beer Data

Metacarpal<- read.csv( “https://krkozak.github.io/MAT160/metacarpal.csv”)head(Metacarpal)

##length height

## 1	45	171
## 2	51	178
## 3	39	157
## 4	41	163
## 5	48	172
## 6	49	183

Code book for Data Frame Metacarpal see problem 2.3.1 in Section 2.3 Homework

• Table #10.1.4 contains the value of the house and the amount of rental income in a year that the house brings in (”Capital and rental,” 2013). Create a scatter plot and find a regression equation between house value and rental income. Then use the regression equation to find the rental income a house worth $230,000 and for a house worth $400,000. Which rental income that you calculated do you think is closer to the true rental income? Why?

 

Table #10.1.4: Data Frame of House Value versus Rental

House<- read.csv( “https://krkozak.github.io/MAT160/house.csv”)head(House)

 

##		capital	rental
##	1	61500	6656
##	2	67500	6864
##	3	75000	4992
##	4	75000	7280
##	5	76000	6656
##	6	77000	4576

Code book for Data Frame House see problem 2.3.2 in Section 2.3 Home- work

• The World Bank collects information on the life expectancy of a person in each country (”Life expectancy at,” 2013) and the fertility rate per woman in the country (”Fertility rate,” 2013). The Data Frame for countries for the year 2011 are in table #10.1.5. Create a scatter plot of the Data Frame and find a linear regression equation between fertility rate and life expectancy in 2011. Then use the regression equation to find the life expectancy for a country that has a fertility rate of 2.7 and for a country with fertility rate of 8.1. Which life expectancy that you calculated do you think is closer to the true life expectancy? Why?

Table #10.1.5: Data Frame of Life Expectancy versus Fertility Rate

Fertility<- read.csv( “https://krkozak.github.io/MAT160/fertility.csv”)head(Fertility)

##country lifexp_2011 fertilrate_2011

##	1	Macao SAR, China	79.91	1.03
##	2	Hong Kong SAR, China	83.42	1.20
##	3	Singapore	81.89	1.20
##	4	Hungary	74.86	1.23
##	5	Korea, Rep.	80.87	1.24
##	6	Romania	74.51	1.25

##lifexp_2000 fertilrate_2000 lifexp_1990 fertilrate_1990

## 1	77.62	0.94	75.28	1.69
## 2	80.88	1.04	77.38	1.27
## 3	78.05	NA	76.03	1.87
## 4	71.25	1.32	69.32	1.84
## 5	75.86	1.47	71.29	1.59
## 6	71.16	1.31	69.74	1.84

Code book for Data Frame Fertility see problem 3.2.3 in Section 3.2 Home-

 

 

work

• The World Bank collected data on the percentage of gross domestic prod- uct (GDP) that a country spends on health expenditures (Current health expenditure (% of GDP), 2019), the fertility rate of the country (Fertility rate, total (births per woman), 2019), and the percentage of woman re- ceiving prenatal care (Pregnant women receiving prenatal care (%), 2019). The Data Frame for the countries where this information is available in ta- ble #10.1.6. Create a scatter plot of the Data Frame and find a regression equation between percentage spent on health expenditure and the percent- age of women receiving prenatal care. Then use the regression equation to find the percent of women receiving prenatal care for a country that spends 5.0% of GDP on health expenditure and for a country that spends 12.0% of GDP. Which prenatal care percentage that you calculated do you think is closer to the true percentage? Why?

Table #10.1.6: Data Frame of Prenatal Care versus Health Expendi- ture

Fert_prenatal<- read.csv( “https://krkozak.github.io/MAT160/fertility_prenatal.csv”)head(Fert_prenatal)

 

##1Lowermiddleincome7.4787.5247.5637.5927.6117.619##2Uppermiddleincome4.7864.6704.5214.3454.1503.950##3Uppermiddleincome6.5006.4806.4606.4406.4206.400##4Lowermiddleincome7.6917.7207.7507.7817.8117.841##5Lowincome6.8806.8776.8756.8726.8676.864##6Lowincome6.1146.1276.1386.1476.1546.160
##Country.Name Country.CodeRegion ## 1AngolaAGOSub-Saharan Africa ## 2ArmeniaARMEurope & Central Asia ## 3BelizeBLZ Latin America & Caribbean ## 4 Cote d’IvoireCIVSub-Saharan Africa ## 5EthiopiaETHSub-Saharan Africa ## 6GuineaGINSub-Saharan Africa ##IncomeGroup f1960 f1961 f1962 f1963 f1964 f1965

 

 

 

 

 

 

##	f1966		f1967	f1968	f1969	f1970	f1971	f1972	f1973	f1974
##	1	7.618	7.613	7.608	7.604	7.601	7.603	7.606	7.611	7.614
##	2	3.758	3.582	3.429	3.302	3.199	3.114	3.035	2.956	2.875
##	3	6.379	6.358	6.337	6.316	6.299	6.288	6.284	6.285	6.287
##	4	7.868	7.893	7.912	7.927	7.936	7.941	7.942	7.939	7.929
##	5	6.867	6.880	6.903	6.937	6.978	7.020	7.060	7.094	7.121
##	6	6.168	6.177	6.189	6.205	6.225	6.249	6.277	6.306	6.337
##		f1975	f1976	f1977	f1978	f1979	f1980	f1981	f1982	f1983
##	1	7.615	7.609	7.594	7.571	7.540	7.504	7.469	7.438	7.413
##	2	2.792	2.712	2.641	2.582	2.538	2.510	2.499	2.503	2.517

 

## 1NANANANANANANANANA## 2NANANANANANANANANA## 3NANANA96NANANANANA## 4NANANANANANA83.2NANA## 5NANANANANANANANANA## 6NANANANA57.6NANANANA

##	3	6.278	6.250	6.195	6.109	5.992	5.849	5.684	5.510	5.336
##	4	7.910	7.877	7.828	7.763	7.682	7.590	7.488	7.383	7.278
##	5	7.143	7.167	7.195	7.230	7.271	7.316	7.360	7.397	7.424
##	6	6.369	6.402	6.436	6.468	6.500	6.529	6.557	6.581	6.602
##		f1984	f1985	f1986	f1987	f1988	f1989	f1990	f1991	f1992
##	1	7.394	7.380	7.366	7.349	7.324	7.291	7.247	7.193	7.130
##	2	2.538	2.559	2.578	2.591	2.592	2.578	2.544	2.484	2.400
##	3	5.170	5.019	4.886	4.771	4.671	4.584	4.508	4.436	4.363
##	4	7.176	7.078	6.984	6.892	6.801	6.710	6.622	6.536	6.454
##	5	7.437	7.435	7.418	7.387	7.347	7.298	7.246	7.193	7.143
##	6	6.619	6.631	6.637	6.637	6.631	6.618	6.598	6.570	6.535
##		f1993	f1994	f1995	f1996	f1997	f1998	f1999	f2000	f2001
##	1	7.063	6.992	6.922	6.854	6.791	6.734	6.683	6.639	6.602
##	2	2.297	2.179	2.056	1.938	1.832	1.747	1.685	1.648	1.635
##	3	4.286	4.201	4.109	4.010	3.908	3.805	3.703	3.600	3.496
##	4	6.374	6.298	6.224	6.152	6.079	6.006	5.932	5.859	5.787
##	5	7.094	7.046	6.995	6.935	6.861	6.769	6.659	6.529	6.380
##	6	6.493	6.444	6.391	6.334	6.273	6.211	6.147	6.082	6.015
##		f2002	f2003	f2004	f2005	f2006	f2007	f2008	f2009	f2010
##	1	6.568	6.536	6.502	6.465	6.420	6.368	6.307	6.238	6.162
##	2	1.637	1.648	1.665	1.681	1.694	1.702	1.706	1.703	1.693
##	3	3.390	3.282	3.175	3.072	2.977	2.893	2.821	2.762	2.715
##	4	5.717	5.651	5.589	5.531	5.476	5.423	5.372	5.321	5.269
##	5	6.216	6.044	5.867	5.690	5.519	5.355	5.201	5.057	4.924
##	6	5.947	5.877	5.804	5.729	5.653	5.575	5.496	5.417	5.336
##		f2011	f2012	f2013	f2014	f2015	f2016	f2017	p1986	p1987
##	1	6.082	6.000	5.920	5.841	5.766	5.694	5.623	NA	NA
##	2	1.680	1.664	1.648	1.634	1.622	1.612	1.604	NA	NA
##	3	2.676	2.642	2.610	2.578	2.544	2.510	2.475	NA	NA
##	4	5.216	5.160	5.101	5.039	4.976	4.911	4.846	NA	NA
##	5	4.798	4.677	4.556	4.437	4.317	4.198	4.081	NA	NA
##	6	5.256	5.175	5.094	5.014	4.934	4.855	4.777	NA	NA
##		p1988	p1989	p1990	p1991	p1992	p1993	p1994	p1995	p1996

 

 

 

 

 

 

##	p1997		p1998	p1999	p2000	p2001	p2002	p2003	p2004	p2005
##	1	NA	NA	NA	NA	65.6	NA	NA	NA	NA
##	2	82	NA	NA	92.4	NA	NA	NA	NA	93.0
##	3	NA	98	95.9	100.0	NA	98	NA	NA	94.0
##	4	NA	NA	84.3	87.6	NA	NA	NA	NA	87.3
##	5	NA	NA	NA	26.7	NA	NA	NA	NA	27.6
##	6	NA	NA	70.7	NA	NA	NA	84.3	NA	82.2

 

 

##p2006 p2007 p2008 p2009 p2010 p2011 p2012 p2013 p2014

## 1		NA	79.8		NA	NA	NA	NA	NA	NA	NA
## 2		NA	NA		NA	NA 99.1		NA	NA	NA	NA
## 3		94.0	99.2		NA	NA	NA	96.2	NA	NA	NA
## 4		84.8	NA		NA	NA	NA	NA	90.6	NA	NA
## 5		NA	NA		NA	NA	NA	33.9	NA	NA	41.2
## 6		NA	88.4		NA	NA	NA	NA	85.2	NA	NA
##		p2015 p2016 p2017 p2018					e2000		e2001	e2002
##	1	NA	81.6		NA	NA 2.334435 5.483824 4.072288
##	2	NA	99.6		NA	NA 6.505224 6.536262 5.690812
##	3	97.2	97.2		NA	NA 3.942030 4.228792 3.864327
##	4	NA	93.2		NA	NA 5.672228 4.850694 4.476869
##	5	NA	62.4		NA	NA 4.365290 4.713670 4.705820
##	6	NA	84.3		NA	NA 3.697726 3.884610 4.384152
##	e2003				e2004	e2005	e2006		e2007	e2008
##	1	4.454100		4.757211		3.734836	3.366183		3.211438	3.495036
##	2	5.610725		8.227844		7.034880	5.588461		5.445144	4.346749
##	3	4.260178		4.091610		4.216728	4.163924		4.568384	4.646109
##	4	4.645306		5.213588		5.353556	5.808850		6.259154	6.121604
##	5	4.885341		4.304562		4.100981	4.226696		4.801925	4.280639
##	6	3.651081		3.365547		2.949490	2.960601		3.013074	2.762090
##		e2009		e2010		e2011	e2012		e2013	e2014
##	1	3.578677		2.736684		2.840603	2.692890		2.990929	2.798719
##	2	4.689046		5.264181		3.777260	6.711859		8.269840	10.178299
##	3	5.311070		5.764874		5.575126	5.322589		5.727331	5.652458
##	4	6.223329		6.146566		5.978840	6.019660		5.074942	5.043462
##	5	4.412473		5.466372		4.468978	4.539596		4.075065	4.033651
##	6	2.936868		3.067742		3.789550	3.503983		3.461137	4.780977

##e2015e2016 ## 1 2.950431 2.877825

## 2 10.117628 9.927321

## 3 5.884248 6.121374

## 4 5.262711 4.403621

## 5 3.975932 3.974016

## 6 5.827122 5.478273

Code book for Data Frame Fert_prenatal see problem 2.3.4 in Section

• Homework
• The height and weight of baseball players are in table #10.1.7 (”MLB heightsweights,” 2013). Create a scatter plot and find a regression equa- tion between height and weight of baseball players. Then use the regres- sion equation to find the weight of a baseball player that is 75 inches tall and for a baseball player that is 68 inches tall. Which weight that you calculated do you think is closer to the true weight? Why?

Table #10.1.7: Heights and Weights of Baseball Players

 

 

Baseball <- read.csv( “https://krkozak.github.io/MAT160/baseball.csv”)head(Baseball)

##player height weight

## 1	1	65.78 112.99
## 2	2	71.52 136.49
## 3	3	69.40 153.03
## 4	4	68.22 142.34
## 5	5	67.79 144.30
## 6	6	68.70 123.30

Code book for Data Frame Baseball Description

The heights and weights of MLB players.

Format

This Data Frame contains the following columns:

Player: Player in the sample

height: height of baseball player (inches) weight: weight of baseball player (pounds)

Source MLB heightsweights. (2013, November 16). Retrieved from http:// wiki.stat.ucla.edu/socr/index.php/SOCR_Data_MLB_HeightsWeights

References SOCR Data Frame of MLB Heights Weights from UCLA.

• Different species have different body weights and brain weights are in table #10.1.8. (”Brain2bodyweight,” 2013). Create a scatter plot and find a regression equation between body weights and brain weights. Then use the regression equation to find the brain weight for a species that has a body weight of 62 kg and for a species that has a body weight of 180,000 kg. Which brain weight that you calculated do you think is closer to the true brain weight? Why?

Table #10.1.8: Body Weights and Brain Weights of Species

Body <- read.csv( “https://krkozak.github.io/MAT160/body.csv”)head(Body)

 

##	species	bodyweight	brainweight
## 1	Newborn_Human	3.20	0.374984813
## 2	Adult_Human	73.00	1.349981613
## 3	Pithecanthropus_Man	70.00	0.925010921
## 4	Squirrel	0.80	0.007620352

 

 

##	5	Hamster	0.15 0.001406136
##	6	Chimpanzee	50.00 0.419981176
##brainbodyproportion
##	1	0.117182754
##	2	0.018492899
##	3	0.013214442
##	4	0.009525440
##	5	0.009374242
##	6	0.008399624

Code book for Data Frame Body Description

The body weight, brain weight, and brain/body proportion of different species of animals.

Format

This Data Frame contains the following columns:

species: species of animal

bodyweight: the body weight of the species (kg) brainweight: the brain weight of the species (kg)

brainbodyproportion: the ratio of brain weight to body weight of the species

Source Brain2bodyweight. (2013, November 16). Retrieved from http://wiki. stat.ucla.edu/socr/index.php/SOCR_Data_Brain2BodyWeight

References SOCR Data of species body weights and brain weights from UCLA.

• A sample of hot dogs was taken and the amount of sodium (in mg) and calories were measured. (”Data hotdogs,” 2013) The Data Frame are in table #10.1.9. Create a scatter plot and find a regression equation between amount of calories and amount of sodium. Then use the regression equation to find the amount of sodium a hot dog has if it is 170 calories and if it is 120 calories. Which sodium level that you calculated do you think is closer to the true sodium level? Why?

## 1 Beef186495## 2 Beef181477## 3 Beef176425## 4 Beef149322
Table #10.1.9: Calories and Sodium Levels in Hotdogs

Hotdog<-read.csv( “https://krkozak.github.io/MAT160/hotdog_beef_poultry.csv”)head(Hotdog)

##type calories sodium

 

 

## 5 Beef	184	482
## 6 Beef	190	587

Code book for data frame Hotdog see Example 9.3.2 in Section 9.3

• Per capita income in 1960 dollars for European countries and the percent of the labor force that works in agriculture in 1960 are in table #10.1.10 (”OECD economic development,” 2013). Create a scatter plot and find a regression equation between percent of labor force in agriculture and per capita income. Then use the regression equation to find the per capita income in a country that has 21 percent of labor in agriculture and in a country that has 2 percent of labor in agriculture. Which per capita income that you calculated do you think is closer to the true income? Why?

Table #10.1.10: Percent of Labor in Agriculture and Per Capita In- come for European Countries

Agriculture <- read.csv( “https://krkozak.github.io/MAT160/agriculture.csv”)head(Agriculture)

##country percapita agriculture industry services

##	1	SWEEDEN	1644	14	53	33
##	2	SWITZERLAND	1361	11	56	33
##	3	LUXEMBOURG	1242	15	51	34
##	4	U. KINGDOM	1105	4	56	40
##	5	DENMARK	1049	18	45	37
##	6	W. GERMANY	1035	15	60	25

Code book for Data Frame Agriculture Description

The per capita income and percent in different industries in European countries

Format

This Data Frame contains the following columns:

country: country in Europe percapita: per captia income

agriculture: percentage of workforce in agriculture industry: percentage of workforce in industry services: percentage of workforce in services

Source OECD economic development. (2013, December 04). Retrieved from http://lib.stat.cmu.edu/DASL/Datafiles/oecdat.html

References Data And Story Library

 

 

• Cigarette smoking and cancer have been linked. The number of deaths per one hundred thousand from bladder cancer and the number of cigarettes sold per capita in 1960 are in table #10.1.11 (”Smoking and cancer,” 2013). Create a scatter plot and find a regression equation between number of cigarettes smoked and number of deaths from bladder cancer. Then use the regression equation to find the number of deaths from bladder cancer when the cigarette sales were 20 per capita and when the cigarette sales were 6 per capita. Which number of deaths that you calculated do you think is closer to the true number? Why?

Table #10.1.11: Number of Cigarettes and Number of Bladder Can- cer Deaths

Cancer <- read.csv( “https://krkozak.github.io/MAT160/cancer_1.csv”)head(Cancer)

 

##statecig bladder lung kidney leukemia

## 1	AL 18.20	2.90 17.05	1.59	6.15
## 2	AZ 25.82	3.52 19.80	2.75	6.61
## 3	AR 18.24	2.99 15.98	2.02	6.94
## 4	CA 28.60	4.46 22.07	2.66	7.06
## 5	CT 31.10	5.11 22.83	3.35	7.20
## 6	DE 33.60	4.78 24.55	3.36	6.45

Code book for Data Frame Cancer

Description This data frame contains the number of cigarette sales (per Capita), number of cancer Deaths (per 100 Thousand) from bladder, lung, kidney, and leukemia.

Format

This Data Frame contains the following columns:

state: state in US

cig: the number of cigarette sales (per capita)

bladder: number of deaths per 100 thousand from bladder cancer lung: number of deaths per 100 thousand from lung cancer kidney: number of deaths per 100 thousand from kidney cancer leukemia: number of deaths per 100 thousand from leukemia

Source Smoking and cancer. (2013, December 04). Retrieved from http://lib. stat.cmu.edu/DASL/Datafiles/cigcancerdat.html

References Data And Story Library

 

 

• The weight of a car can influence the mileage that the car can obtain. A random sample of cars’ weights and mileage was collected and are in table #10.1.12 (”us auto mileage,” 2019). Create a scatter plot and find a regression equation between weight of cars and mileage. Then use the regression equation to find the mileage on a car that weighs 3800 pounds and on a car that weighs 2000 pounds. Which mileage that you calculated do you think is closer to the true mileage? Why?

Table #10.1.12: Weights and Mileages of Cars

Cars<- read.csv( “https://krkozak.github.io/MAT160/cars.csv”)head(Cars)

 

##	make vol hp mpg	spwt
## 1	GM/GeoMetroXF1 89 49 65.4	96 17.5
## 2	GM/GeoMetro 92 55 56.0	97 20.0
## 3	GM/GeoMetroLSI 92 55 55.9	97 20.0
## 4	SuzukiSwift 92 70 49.0	105 20.0
## 5	DaihatsuCharade 92 53 46.5	96 20.0
## 6	GM/GeoSprintTurbo 89 70 46.2	105 20.0

Code book for Data Frame Cars

Description Variation in gasoline mileage among makes and models of auto- mobiles is influenced substantially by the weight and horsepower of the vehicles. When miles per gallon and horsepower are transformed to logarithms, the lin- earity of the regression is improved. A negative second order term is required to fit the logarithmic mileage to weight relation. If the logged variables are stan- dardized, the coefficients of the first order terms indicate the standard units change in log mileage per one standard unit change in the predictor variable at the (logarithmic) mean. This change is constant in the case of mileage to horsepower, but not for mileage to weight. The coefficient of the second order weight term indicates the change in standardized slope associated with a one standard deviation increase in the logarithm of weight.

Format

This Data Frame contains the following columns: make: the type of car

vol: cubic feet of cab space hp: engine horsepower

npg: the average mileage of the car sp: top speed (mph)

wt: the weight of the car (100 pounds)

 

 

Source (n.d.). Retrieved July 21, 2019, from https://www3.nd.edu/~busiforc/ handouts/Data and Stories/regression/us auto mileage/usautomileage.html

References R.M. Heavenrich, J.D. Murrell, and K.H. Hellman, Light Duty Automotive Technology and Fuel Economy Trends Through 1991, U.S. Envi- ronmental Protection Agency, 1991 (EPA/AA/CTAB/91-02).

 

 

 

 

Correlation

A correlation exists between two variables when the values of one variable are somehow associated with the values of the other variable. Correlation is a word used in every day life and many people think they understand what it means. What it really means is a measure of how strong the linear relationship is between two variables. However, it doesn’t give you a strong indication of how well the relationship explains the variability in the response variable, and it doesn’t work for more than two variables. There is a way to look at how much of the relationship explains the variability in the response variable.

There are many types of patterns one can see in the data. Common types of patterns are linear, exponential, logarithmic, or periodic. To see this pattern, you can draw a scatter plot of the data.This course will mostly deal with linear, but the other patterns exist.

There is some variability in the response variable values, such as calories. Some of the variation in calories is due to alcohol content and some is due to other factors. How much of the variation in the calories is due to alcohol content?

When considering this question, you want to look at how much of the variation in calories is explained by alcohol content and how much is explained by other variables. Realize that some of the changes in calories have to do with other in- gredients. You can have two beers at the same alcohol content, but one beer has higher calories because of the other ingredients. Some variability is explained by the model and some variability is not explained. Together, both of these give the total variability. This is (total variation)=(explained variation)+(unexplained variation)

𝑅2 =
The proportion of the variation that is explained by the model is

explained variation

total variation

This is known as the coefficient of determination.

To find the coefficient of determination, on R Studio, the command is to first name the linear model to be something and then you can find the summary of that named linear model. The next example shows how to find this.

 

Example: Find the coefficient of determinatin 𝑅2

How well does the alcohol content of a beer explain the variability in the number of calories in 12-ounce beer? To determine this a sample of beer’s alcohol content and calories (Find Out How Many Calories in Beer?, 2019), is in table #10.1.1.

Solution:

State random variables

x = alcohol content in the beer

y = calories in 12 ounce beer

First find and name the linear model, how about lm.out, so you can refer to it. Then use the command

summary(lm.out)

 

For this example this is

lm.out<-lm(calories~alcohol, data=Beer) #creates the linear model

and calls it lm.out summary(lm.out) #creates a summary of the linear model created.

 

##

## Call:

## lm(formula = calories ~ alcohol, data = Beer) ##

## Residuals:

## Min 1Q Median 3Q Max ## -51.869 -11.421 -0.145 11.855 55.473 ##

## Coefficients:

##Estimate Std. Error t value Pr(>|t|) ## (Intercept)14.5274.5773.174 0.00171 **

## alcohol2672.35984.478 31.634 < 2e-16 *** ## —

## Signif. codes:

## 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ‘ 1 ##

## Residual standard error: 17.62 on 225 degrees of freedom ## Multiple R-squared: 0.8164,Adjusted R-squared: 0.8156 ## F-statistic: 1001 on 1 and 225 DF, p-value: < 2.2e-16

0.8164
𝑅𝑅 =
The multiple R-squared value is 2 the coefficient of determination. So 2

Thus, 81.64% of the variation in calories is explained to the linear rela- tionship between alcohol content and calories. The other 18.36% of the variation is due to other factors. A really good coefficient of determination has a very small, unexplained part. So this means you probably don’t need to look for any covariates.

 

 

This is all you have to do to find how much of the variability in the response variable is explained by the linear model.

If you do want to talk about correlation, correlation is another measure of the strength of the linear relationship. The symbol for t√he coefficient correlation

is r. To find r, it is the square root of 𝑅2. So 𝑟 =𝑅2. So all you need to

±
do is take the square root of the answer you already found. The only thing is that when you take a square root of both sides of an equation you need the symbol. This is because you don’t know if you want the positive or negative root. However, remember to read graphs from left to right, the same as you read words. If the graph goes up the correlation is positive and if the graph goes down the correlation is negative. Just affix the correct sign to the r value.

The linear correlation coefficient is a number that describes the strength of the linear relationship between the two variables. It is also called the Pearson correlation coefficient after Karl Pearson who developed it.

Interpretation of the correlation coefficient

−1
−1
r is always between and 1. r = means there is a perfect negative linear correlation and r = 1 means there is a perfect positive correlation. The closer r is to 1 or , the stronger the correlation. The closer r is to 0, the weaker the correlation. CAREFUL: r = 0 does not mean there is no correlation. It just means there is no linear correlation. There might be a very strong curved pattern.

 

Example: Calculating the Linear Correlation Coef- ficient, r

How strong is the positive relationship between the alcohol content and the number of calories in 12-ounce beer? To determine if there is a linear correlation, a random sample was taken of beer’s alcohol content and calories for several different beers (”Calories in beer„” 2011), and the Data Frame are in table #10.2.1. Find the correlation coefficient and interpret that value.

Solution:

State random variables

x = alcohol content in the beer

y = calories in 12 ounce beer

√
0.8164 = 0.9035486
𝑅 = 0.8164𝑟 =𝑟 =
Fr om example 10.2.1, 2. The correlation coefficient = √ 2

The scatter graph went up from left to right, so the correlation coefficient is positive. Since 0.90 is close to 1, then that means there is a fairly strong positive correlation.

Causation

 

 

One common mistake people make is to assume that because there is a corre- lation, then one variable causes the other. This is usually not the case. That would be like saying the amount of alcohol in the beer causes it to have a cer- tain number of calories. However, fermentation of sugars is what causes the alcohol content. The more sugars you have, the more alcohol can be made, and the more sugar, the higher the calories. It is actually the amount of sugar that causes both. Do not confuse the idea of correlation with the concept of causation. Just because two variables are correlated does not mean one causes the other to happen. However, the new theory is showing that if you have a relationship between two variables and a strong correlation, and you can show that there are no other variables that could explain the change, then you can show causation. This is how doctors have shown that smoking causes bladder cancer. Just realize that proving that one caused the other is a difficult process, and causation should not be just assumed.

 

Example: Correlation Versus Causation

• A study showed a strong linear correlation between per capita beer con- sumption and teacher’s salaries. Does giving a teacher a raise cause people to buy more beer? Does buying more beer cause teachers to get a raise?

Solution:

There is probably some other factor causing both of them to increase at the same time. Think about this: In a town where people have little extra money, they won’t have money for beer and they won’t give teachers raises. In another town where people have more extra money to spend it will be easier for them to buy more beer and they would be more willing to give teachers raises.

• A study shows that there is a correlation between people who have had a root canal and those that have cancer. Does that mean having a root canal causes cancer?

Solution:

Just because there is positive correlation doesn’t mean that one caused the other. It turns out that there is a positive correlation between eating carrots and cancer, but that doesn’t mean that eating carrots causes cancer. In other words, there are lots of relationships you can find between two variables, but that doesn’t mean that one caused the other.

Remember a correlation only means a pattern exists. It does not mean that one variable causes the other variable to change.

 

Homework

For each problem, state the random variables.The Data Frame in this section are in section 10.1 and will be used in section 10.3.

 

 

• When an anthropologist finds skeletal remains, they need to figure out the height of the person. The height of a person (in cm) and the length of their metacarpal bone 1 (in cm) were collected and are in table #10.1.3 (”Prediction of height,” 2013). Find the coefficient of determination and the correlation coefficient, then interpret both.

Code book for Data Frame Metacarpal see problem 2.3.1 in Section 2.3 Homework

• Table #10.1.4 contains the value of the house and the amount of rental income in a year that the house brings in (”Capital and rental,” 2013). Find the coefficient of determination and the correlation coefficient, then interpret both.

Code book for Data Frame House see problem 2.3.2 in Section 2.3 Home- work

• The World Bank collects information on the life expectancy of a person in each country (”Life expectancy at,” 2013) and the fertility rate per woman in the country (”Fertility rate,” 2013). The Data Frame for countries for the year 2011 are in table #10.1.5. Find the coefficient of determination and the correlation coefficient, then interpret both.

Code book for Data Frame Fertility see problem 3.2.3 in Section 3.2 Home- work

• The World Bank collected data on the percentage of gross domestic prod- uct (GDP) that a country spends on health expenditures (Current health expenditure (% of GDP), 2019), the fertility rate of the country (Fertility rate, total (births per woman), 2019), and the percentage of woman re- ceiving prenatal care (Pregnant women receiving prenatal care (%), 2019). The Data Frame for the countries where this information is available in table #10.1.6. Find the coefficient of determination and the correlation coefficient, then interpret both.

Code book for Data Frame Fert_prenatal see problem 3.2.4 in Section

• Homework
• The height and weight of baseball players are in table #10.1.7 (”MLB heightsweights,” 2013). Find the coefficient of determination and the cor- relation coefficient, then interpret both.

Code book for Data Frame Baseball see problem 10.1.5 in Section 10.1 Homework

• Different species have different body weights and brain weights are in table #10.1.8. (”Brain2bodyweight,” 2013). Find the coefficient of determina- tion and the correlation coefficient, then interpret both.

Code book for Data Frame Body see problem 10.1.6 in Section 10.1 Home- work

 

 

• A sample of hot dogs was taken and the amount of sodium (in mg) and calories were measured. (”Data hotdogs,” 2013) The Data Frame are in table #10.1.9. Find the coefficient of determination and the correlation coefficient, then interpret both.

Code book for data frame Hotdog see Example 9.3.2 in Section 9.3

• Per capita income in 1960 dollars for European countries and the percent of the labor force that works in agriculture in 1960 are in table #10.1.10 (”OECD economic development,” 2013). Find the coefficient of determi- nation and the correlation coefficient, then interpret both.

Code book for Data Frame Agriculture see problem 10.1.8 in Section 10.1 Homework

• Cigarette smoking and cancer have been linked. The number of deaths per one hundred thousand from bladder cancer and the number of cigarettes sold per capita in 1960 are in table #10.1.11 (”Smoking and cancer,” 2013). Find the coefficient of determination and the correlation coefficient, then interpret both.

Code book for Data Frame Cancer see problem 10.1.9 in Section 10.1 Homework

• The weight of a car can influence the mileage that the car can obtain. A random sample of cars’ weights and mileage was collected and are in table #10.1.12 (”us auto mileage,” 2019). Find the coefficient of determination and the correlation coefficient, then interpret both.

Code book for Data Frame Cars see problem 10.1.10 in Section 10.1 Home- work

• There is a correlation between police expenditure and crime rate. Does this mean that spending more money on police causes the crime rate to decrease? Explain your answer.
• There is a correlation between tobacco sales and alcohol sales. Does that mean that using tobacco causes a person to also drink alcohol? Explain your answer.
• There is a correlation between the average temperature in a location and the morality rate from breast cancer. Does that mean that higher temper- atures cause more women to die of breast cancer? Explain your answer.
• There is a correlation between the length of time a tableware company polishes a dish and the price of the dish. Does that mean that the time a plate is polished determines the price of the dish? Explain your answer.

Inference for Regression and Correlation

The idea behind regression is to find an equation that relates the response variable to the explanatory variables, and then use that equation to predict values of the response variable from values of the explanatory variables. But how do you now how good that estimate is? First you need a measure of the variability in the estimate, called the standard error of the estimate. The definition of the standard error of the estimate is how much error is in the estimate from the regression equation.

Standard Error of the Estimate formula 𝑆_𝑒 = ∑ (𝑦̂ − 𝑦)2

This formula is hard to work with but luckily R Studio will calculate it for you. In fact the command has already been used, so this information is already found. The command is

summary(lm.out)

 

Example: Finding the Standard Error of the Esti- mate

Is there a relationship between the alcohol content and the number of calories in 12-ounce beer? To determine if there is a sample of beer’s alcohol content and calories (Find Out How Many Calories in Beer?, 2019), in table #10.1.1. Find the standard error of the estimate.

Code book for data frame Beer see Example 10.1.1

Solution:

x = alcohol content in the beer

y = calories in 12 ounce beer

First save the linear model with a name like lm.out. Then use summary(lm.out) to find the standard error of the estimate.

lm.out=lm(calories~alcohol, data=Beer)summary(lm.out)

##

## Call:

## lm(formula = calories ~ alcohol, data = Beer) ##

## Residuals:

## Min 1Q Median 3Q Max ## -51.869 -11.421 -0.145 11.855 55.473 ##

## Coefficients:

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

 

 

##	(Intercept)14.5274.5773.174 0.00171 **
##	alcohol2672.35984.478 31.634 < 2e-16 ***
##	—
##	Signif. codes:
##	0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ‘ 1
##
##	Residual standard error: 17.62 on 225 degrees of freedom
##	Multiple R-squared: 0.8164,Adjusted R-squared: 0.8156
##	F-statistic: 1001 on 1 and 225 DF, p-value: < 2.2e-16

The standard error of the estimate is the residual standard error in the output. So the standard error of the estimate for the number of calories in a beer related to alcohol content is 𝑆_𝑒 = 17.62

Prediction Interval

Using the regression equation you can predict the number of calories from the alcohol content. However, you only find one value. The problem is that beers vary a bit in calories even if they have the same alcohol content. It would be nice to have a range instead of a single value. The range is called a prediction interval.

Prediction Interval for an Individual y

𝑦̂±𝐸
Given the fixed value x, the prediction interval for an individual y iswhere

E is the error of the estimate.

R will produce the prediction interval for you. The commands are To calculate the linear model. Note this may already been done.

lm.out = lm(explanatory variable~ response variable)

 

The following will compute a prediction interval. For the prediction variable set to a particular value (put that value in place of the word value), at a particular C level (given as a decimal).

predict(lm.out, newdata=list(prediction variable = value), interval=”prediction”, level=

Example #10.3.2: Find the Prediction Interval

Is there a relationship between the alcohol content and the number of calories in 12-ounce beer? To determine if there is a sample of beer’s alcohol content and calories (Find Out How Many Calories in Beer?, 2019), in table #10.1.1. Find a 95% prediction interval for the number of calories when the alcohol content is 7.0%. content and calories (”Calories in beer„” 2011). The Data Frame are in table #10.2.1.

Solution:

x = alcohol content in the beer

 

 

y = calories in 12 ounce beer

Computing the prediction interval using R Studio

predict(lm.out, newdata=list(alcohol=0.07), interval = “prediction”, level=0.95)

 

##fitlwrupr ## 1 201.5922 166.6664 236.518

the fit = 201.5922 is the number of calories in the beer when the alcohol content is 7.0%. The prediction interval is between the lower (lwr) and upper (upr). The prediction interval is 166.6664 and 236.528. That means that 95% of all beer that has 7.0% alcohol has between 167 and 237 calories.

Hypothesis Test for Correlation or Slope: How do you really say you have a correlation or a linear relationship? Can you test to see if there really is a correlation or linear relationship? Of course, the answer is yes. The hypothesis test for correlation is as follows.

• State the random variables in words.

x = explanatory variable

y = response variable

• State the null and alternative hypotheses and the level of significance

𝐻_𝑜 ∶ there is not a correlation or linear relationship

𝐻_𝑎 ∶ there is a correlation or linear relationship Also, state your 𝛼 level here.

• State and check the assumptions for the hypothesis test

The assumptions for the hypothesis test are the same assumptions for regression and correlation.

• Find the test statistic and p-value

The test statistic and p-value is already in the summary(lm.out) output.

• Conclusion

This is where you write reject 𝐻_𝑜 or fail to reject 𝐻_𝑜. The rule is: if the p-value

< 𝛼, then reject 𝐻_𝑜. If the p-value≥ 𝛼, then fail to reject 𝐻_𝑜

• Interpretation

This is where you interpret in real world terms the conclusion to the test. The conclusion for a hypothesis test is that you either have enough evidence to support 𝐻_𝑎, or you do not have enough evidence to support 𝐻_𝑎.

 

 

Example: Testing the Claim of a Linear Correlation

Is there a linear relationship, or correlation, between the alcohol content and the number of calories in 12-ounce beer? To determine if there is a sample of beer’s alcohol content and calories (Find Out How Many Calories in Beer?, 2019), in table #10.1.1. Test at the 5% level.

Solution:

• State the random variables in words.

x = alcohol content in the beer

y = calories in 12 ounce beer

• State the null and alternative hypotheses and the level of significance

𝐻_𝑜 ∶ there is not a correlation or linear relationship

𝐻_𝑎 ∶ there is a correlation or linear relationship level of significance 𝛼 = 0.05

• State and check the assumptions for the hypothesis test

The assumptions for the hypothesis test were already checked in example #10.1.2.

• lm.out<-lm(calories~alcohol, data=Beer)summary(lm.out)
  Find the test statistic and p-value The command on R Studio is

 

 

##

## Call:

## lm(formula = calories ~ alcohol, data = Beer) ##

## Residuals:

## Min 1Q Median 3Q Max ## -51.869 -11.421 -0.145 11.855 55.473 ##

## Coefficients:

##Estimate Std. Error t value Pr(>|t|) ## (Intercept)14.5274.5773.174 0.00171 **

## alcohol2672.35984.478 31.634 < 2e-16 *** ## —

## Signif. codes:

## 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ‘ 1 ##

## Residual standard error: 17.62 on 225 degrees of freedom ## Multiple R-squared: 0.8164,Adjusted R-squared: 0.8156

 

 

## F-statistic: 1001 on 1 and 225 DF, p-value: < 2.2e-16

The test statistic is the t value in the table for the explanatory variable. In this case that is 31.634. The p-value is the Pr(>|t|) which is 2.2𝑋10−16.

• Conclusion

Reject 𝐻_𝑜 since the p-value is less than 0.05.

• Interpretation

There is enough evidence to show that there is a correlation and linear rela- tionship between alcohol content and number of calories in a 12-ounce bottle of beer.

 

Homework

** For each problem, state the random variables. The Data Frame in this section are in the homework for section 10.1 and were also used in section 10.2.**

• When an anthropologist finds skeletal remains, they need to figure out the height of the person. The height of a person (in cm) and the length of their metacarpal bone 1 (in cm) were collected and are in table #10.1.3 (”Prediction of height,” 2013).

Code book for Data Frame Metacarpal see problem 2.3.1 in Section 2.3 Homework

• Find the standard error of the estimate.
• Compute a 99% prediction interval for height of a person with a metacarpal length of 44 cm.
• Test at the 1% level for a correlation or linear relationship between length of metacarpal bone 1 and height of a person.
• Table #10.1.4 contains the value of the house and the amount of rental income in a year that the house brings in (”Capital and rental,” 2013).

Code book for Data Frame House see problem 2.3.2 in Section 2.3 Home- work

• Find the standard error of the estimate.
• Compute a 95% prediction interval for the rental income on a house worth

$230,000.

• Test at the 5% level for a correlation or linear relationship between house value and rental amount.
• The World Bank collects information on the life expectancy of a person in each country (”Life expectancy at,” 2013) and the fertility rate per woman in the country (”Fertility rate,” 2013). The Data Frame for countries for the year 2011 are in table #10.1.5.

 

Code book for Data Frame Fertility see problem 3.2.3 in Section 3.2 Home- work

• Find the standard error of the estimate.
• Compute a 99% prediction interval for the life expectancy for a country that has a fertility rate of 2.7.
• Test at the 1% level for a correlation or linear relationship between fertility rate and life expectancy.
• The World Bank collected data on the percentage of gross domestic prod- uct (GDP) that a country spends on health expenditures (Current health expenditure (% of GDP), 2019), the fertility rate of the country (Fertility rate, total (births per woman), 2019), and the percentage of woman re- ceiving prenatal care (Pregnant women receiving prenatal care (%), 2019). The Data Frame for the countries where this information is available in table #10.1.6.

Code book for Data Frame Fert_prenatal see problem 3.2.4 in Section

• Homework
• Find the standard error of the estimate.
• Compute a 95% prediction interval for the percentage of woman receiving prenatal care for a country that spends 5.0 % of GDP on health expendi- ture.
• Test at the 5% level for a correlation or linear relationship between percent- age spent on health expenditure and the percentage of women receiving prenatal care.
• The height and weight of baseball players are in table #10.1.7 (”MLB heightsweights,” 2013).

Code book for Data Frame Baseball see problem 10.1.5 in Section 10.1 Homework

• Find the standard error of the estimate.
• Compute a 95% prediction interval for the weight of a baseball player that is 75 inches tall.
• Test at the 5% level for a correlation or linear relationship between height and weight of baseball players.
• Different species have different body weights and brain weights are in table #10.1.8. (”Brain2bodyweight,” 2013).

Code book for Data Frame Body see problem 10.1.6 in Section 10.1 Home- work

• Find the standard error of the estimate.
• Compute a 99% prediction interval for the brain weight for a species that has a body weight of 62 kg.
• Test at the 1% level for a correlation or linear relationship between body weights and brain weights.

 

• A sample of hot dogs was taken and the amount of sodium (in mg) and calories were measured. (”Data hotdogs,” 2013) The Data Frame are in table #10.1.9.

Code book for data frame Hotdog see Example 9.3.2 in Section 9.3

• Find the standard error of the estimate.
• Compute a 95% prediction interval for the amount of sodium a beef hot dog has if it is 170 calories.
• Test at the 5% level for a correlation or linear relationship between amount of calories and amount of sodium.
• Per capita income in 1960 dollars for European countries and the percent of the labor force that works in agriculture in 1960 are in table #10.1.10 (”OECD economic development,” 2013).

Code book for Data Frame Agriculture see problem 10.1.8 in Section 10.1 Homework

• Find the standard error of the estimate.
• Compute a 90% prediction interval for the per capita income in a country that has 21 percent of labor in agriculture.
• Test at the 5% level for a correlation or linear relationship between percent of labor force in agriculture and per capita income.
• Cigarette smoking and cancer have been linked. The number of deaths per one hundred thousand from bladder cancer and the number of cigarettes sold per capita in 1960 are in table #10.1.11 (”Smoking and cancer,” 2013).

Code book for Data Frame Cancer see problem 10.1.9 in Section 10.1 Homework

• Find the standard error of the estimate.
• Compute a 99% prediction interval for the number of deaths from bladder cancer when the cigarette sales were 20 per capita.
• Test at the 1% level for a correlation or linear relationship between cigarette smoking and deaths of bladder cancer.
• The weight of a car can influence the mileage that the car can obtain. A random sample of cars’ weights and mileage was collected and are in table #10.1.12 (”us auto mileage,” 2019).

Code book for Data Frame Cars see problem 10.1.10 in Section 10.1 Home- work

• Find the standard error of the estimate.
• Compute a 95% prediction interval for the mileage on a car that weighs 3800 pounds.
• Test at the 5% level for a correlation or linear relationship between the weight of cars and mileage.

 

 

 

Data Source:

Find Out How Many Calories in Beer? (n.d.). Retrieved July 21, 2019, from https://www.beer100.com/beer-calories/

MLB heightsweights. (2013, November 16). Retrieved from http://wiki.stat. ucla.edu/socr/index.php/SOCR_Data_MLB_HeightsWeights

Brain2bodyweight. (2013, November 16). Retrieved from http://wiki.stat.ucla. edu/socr/index.php/SOCR_Data_Brain2BodyWeight

OECD economic development. (2013, December 04). Retrieved from http:

//lib.stat.cmu.edu/DASL/Datafiles/oecdat.html

Smoking and cancer. (2013, December 04). Retrieved from http://lib.stat.cmu. edu/DASL/Datafiles/cigcancerdat.html

(n.d.).Retrieved July 21, 2019, from https://www3.nd.edu/~busiforc/ handouts/Data and Stories/regression/us auto mileage/usautomileage.html

 

 

 

 

 

Chapter 11