Download presentation

Presentation is loading. Please wait.

Published byEmory Nash Modified over 6 years ago

1
1 Objective Investigate how two variables (x and y) are related (i.e. correlated). That is, how much they depend on each other. Section 10.2 Correlation between two variables (x and y)

2
2 Definitions A correlation exists between two variables when the values of one appears to somehow affect the values of the other in some way. In this class, we are only interested in linear correlation

3
3 Linear correlation coefficient : r A numerical measure of the strength of the linear relationship between two variables, x and y, representing quantitative data. r always belongs in the interval (-1,1) ( i.e. –1 r 1 ) We use this value to conclude if there is (or is not) a linear correlation between the two variables. Definitions

4
4 Exploring the Data We can often see a relationship between two variables by constructing a scatterplot.

5
5 Positive Correlation We say the data has positive correlation if the data follows a line (with a positive slope). The correlation coefficient (r) will be close to +1

6
6 Negative Correlation We say the data has negative correlation if the data follows a line (with a negative slope). The correlation coefficient (r) will be close to –1

7
7 We say the data has no correlation if the data does not seem to follow any line. The correlation coefficient (r) will be close to 0 No Correlation

8
8 r ≈ 1 Strong positive linear correlation r ≈ 0 Weak linear correlation r ≈ -1 Strong negative linear correlation Interpreting r

9
9 Nonlinear Correlation The data may follow a curve, but if the data is not linear, the linear correlation coefficient (r) will be close to zero.

10
10 1. The sample of paired ( x, y ) data is a random sample of quantitative data. 2. Visual examination of the scatterplot must confirm that the points approximate a straight-line pattern. 3. The outliers must be removed if they are known to be errors. (Note: We will not do this in this course) Requirements

11
11 n Number of pairs of sample data Denotes the addition of the items x The sum of all x- values x = x 1 + x 2 +…+ x n y The sum of all y- values y = y 1 + y 2 +…+ y n Notation

12
12 x 2 The sum of the squares for all x - values x 2 = x 1 2 + x 2 2 +…+ x n 2 ( x ) 2 The sum of the x -values, then squared ( x) 2 = (x 1 + x 2 +…+ x n ) 2 xy The sum of the products of x and y xy = x 1 y 1 + x 1 y 2 +…+ x n y n Notation

13
13 r Sample linear correlation coefficient Population linear correlation coefficient (i.e. the linear correlation between the two populations) Correlation Coefficient n( xy) – ( x)( y) n( x 2 ) – ( x) 2 n( y 2 ) – ( y) 2 r =r = r measures the strength of a linear relationship between the paired values in a sample. We use StatCrunch compute r (Don’t panic!)

14
14 Make a scatterplot for the heights of mother, daughter Enter data on StatCrunch (Mother in 1 st column, daughter in 2 nd column) 1 Example 1a

15
15 Graphics – Scatter Plot Make a scatterplot for the heights of mother, daughter 2 Example 1a

16
16 Select var1 for X variable (height of mother) Select var2 for Y variable (height of daughter) Click Create Graph! Make a scatterplot for the heights of mother, daughter 3 Example 1a

17
17 Voila! (Does there appear to be correlation?) Make a scatterplot for the heights of mother, daughter 4 Example 1a

18
18 Find the linear correlation coefficient of the heights Example 1b Stat – Summary Stats – Correlation 1

19
19 Select var1 and var2 so they appear in the right box Click Calculate 2 Find the linear correlation coefficient of the heights Example 1b

20
20 Find the linear correlation coefficient of the heights Example 1b 3 The Correlation Coefficient is r = 0.802 (rounded)

21
21 Determining if Correlation Exists We determine whether a population is correlated via a two-tailed test on a sample using a significance level (α) H 0 : ρ = 0 (i.e. not correlated) H 1 : ρ ≠ 0 (i.e. is correlated) Again, two methods available: Critical Regions(Use Table A-6) P-value(Use StatCrunch) Note: In most cases we use significance level = 0.05

22
22 Use Table A-6 to find the critical values, which depends on the sample size n. Use both positive a negative values (two- tailed) ● If the r is in the critical region, we conclude that there is a linear correlation. (Since H 0 is rejected) ● If the r is not in the critical region, we say there is insufficient evidence of correlation. (Since we fail to reject H 0 ) Using Critical Regions 1 0

23
23

24
24 Use a 0.05 significance level to determine if the heights are linearly correlated. Example 1c Using Critical Regions ● From Example 1b, we found r = 0.802 ● Since n = 20 and α = 0.05, using Table A-6, we find the critical values to be: 0.444, -0.444 Since r is in the critical region (reject H 0 ), we conclude the data is linearly correlated (under 0.05 significance).

25
25 Use StatCrunch to calculate the two-tailed P-value from a sample set (see Example 1c) ● If the P-value is less than α, we conclude that there is a linear correlation. (Since H 0 is rejected) ● If the P-value is greater than α, we say there is insufficient evidence of correlation. (Since we fail to reject H 0 ) Using P-value

26
26 Use a 0.05 significance level to determine if the heights are linearly correlated. Example 1c Using P-value ● On StatCrunch: Stat – Summary Stats – Correlation ● Select var1, var2 so they appear in right box Click Next ● Check “Display two-sides P-value from sig. test” Click Calculate ● Result: P-value < 0.0001 Since P-value is less than α=0.05 (reject H 0 ), we conclude the data is linearly correlated

27
27 Caution! Know that the methods of this section apply only to a linear correlation. If you conclude that there is no linear correlation, it is possible that there is some other association that is not linear.

28
28 Round r to three decimal places so that it can be compared to critical values in Table A-6 Rounding the Linear Correlation Coefficient

29
29 Properties of the Linear Correlation Coefficient r 1. –1 r 1 2.If all values of either variable are converted to a different scale, the value of r does not change. 3. The value of r is not affected by the choice of x and y. Interchange all x- values and y- values and the value of r will not change. 4. r measures strength of a linear relationship. 5. r is very sensitive to outliers, they can dramatically affect its value.

30
30 A new medication for high blood pressure was tested on a batch of 18 patients with different ages. The results were as follows: (a) Plot the points (b) Find the correlation coefficient (c) Use a 0.05 significance level to test linear correlation Example 2 123456789101112131415161718 Age 563476125633676922116543236619842739 Blood Pressure 1941332507120113322723012468219157123222182298113146

31
31 ● Enter data in StatCrunch ● Go to: Graphics – Scatter Plot ● Select var1 and var2, hit Create Graph! Example 2a Plot the points 123456789101112131415161718 Age 563476125633676922116543236619842739 Blood Pressure 1941332507120113322723012468219157123222182298113146

32
32 ● Go to: Stats – Summary Stats – Correlation ● Select var1 and var2, hit Calculate Example 2b r = 0.964 123456789101112131415161718 Age 563476125633676922116543236619842739 Blood Pressure 1941332507120113322723012468219157123222182298113146 Find Correlation Coefficient

33
33 Using Critical Values Given n=18 and α=0.05, using Table A-6, the critical values are 0.468, -0.468 r = 0.964 Example 2c Test for Correlation (α = 0.05) Since r is in the critical region, (reject H 0 ), we conclude there is linear correlation 123456789101112131415161718 Age 563476125633676922116543236619842739 Blood Pressure 1941332507120113322723012468219157123222182298113146

34
34 Using P-value ● Go to: Stats – Summary Stats – Correlation ● Select var1 and var2, hit Next ● Check box, hit Calculate Example 2c Test for Correlation (α = 0.05) P-value < 0.0001 Since P-value less than α=0.05 (reject H 0 ), we conclude there is linear correlation r = 0.964 123456789101112131415161718 Age 563476125633676922116543236619842739 Blood Pressure 1941332507120113322723012468219157123222182298113146

35
35 A survey of 15 people was conducted to see how many friends people had on Facebook vs. their shoe size. The results were as follows: (a) Plot the points (b) Find the correlation coefficient (c) Use a 0.05 significance level to test linear correlation Example 3 123456789101112131415 Friends on FB170 680510680425 68085680850 68017510 Shoe size8.49.09.17.88.88.78.89.18.59.69.18.29.38.09.4

36
36 ● Enter data in StatCrunch ● Go to: Graphics – Scatter Plot ● Select var1 and var2, hit Create Graph! 123456789101112131415 Friends on FB170 680510680425 68085680850 68017510 Shoe size8.49.09.17.88.88.78.89.18.59.69.18.29.38.09.4 Example 3a Plot the points

37
37 ● Go to: Stats – Summary Stats – Correlation ● Select var1 and var2, hit Calculate Example 3b 123456789101112131415 Friends on FB170 680510680425 68085680850 68017510 Shoe size8.49.09.17.88.88.78.89.18.59.69.18.29.38.09.4 Find Correlation Coefficient r = 0.409

38
38 Using Critical Values Given n=15 and α=0.05, using Table A-6, the critical values are 0.514, -0.514 123456789101112131415 Friends on FB170 680510680425 68085680850 68017510 Shoe size8.49.09.17.88.88.78.89.18.59.69.18.29.38.09.4 r = 0.409 Example 3c Test for Correlation (α = 0.05) Since r is not in the critical region, (fail to reject H 0 ), we conclude there is no correlation

39
39 Using P-value ● Go to: Stats – Summary Stats – Correlation ● Select var1 and var2, hit Next ● Check box, hit Calculate 123456789101112131415 Friends on FB170 680510680425 68085680850 68017510 Shoe size8.49.09.17.88.88.78.89.18.59.69.18.29.38.09.4 Example 3c Test for Correlation (α = 0.05) P-value = 0.1297 Since P-value greater than α=0.05 (fail to reject H 0 ), we conclude there is no correlation r = 0.409

40
40 Interpreting r : Explained Variation The value of r 2 is the proportion of the variation in y that is explained by the linear relationship between x and y. r = 0.623 r 2 = 0.388 r = 0.998 r 2 = 0.996 Low varianceHigh variance

41
41 Using the data in Example 2 (blood pressure vs. age), we found that the linear correlation coefficient is r = 0.964 What proportion of the variation in the patients’ blood pressure can be explained by the variation in the patients’ age? With r = 0.964, we get r 2 = 0.929 We conclude that 0.929 (or about 93%) of the variation in blood pressure can be explained by the linear relationship between the age and blood pressure. This implies about 7% of the variation in blood pressure cannot be explained by the age. Example 4

42
42 Common Errors Involving Correlation 1. Causation: It is wrong to conclude that correlation implies causality. 2. Linearity: There may be some relationship between x and y even when there is no linear correlation.

43
43 Caution!!! Know that correlation does not imply causality. There may be correlation without causality.

Similar presentations

© 2021 SlidePlayer.com Inc.

All rights reserved.

To make this website work, we log user data and share it with processors. To use this website, you must agree to our Privacy Policy, including cookie policy.

Ads by Google