Scribd
Upload
79 views23 pages

Association Between Variables Measured at The Interval-Ratio Level

This chapter discusses interpreting the correlation coefficient r and scatterplots. Scatterplots graphically display the relationship between two interval or ratio variables with one variable on the x-axis and the other on the y-axis. The regression line approximates the relationship between the variables and its slope and position indicate the strength and direction of the correlation. Pearson's r measures the strength and direction of the linear association between two variables and can range from -1 to 1. Sample computations are provided to demonstrate calculating r, the regression line formula, and interpreting the results.

Uploaded by

David Renz Pila Bonifacio
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
79 views23 pages

Association Between Variables Measured at The Interval-Ratio Level

This chapter discusses interpreting the correlation coefficient r and scatterplots. Scatterplots graphically display the relationship between two interval or ratio variables with one variable on the x-axis and the other on the y-axis. The regression line approximates the relationship between the variables and its slope and position indicate the strength and direction of the correlation. Pearson's r measures the strength and direction of the linear association between two variables and can range from -1 to 1. Sample computations are provided to demonstrate calculating r, the regression line formula, and interpreting the results.

Uploaded by

David Renz Pila Bonifacio
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
You are on page 1/ 23

Chapter 15

Association Between Variables Measured at the Interval-Ratio Level

Chapter Outline

Interpreting the Correlation Coe icient! r " #he Correlation Matri$ #esting %earson&s r or 'igni icance Interpreting 'tatistics! #he Correlates o Cri(e

'cattergra(s

'cattergra(s have two di(ensions!

#he ) *independent+ variable is arra,ed along the hori-ontal a$is. #he / *dependent+ variable is arra,ed along the vertical a$is.

'cattergra(s

0ach dot on a scattergra( is a case. #he dot is placed at the intersection o the case&s scores on ) and /.

'cattergra (s

'hows the relationship between 1 College 0ducated *)+ and Voter #urnout */+ on election da, or the 52 states.
Turnout By % College
73 68

63

58

53

48

43

15

17

19

21

23

25 % College

27

29

31

33

35

'cattergra(s

3ori-ontal ) a$is - 1 o population o a state with a college education. 'cores range ro( 15.41 to 45.61 and increase ro( le t to right.
Turnout By % College
73 68 63 58 53 48 43 15 17 19 21 23 25 % College 27 29 31 33 35

'cattergra(s

Vertical */+ a$is is voter turnout. 'cores range ro( 55.11 to 72.51 and increase ro( botto( to top
Turnout By % College
73 68 63 58 53 48 43 15 17 19 21 23 25 % College 27 29 31 33 35

'cattergra(s! Regression Line


A single straight line that co(es as close as possible to all data points. Indicates strength and direction o the relationship.
Turnout By % College
73 68 63 58 53 48 43 15 17 19 21 23 25 % College 27 29 31 33 35

'cattergra(s! 'trength o Regression Line

#he greater the e$tent to which dots are clustered around the regression line8 the stronger the relationship. #his relationship is wea9 to (oderate in strength.
Turnout By % College
73 68 63 58 53 48 43 15 17 19 21 23 25 % College 27 29 31 33 35

'cattergra(s! :irection o Regression Line


%ositive! regression line rises le t to right. ;egative! regression line alls le t to right. #his a positive relationship! As 1 college educated increases8 turnout increases.
Turnout By % College
73 68 63 58 53 48 43 15 17 19 21 23 25 % College 27 29 31 33 35

'cattergra(s

Inspection o the scattergra( should alwa,s be the irst step in assessing the correlation between two I-R variables
Turnout By % College
73 68 63 58 53 48 43 15 17 19 21 23 25 % College 27 29 31 33 35

#he Regression Line! <or(ula

#his or(ula de ines the regression line!


Y = a + bX =here! / > score on the dependent variable a > the / intercept or the point where the regression line crosses the / a$is. b > the slope o the regression line or the a(ount o change produced in / b, a unit change in ) ) > score on the independent variable

Regression Anal,sis

Be ore using the or(ula or the regression line8 a and b (ust be calculated. Co(pute b irst8 using <or(ula 15.4 *we won&t do an, calculation or this chapter+

Regression Anal,sis

#he / intercept *a+ is co(puted ro( <or(ula 15.5!

Regression Anal,sis

<or the relationship between 1 college educated and turnout!


Regression or(ula! / > 52.24 ? .5" ) A slope o .5" (eans that turnout increases b, .5" *less than hal a percent+ or ever, unit increase o 1 in 1 college educated. #he / intercept (eans that the regression line crosses the / a$is at / > 52.24.

b *slope+ > .5" a */ intercept+> 52.24

%redicting /

=hat turnout would be e$pected in a state where onl, 121 o the population was college educated@ =hat turnout would be e$pected in a state where 721 o the population was college educated@ #his is a positive relationship so the value or / increases as ) increases!

<or ) >128 / > 52.4 ?.5"*12+ > 55.5 <or ) >728 / > 52.4 ? .5"*72+ > 7A.7

%earson correlation coe icient

But o course8 this is Bust an esti(ate o turnout based on 1 college educated8 and (an, other actors also a ect voter turnout. 3ow (uch o the variation in voter turnout depends on 1 college educated@ #he relevant statististic is the coe icient o deter(ination *r sCuared+8 but irst we need to learn about %earson&s correlation coe icient *r+.

%earson&s r

%earson&s r is a (easure o association or I-R variables. It varies ro( -1.2 to ?1.2 Relationship (a, be positive *as ) increases8 / increases+ or negative *as ) increases8 / decreases+ <or the relationship between 1 college educated and turnout8 r >.4". #he relationship is positive! as level o education increases8 turnout increases. 3ow strong is the relationship@ <or that we use R sCuared8 but irst8 let&s loo9 at the calculation process

0$a(ple o Co(putation

#he co(putation and interpretation o a8 b8 and %earson&s r will be illustrated using %roble( 15.1. #he variables are!

Voter turnout */+ Average ,ears o school *)+ #his is onl, to si(pli , co(putations8 5 is (uch too s(all a sa(ple or serious research.

#he sa(ple is 5 cities.

0$a(ple o Co(putation
Cit, A B C : 0 ) 11.A 1".1 1".7 1".D 14.2 / 55 62 65 6D 72

#he scores on each variable are displa,ed in table or(at!


/ > #urnout ) > /ears o 0ducation

0$a(ple o Co(putation
) 11.A 1".1 1".7 1".D 14.2 62.5 / 55 62 65 6D 72 318 )
"

/" 42"5 4622 5""5 56"5 5A22 20374

)/ 655.5 7"6 D"5.5 D72.5 A12 3986.4

151.61 156.51 161."A 164.D5 16A 782.15

Sums are needed to co(pute b8 a8 and %earson&s r.

Interpreting %earson&s r

An r o 2.AD indicates an e$tre(el, strong relationship between average ,ears o education and voter turnout or these ive cities. #he coe icient o deter(ination is r" > .A6. Enowing education level i(proves our prediction o voter turnout b, A61. #his is a %R0 (easure *li9e la(bda and ga((a+ =e could also sa, that education e$plains A61 o the variation in voter turnout.

Interpreting %earson&s r

Our irst e$a(ple provides a (ore realistic value or r.

#he r between turnout and 1 college educated or the 52 states was!


r > .4" #his is a wea9 to (oderate8 positive relationship.

#he value o r" is .12. %ercent college educated e$plains 121 o the variation in turnout.

You might also like