Chapter 18 Business Research Methods Donald Cooper Pamela

Chapter 18 Business Research Methods ä Donald Cooper ä Pamela Schindler Irwin/Mc. Graw-Hill ©The Mc. Graw-Hill Companies, Inc. , 2001

Chapter 18 Measures of Association Irwin/Mc. Graw-Hill ©The Mc. Graw-Hill Companies, Inc. , 2001

Bivariate Correlation vs. Nonparametric Click to edit Master title style Measures of Association Slide 18 - 1 ä Parametric correlation requires two continuous variables measured on an interval or ratio scale ä The coefficient does not distinguish between independent and dependent variables Irwin/Mc. Graw-Hill ÓThe Mc. Graw-Hill Companies, Inc. , 2001

Click to edit Master title. Analysis style Bivariate Correlation Pearson correlation coefficient Slide 18 - 2 är symbolized the coefficient's estimate of linear association based on sampling data äCorrelation coefficients reveal the magnitude and direction of relationships äCoefficient’s sign (+ or -) signifies the direction of the relationship ä Assumptions of r Linearity Bivariate normal distribution Irwin/Mc. Graw-Hill ÓThe Mc. Graw-Hill Companies, Inc. , 2001

Click to edit Master title. Analysis style Bivariate Correlation Slide 18 - 3 Scatterplots äProvide a means for visual inspection of data äthe direction of a relationship äthe shape of a relationship äthe magnitude of a relationship (with practice) Irwin/Mc. Graw-Hill ÓThe Mc. Graw-Hill Companies, Inc. , 2001

Interpretation of Coefficients Click to edit Master title style Slide 18 - 4 ä Relationship does not imply causation ä Statistical significance does not imply a relationship is practically meaningful Irwin/Mc. Graw-Hill ÓThe Mc. Graw-Hill Companies, Inc. , 2001

Interpretation of Coefficients Click to edit Master title style Slide 18 - 5 ä Suggests alternate explanations for correlation results äX causes Y. . . or äY causes X. . . or äX & Y are activated by one or more other variables. . . or äX & Y influence each other reciprocally Irwin/Mc. Graw-Hill ÓThe Mc. Graw-Hill Companies, Inc. , 2001

Interpretation of Coefficients Click to edit Master title style Slide 18 - 6 ä Artifact Correlations ä Goodness of fit F test Coefficient of determination Correlation matrix äused to display coefficients for more than two variables Irwin/Mc. Graw-Hill ÓThe Mc. Graw-Hill Companies, Inc. , 2001

Bivariate Click to edit. Linear Master. Regression title style Slide 18 -7 ä Used to make simple and multiple predictions Regression coefficients Slope Intercept Error term Method of least squares Irwin/Mc. Graw-Hill ÓThe Mc. Graw-Hill Companies, Inc. , 2001

Click to edit Master style Interpreting Lineartitle Regression Slide 18 - 8 Residuals äwhat remains after the line is fit or (Yi-Y^ i) Prediction and confidence bands Irwin/Mc. Graw-Hill ÓThe Mc. Graw-Hill Companies, Inc. , 2001

Click to edit Master style Interpreting Linear title Regression Goodness of fit Slide 18 - 9 äZero slope äY completely unrelated to X and no systematic pattern is evident äconstant values of Y for every value of X ädata are related, but represented by a nonlinear function Irwin/Mc. Graw-Hill ÓThe Mc. Graw-Hill Companies, Inc. , 2001

Nonparametric Measures of Click to edit Master title style Association Slide 18 - 10 ä Measures for nominal data äWhen there is no relationship at all, coefficient is 0 äWhen there is complete dependency, the coefficient displays unity or 1 Irwin/Mc. Graw-Hill ÓThe Mc. Graw-Hill Companies, Inc. , 2001

Nonparametric Measures of Click to edit Master title style Association Slide 18 - 11 Chi-square based measure äPhi Cramer’s V Contingency coefficient of C Proportional reduction in error (PRE) äLambda äTau Irwin/Mc. Graw-Hill ÓThe Mc. Graw-Hill Companies, Inc. , 2001

Click to edit Master title style Characteristics of Ordinal Data Slide 18 - 12 Concordant- subject who ranks higher on one variable also ranks higher on the other variable Discordant- subject who ranks higher on one variable ranks lower on the other variable Irwin/Mc. Graw-Hill ÓThe Mc. Graw-Hill Companies, Inc. , 2001

Click to edit Master title style Measures for Ordinal Data Slide 18 - 13 ä No assumption of bivariate normal distribution ä Most based on concordant/discordant pairs ä Values range from +1. 0 to -1. 0 Irwin/Mc. Graw-Hill ÓThe Mc. Graw-Hill Companies, Inc. , 2001

Click to edit Master title style Measures for Ordinal Data Slide 18 - 14 Tests Gamma Somer’s d Spearman’s rho Kendall’s tau b Kendall’s tau c Irwin/Mc. Graw-Hill ÓThe Mc. Graw-Hill Companies, Inc. , 2001