Identifying and Measuring the Associations among Two Sets of Variables: An Application Canonical Correlation Analysis


Tüzüntürk S.

1st International Congress on Economics Public Finance Business & Social Sciences, Kütahya, Turkey, 9 - 11 November 2023, pp.62-63

  • Publication Type: Conference Paper / Summary Text
  • City: Kütahya
  • Country: Turkey
  • Page Numbers: pp.62-63
  • Bursa Uludag University Affiliated: Yes

Abstract

A statistical relationship between any two variables, whether causal or not, refers to correlation. In correlation analysis, correlation coefficients such as Pearson or Spearman correlation coefficients are calculated to measure the extent to which two variables are related.  This type of analysis is known as bivariate analysis in statistics. On the other hand, multivariate analysis methods are used in cases where there are more than two variables. When it is desired to determine and measure the relationships between two sets of variables, the canonical correlation analysis method, one of the multivariate analysis methods, is used.

In this study, it is aimed to investigate the relationships between variable sets used in measuring the concepts of perceived quality, perceived value, loyalty and image regarding the services provided to the customers of a luxury car brand in Bursa by using the canonical correlation analysis method. For this purpose, a survey was conducted among automobile service customers and the obtained data was analyzed in the SPSS package program. Research findings show that the perceived quality variable set has a correlation of 0.84 with the perceived value variable set, 0.76 with the loyalty variable set and 0.86 with the image variable set. On the other hand, it was found that the perceived value variable set had a correlation of 0.86 with the loyalty variable set and 0.89 with the image variable set. It was found that the loyalty variable set had a correlation of 0.85 with the image variable set. As a result, all sets of variables appear to be highly correlated with each other.