13.07.2015 Views

Contents

Contents

Contents

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

166 PART II: Descriptive MethodsTABLE 5.1EXAMPLE OF CONSTRUCT VALIDITY*SWLS LS-5 PASWLS (.88)LS-5 .77 (.90)PA .42 .47 (.81)*Data from Lucas et al. (1996), Table 3.Note: SWLS Satisfaction with Life Scale; LS-5 5-item Life Satisfaction scale; PA Positive Affect scale.Table 5.1 presents data showing how we might assess the construct validityof a measure of “life satisfaction.” Lucas, Diener, and Suh (1996) note thatpsychologists are increasingly examining factors such as happiness, life satisfaction,self-esteem, optimism, and other indicators of well-being. However,it’s not clear whether these different indicators all measure the same construct(e.g., well-being) or whether each is a distinguishable construct. Lucas andhis colleagues conducted several studies in which they asked individuals tocomplete questionnaire measures of these different indicators of well-being.For our purposes we will focus on a portion of their data from their thirdstudy, in which they asked participants to complete three scales: two lifesatisfaction measures, the Satisfaction with Life Scale (SWLS) and a 5-itemLife Satisfaction measure (LS-5); and a measure of Positive Affect (PA). Atissue in this example is whether the construct of life satisfaction—the qualityof being happy with one’s life—can be distinguished from being happy moregenerally (positive affect).The data in Table 5.1 are presented in the form of a correlation matrix. A correlationmatrix is an easy way to present a number of correlations. Look first atthe values in parentheses that appear on the diagonal. These parenthesized correlationcoefficients represent the values for the reliability of each of the threemeasures. As you can see, the three measures show good reliability (each isabove .80). Our focus, however, is on measuring the construct validity of “lifesatisfaction,” so let’s look at what else is in Table 5.1.It is reasonable to expect that scores on the Satisfaction with Life Scale (SWLS)should correlate with scores on the 5-item Life Satisfaction measure; after all,both measures were designed to assess the life satisfaction construct. In fact,Lucas et al. observed a correlation between these two measures of .77, whichindicates that they correlate as expected. This finding provides evidence for convergentvalidity of the measures; the two measures converge (or “go together”)as measures of life satisfaction.The case for the construct validity of life satisfaction can be made even morestrongly when the measures are shown to have discriminant validity. As canbe seen in Table 5.1, the correlations between the Satisfaction with Life Scale(SWLS) and Positive Affect (.42) and between the 5-item Life Satisfaction measure(LS-5) and Positive Affect (.47) are lower. These findings show that lifesatis faction measures do not correlate as well with a measure of another theoreticalconstruct—namely, positive affect. The lower correlations between thelife satisfaction tests and the positive affect test indicate that different constructsare being measured. Thus, there is evidence for discriminant validity of the life

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!