Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Attitude Measurement


Published on

This presentation was part of my graduate school studies.

  • Be the first to comment

Attitude Measurement

  1. 1. Attitude Measurement Matt Hendrickson Sopsy 640: Attitudes Dr. White August 31, 2006
  2. 2. Attitude Measurement <ul><li>There are numerous scales to measure attitudes </li></ul><ul><li>Here are a few of the most common </li></ul><ul><ul><li>Thurstone </li></ul></ul><ul><ul><li>Likert/Likert-Type </li></ul></ul><ul><ul><li>Guttman </li></ul></ul><ul><ul><li>Semantic Differential </li></ul></ul>
  3. 3. Factor Analysis Primer From Whitley (2002) <ul><li>Many of these scales use factor analysis as a basis for their usability </li></ul><ul><li>This method creates subsets of variables; with the subset items being correlated with each other, but not the other subsets </li></ul><ul><li>There are two main uses for factor analysis </li></ul><ul><ul><li>Data reduction </li></ul></ul><ul><ul><li>Scale development </li></ul></ul>
  4. 4. Factor Analysis cont. <ul><li>Data reduction </li></ul><ul><ul><li>Condenses large numbers of variables into a few for analysis simplification </li></ul></ul><ul><li>Scale development </li></ul><ul><ul><li>Scales should represent only one hypothetical construct </li></ul></ul><ul><ul><li>FA is used after the data has been collected on a pilot run and it determines if the items are intercorrelated </li></ul></ul><ul><ul><li>If more than one factor is present, the scale is not measuring just one construct </li></ul></ul>
  5. 5. Thurstone (1928) <ul><li>Was intended to create a validated measure for assessing attitudes </li></ul><ul><li>See summary in Thurstone (1928) p. 552 </li></ul><ul><ul><li>Specify the attitude to be measured </li></ul></ul><ul><ul><li>Collect a variety of options </li></ul></ul><ul><ul><li>Edit list to ≈ 100 statements of opinion </li></ul></ul><ul><ul><li>Sort statements into imaginary scale (300 p’s) </li></ul></ul><ul><ul><li>Calculate scale value of each statement </li></ul></ul><ul><ul><li>Eliminate statements due to ambiguity </li></ul></ul><ul><ul><li>Eliminate statements due to irrelevancy </li></ul></ul><ul><ul><li>Select ≈ 20 statements evenly across scale </li></ul></ul>
  6. 6. Thurstone <ul><li>Although this measure was the first to apply principles of scaling to attitudes (Whitley, 2002), its popularity has diminished </li></ul><ul><li>These scales are less valid than Likert scales (of the same length) and take much longer to construct (Whitley, 2002) </li></ul><ul><li>This scale has been noted for its influence and for drawing attention to measuring attitude constructs (Dawes, 1994) </li></ul><ul><li>Lack external validity—not as useful outside of the classroom (Likert, 1970; p. 153) </li></ul><ul><li>There are alternative scorings for Thurstone scales that increase the validity from .76  .85 (Likert, Roslow, & Murphy, 1993) </li></ul>
  7. 7. Thurstone example In Whitley (2002) p. 364
  8. 8. Likert <ul><li>A.K.A. summated rating scales (Whitley, 2002) </li></ul><ul><li>Most widely used scale </li></ul><ul><li>Most direct and the easiest way to measure attitudes (Feldman, 2001) </li></ul><ul><li>Scored based on the sums of their responses (Whitely, 2002) </li></ul><ul><li>There has been discussion to the number of possible responses </li></ul><ul><ul><li>Anchoring-the verbal labels put to the numbers on measurement scales </li></ul></ul><ul><ul><li>Should one use 1-5; 1-7; - 3- + 3? </li></ul></ul>
  9. 9. Likert vs. Likert-Type <ul><li>Difference between Likert and Likert-Type </li></ul><ul><ul><li>Likert: must go through development process </li></ul></ul><ul><ul><li>Likert-type: just add numbers to an ordinal plane </li></ul></ul><ul><li>To be a true Likert scale, the scale must utilize all four of these steps (Whitley, 2002) </li></ul><ul><ul><li>Write a large number of items for the variable to be measured, which must include both extremes; include numbers with the anchors </li></ul></ul><ul><ul><li>Administer the items to a large number of respondents (minimum of 100) </li></ul></ul><ul><ul><li>Conduct an item analysis (discriminates between high and low scorers)—ensures internal consistency (.7<) </li></ul></ul><ul><ul><li>Highest item-total correlations create the final scale </li></ul></ul><ul><li>Likert scales are unidimensional (assumption) </li></ul><ul><ul><li>Can get at multidimensional constructs, but must use sub-scales </li></ul></ul>
  10. 10. Likert Scaling In Shaughnessy, Zechmeister, & Zechmeister (2003) p. 150
  11. 11. Likert example In Whitley (2002) p. 362
  12. 12. Guttman <ul><li>Uses a gradation of attidudes, from least to most extreme (Feldman, 2001) </li></ul><ul><li>Based on the premise that you endorse all of the items up to a point, and none after that point (Feldman, 2001) </li></ul><ul><li>The last point endorsed is the p’s score (Whitley, 2002) </li></ul><ul><li>Primarily useful when the attitude being assessed follows a stepwise sequence. </li></ul><ul><ul><li>Rarely used for this reason, most behaviors are not stepwise sequences </li></ul></ul>
  13. 13. Guttman example In Whitley (2002) p. 365
  14. 14. Semantic Differential <ul><li>Allows the targeting of an evaluative component of the attitude, as well as accessing broad concepts, rather than specific ones (Feldman, 2001) </li></ul><ul><li>Has p’s rate the concept on sets of bipolar adjective pairs; for instance, good-bad , active passive , strong-weak (Whitley, 2002) </li></ul><ul><li>Scored by adding up item values (7 point scale, usually - 3 to + 3; Whitley, 2002) </li></ul><ul><li>Also have false SD scales, use of arbitrary pairs </li></ul><ul><li>Osgood et al. (1957, pp. 53-55; in Whitley, 2002) have developed a large listing of pairs that have been tested and can be used for this task </li></ul>
  15. 15. Semantic Differential example In Feldman (2001) p. 337
  16. 16. Likert, SD, & Guttman examples In Feldman (2001) p. 337
  17. 17. References: <ul><li>Dawes, R. M. (1994). Psychological Measurement. Psychological Review, 101, 278-281. </li></ul><ul><li>Feldman, R. S. (2001). Social Psychology (3 rd ed.). Upper Saddle River, New Jersey: Prentice Hall. </li></ul><ul><li>Likert, R. (1970). A technique for the measurement of attitudes. In G. F. Summers (Ed.), Attitude assessment (pp. 149-158). Chicago: Rand-McNally. (Excerpted from: A technique for the measurement of attitudes, Archives of Psychology , 1932, No. 140). </li></ul><ul><li>Likert, R., Roslow, S., & Murphy, G. (1993). A simple and reliable method of scoring the Thurstone Attitude Scales. Personnel Psychology, 46, 689-690. </li></ul><ul><li>Shaughnessy, J. J., Zechmeister, E. B., & Zechmeister, J. S. (2003). Research Methods in Psychology (6 th ed.). New York: McGraw-Hill. </li></ul><ul><li>Thurstone, L. L. (1928). Attitudes can be measured. American Journal of Sociology, 33, 529-554. </li></ul><ul><li>Whitley, B. E. Jr. (2002). Principles of Research in Behavioral Science (2 nd ed.). New York: McGraw-Hill. </li></ul>