Assignment 2: Methodology - PTSD

Page 1 of 1 Research and Evaluation Design

© 2013 Argosy University

Scales of Measurement The type of statistical analysis that is used depends on the type of data that the researcher has. Types of data are referred to as scales of measurement, and there are four of them: nominal, ordinal, interval, and ratio. Nominal data assigns numbers to the data without assigning any particular meaning to the numbers, such as ethnicity (African = 1, Asian = 2, Caucasian = 3, Hispanic = 4, etc.) and gender (female = 1 and male = 2). The numbers are used simply to distinguish among the data and not to suggest any comparative differences among them. For example, assigning African a “1” does not suggest that the African ethnicity is higher or more meaningful than the others. Additionally, dichotomous variables, such as yes or no questions, constitute nominal data. Data at the ordinal level is when there are comparative differences between the data, such as most/least satisfied. Ordinal-level data is also used to order participants on the basis of differences but with unequal spacing between participants. For example, ordinal data would be used to clarify the order in which participants finished a project, for example, first place = 1, second place = 2, third place = 3, and so on. The data is ordinal because there is an order to it, but the ranking may have very unequal spacing, for example, the first-place time may be 14 minutes, the second-place time may be 17 minutes, and the third-place time may be 25 minutes. Interval data has some order among the data and equal intervals between the numbers, such as temperature or Likert scale data. For example, a temperature of 60 degrees is twice as hot as a temperature of 30 degrees. Similarly, on a Likert scale that assigns numbers according to most satisfied and least satisfied, it could be said that a score of 6 indicates twice as much satisfaction as a score of 3. Ratio data is essentially the same as interval data except it has a true zero, such as weight, because it is possible for something to have zero (or no) weight. Most research is at the ordinal or interval level of data.