Other specific considerations

krajcsi edited this page Jun 8, 2018 · 7 revisions

Names of the measurement levels

Usually nominal and ordinal terms are used for these two measurement levels.

However, for a third level interval, scale (e.g., in SPSS) or continuous (e.g. in jamovi) terms are also used. CogStat uses the interval term.

  • Scale is not recommended, because it might be ambiguous, also noting the meaning of measurement tools.
  • Continuous is imprecise, because there could be interval scales that are discrete. In fact, the term is orthogonal to the measurement level, even if the two dimensions correlate.
  • There are ratio level variables, but in practice they are usually handled as interval variables, and CogStat (similar to other statistical software packages) does not offer ratio-specific statistical calculations.

Using "ordinal" information for interval variables

CogStat calculates some statistics for interval variables that are seemingly ordinal statistics, e.g., median or Spearman correlation coefficient. This is appropriate, because first, a variable can be handled as a lower measurement level variable, e.g., an interval variable can be handled as an ordinal or nominal variable, although in those cases we drop some part of the information the data include. Second, ordinal statistics have some attractive features, such as not sensitive to outliers, not sensitive to violation of normality, etc.

Clone this wiki locally
You can’t perform that action at this time.
You signed in with another tab or window. Reload to refresh your session. You signed out in another tab or window. Reload to refresh your session.
Press h to open a hovercard with more details.