variables, levels and measurement

• constant error


  • occurs when we have a systematic effect of some extraneous variables on one of our experimental conditions
  • seriously damages what we can say about the results because we may be measuring an effect that is not really present
• minimizing constant error


  • randomly assigning participants to conditions
• random error


  • where we have an effect from an extraneous variable on our measurement of the relationship between the IV and DV that is random
  • obscures the effect we are interested in identifying
• when we have a biased result due to constant error from an extraneous variable we say our results have been confounded
• confounding occurs when


  a real effect between the IV and the DV is biased or distorted by another variable
• ways confounding variables can act on the relationship between an IV and DV


  • an effect being observed in the DV on the basis of the IV when there was no effect produced by the IV
  • can obscure a real effect between an IV and DV
• nominal

  • a way of labelling categories of data
  • at this level we assign numbers to meaningful categories
• ordinal


  • simplest scale where phenomena are ordered along some continuum
  • phenomena ranked or ordered along some dimension of meaning
• interval


  • the differences between items on the scale really mean something
  • interval scales have intervals equal in amount across the entire scale
  • most common form of measurement used in psychological research
• ratio


  • unlike interval has a true zero point
  • results in a scale where the relative proportions make sense when compared with each other
• discrete variables


  • take on only a limited number of values
• continuous variables 


  • can have any value between the lowest and highest points on the scale