1. nominal scales
    • coding of categorical data
    • (i.e. males=1 females=2)
  2. Ordinal Scale
    • ranked order of data
    • (i.e. class rank from highest to lowest)
  3. Interval Scale
    • ranked order with the difference between values
    • (i.e. 2, 3, 4 have an interval of 1 between each and no absolute value)
  4. Ratio Scale
    • comparative data that is interval-level with an absolute zero
    • (i.e. person a is 5' and person b is 4' then person a is 1' taller than person b)
  5. Normal Distribution
    the variance between individual scores/scale values are normallu distributed in the general population (i.e. 95% distributed symetrically under a bell curve)
  6. Alpha Level
    • Reports that data found would occur due to chance X times out of 100.
    • Correspond to p-values (i.e. 5 out of 100 times due to chance = .05)
  7. Sample Distributions
    a distribution of actual data or test scores (mean, SD, variance)
  8. Sampling Distribution
    • theoretical distibutions with mean, variance, & SD.
    • used to set alpha levels for a statistical test; used to determine acceptance or rejection of the null hypothesis.
  9. Correlation
    • describes the strength and direction (vector) of a linear value between two variables.
    • if one variable rises in value and the second rises proportionally then it is a positive manner correlation. However, if one variable rises in value and one decreases in value the the correlation is negative manner.
  10. Linear Regression
    requires the indication of one variable is dependant on another variable (a predictor).
  11. z score
    • standardized score with a mean of zero and a SD of 1.
    • z score = (raw score - mean)/SD
  12. T score
    • converted z score.
    • New distribution score = (z score x SD new) + New Mean.
  13. Construct
    "an idea or perception resulting from a synthesis of sense impressions."
  14. construct cleanliness
    • occures when the construct evaluates what they are supposed to.
    • when imperfections occur there are two types of imperfections: deficiency (when a measure does not cover the domain of a construct in its entirety); and contamination (when a measure contains material that should not be a part of the construct).
Card Set
Basic concepts for Psychometrics