Wikiversity

Survey research and design in psychology/Tutorials/Psychometrics/General steps

< Survey research and design in psychology‎ | Tutorials‎ | Psychometrics
Psychometrics tutorial - General steps

General steps edit

View the accompanying screencasts: [1] [2]

Recommended psychometric instrument development steps emphasised in this tutorial are:

  1. Check and prepare the data
    1. Reverse code (if necessary) so that all variables are scored in the same direction
    2. Examine the univariate properties (e.g., histogram, mean, standard deviation, skewness, and kurtosis)
    3. Examine the bivariate properties (e.g., scatterplots and correlations) to help understand the data.
  2. Exploratory factor analysis
    1. Decide on a purpose and ask a research question
      (e.g., what are the underlying dimensions of X?)
    2. Examine assumptions (sample size, linear relations, and factorability)
    3. Choose method of:
    4. Determine the number of factors
    5. Name and describe each factor
    6. Identify which items belong to each factor, drop problematic items as necessary, and repeat analysis
  3. Analyse internal consistency for each factor
  4. Examine correlations amongst the factors
  5. Compute composite scores (if needed)
  6. Check factor structure across sub-groups

Retrieved from "https://en.wikiversity.org/w/index.php?title=Survey_research_and_design_in_psychology/Tutorials/Psychometrics/General_steps&oldid=1664388"