LESSON 8

Dr. John M. Ritz

Correlation vs True Causality

    * Correlation is only a statistical degree of relationship.
    * We do not know whether one variable causes or influences the other.
    * Correlation means relationship and not cause.

Methods of Linear Correlation

    * Pearson r (product moment correlation)
    * Spearman Rho (ranked order correlation)

Pearson's r Product Moment Correlation

    * Used to determine if there is a relationship between two sets of paired numbers.
    * The basic requirement is that both sets of measurements be interval data (generally raw scores).
    * Examples: IQ to grade point average, salary to years of education, TVs to income, reading score to IQ, GRE to graduate grade point average, etc.

Calculating r

    * Determine the number of pairs of scores.
    * Sum the product of the pairs of scores or Sigma XY.
    * Sum the scores on Test X.
    * Sum the scores on Test Y.
    * Sum the squared scores from Test X.
    * Sum the squared scores from Test Y.
    * Plug and Chug!

Interpret r

    * Check the degree of freedom ( is the data good and free of error?).
    * Check for significance of results.

Spearman's Rho

    * Used when an experimenter wishes to determine whether two sets of ranked ordered data are related.
    * It is very useful for correlating rankings and other ordinal data such as ranked leadership ability and IQ, achievement potential and IQ, class standing and income, etc.

Interpret Rho

    * Check for degree of freedom (error).
    * Check for level of correlation.