1. Correlations

• 1. Correlation:

  Relationship between 2 co-variables.
• 2. Positive and Negative Correlations

  Positive - 1 variable increases, other increases too.

  Negative - 1 variable increases, other decreases.
• 3. Name 2 things correlations presented on:

  Graphically (scattergraph).

  Statistically (Spearman’s rank correlation coefficient).
• 4. Correlations (vs Experiments)

  Don’t manipulate variables.

  Only measures pre-existing variables (e.g. age).
• 5. Strengths:

  • Useful when not ethical/practical to do experiment (bc variables can’t be manipulated).
  • Quick/cheap - no need for controlled environment, uses secondary data, etc.
  • Allows researchers make predictions about similar situations.
• 6. Limitations:

  • Can’t establish cause + effect (bc only measures pre-existing variables).
  • Can misinterpret correlation; if 2 variables related, 1 must cause the other - can’t determine this bc extraneous variable not held constant.
  • ‘3rd variable problem’; any other factor influencing correlation.
• 7. The Correlation Coefficient - Statistical Display

  This is the figure calculated after doing Spearman’s rank test.

  • Statistical tests used to enhance accuracy of correlations.
  • E.g. launching new drug, must be confident correlation correct = rigorous testing.
• 7a. The Correlation Coefficient - Statistical Display

  • Perfect Correlation = 1
  • Strong Correlation = 0.7, 0.8, 0.9
  • Moderate C = 0.4, 0.5, 0.6
  • Weak C = 0.1, 0.2, 0.3
  • No Correlation = 0
• 7b. The Correlation Coefficient - Statistical Display

  Sign tells direction of correlation:

  • Positive (+)
  • Negative (-)

  Number tells if correlation strong/mod/weak.

  • e.g. +0.8, strong positive.
  • e.g. -0.3, weak negative.

  Closer coefficient to +1/-1, stronger the relationship.

  Closer to zero, weaker the relationship.
• 8. Scattergraphs - Graphical Display of Correlation

  Positive Correlation - increase/increase.

  Negative Correlation - increase/decrease.

  No Correlation (variable ‘A’ no effect on variable ‘B’).
• 8a. Scattergraph - Graphical Display of Correlation

  Not all correlations linear; there’s still relationship but it’s not linear.

  • e.g. eating carbohydrates = more energy to run (positive), but eating too much means ability to run becomes less (negative).
• 8b. Scattergraph - Graphical Display of Correlation

  Line of Best Fit:

  • Line through scattergraph of data points.
  • Expresses relationship between points.
  • Concentration of data points around line of best fit = analyse if relationship strong/weak.