Analysis of Paired Data

• Definition of Paired Data

  • Paired or dependent samples come from the same individual or matched individuals.

  • Examples include:

    • Measurements before and after a treatment/intervention (e.g., blood pressure before and after medication).

    • Comparing measurements within one individual (e.g., left vs. right arm length or eye color).

• Characteristics of Paired Data

  • Expected similarities in measurements (e.g., blood pressure, arm length) when no treatment effect is present.

  • Differences can signal an intervention's effectiveness or natural variances in paired measurements.

• Determining Pairing

  • Example: Height comparison between spouses.

    • There may be no strong correlation between male and female heights, suggesting the data may not be paired.

  • Example: Dietary comparison within couple (e.g., fat intake).

    • If couples share similar dietary habits, data can be considered paired.

• Statistical Advantages of Paired Data

  • Reduces variability by allowing individuals to serve as their own controls.

  • Example: Blood pressure before and after treatment.

    • Variability between individuals is minimized, enhancing clarity of treatment effects