One-Way Independent ANOVA: Notes

One-Way Independent ANOVA Overview

• Focus on the impact of superhero costumes on injury severity in children.
• Hypothesis: Flying superhero costumes result in more severe injuries than non-flying ones.

Data Summary

• Data collected from 30 children involving superhero costumes and injury severity (scale from 0= no injury to 100= death).
• Costumes analyzed:
  • Superman: Injuries recorded: 69, 32, 85, 66, 58, 52
  • Spiderman: Injuries recorded: 51, 31, 58, 20, 47, 37, 49, 40
  • Hulk: Injuries recorded: 26, 43, 10, 45, 30, 35, 53, 41
  • Ninja Turtle: Injuries recorded: 18, 18, 30, 30, 30, 41, 18, 25

Hypothesis Testing

• Hypothesis 1: Flying superheroes (Superman, Spiderman) lead to higher injuries than non-flying (Hulk, Ninja Turtle).
• Hypothesis 2: Injury severity ranking: Superman > Spiderman > Hulk > Ninja Turtle.

Contrast Generation Rules

1. Sensible Comparisons: Only compare two groups at a time.
2. Assign Weights: Positive weights for one set of contrasts and negative for the others.
3. Weight Sum: The total weight of each contrast must be zero.
4. Weight for Non-Including Groups: Assign a weight of 0 to groups not involved in the contrast.
5. Equal Weights: For one chunk of variance, the weights should equal the number of groups in the other chunk.

Example Weights Calculation for Contrasts

Effect Size Calculation - Cohen’s d

• Formula: $d = \frac{\bar{X<em>1} - \bar{X</em>2}}{s}$
• When comparing means:
  • Example of Superman vs. Ninja Turtle:
  • Superman mean: $60.33$, Ninja Turtle mean: $26.25$
  • Standard deviation (Control): $s = 8.16$
  • $d = \frac{60.33 - 26.25}{8.16} = 4.18$ (large effect)
• Cohen’s d interpretations:
  • 0.2: Small effect
  • 0.5: Medium effect
  • 0.8: Large effect

Running One-Way ANOVA in SPSS

1. Enter data into SPSS with columns for costumes and injury severity.
2. Define Dependent Variable (injury severity) & Independent Variable (costume).
3. Use related menu options for planned comparisons.
4. Specify appropriate contrasts weights as outlined above.
5. Run the analysis and check for homogeneity of variance using Levene’s test.

Post Hoc Analysis in SPSS

• Conducted if no prior hypotheses were made.
• If homogeneity assumed: use Gabriel's test/ Games-Howell procedure if not.
• Interpret p-values along with effect sizes.

Results Reporting

• Report F-ratio, degrees of freedom, and significance levels.
• Example:
  • Results: F(3, 26) = 8.32, p < 0.001 (Significant impact of costume on injury severity).

Visual Representation

• Utilize error bar charts for visual representation of means and confidence intervals.
• Check overlap between confidence intervals to assess significance of group differences.

Summary and Implications

• Higher injury severity associated with certain superhero costumes.
• Educational implications about children's perception of superheroes and injury risks.
• Understand the importance of using statistical software for data analysis.

Self-Test and Practical Applications

• Compute Cohen’s d for various contrasts (Superman vs. Hulk & Ninja Turtle costumes).
• Conduct additional analyses and report findings appropriately as per guidelines.