Help shape GoConqr. What should we add next?

Mind map

stats week 5

Abbey Wrigley Statistics Degree English

Map content

stats week 5

  1. ANOVA with. more than one IV

    • 2 way indpendent ANOVA, 3 way mixed ANOVA, etc - number denotes number of IVS
      • factor/IV interchangeable
        • 2*2 ANOVA - 2 IVs with 2 levels each - 4*2*2 = 3 IVs, one with 4 levels, 2 with 2 levels
    • example - Men and women like chocolate Questions: Is there a gender difference in enjoyment of chocolate? Does enjoyment of chocolate differ according to texture? Are gender differences in enjoyment of chocolate dependent on the texture? - IV 1 = gender differences, IV 2 = texture differences
      • gender effect and texture effect = main effects - gender effects dependent on texture = interaction
    • first IV is main IV, second is secondary IV
    • Cell means = individual mean for each condition, marginal means = mean scores for single IV levels (e.g overall score for women)
  2. variance

    • variance between IV levels = IV 1, IV 2, interaction (gender*texture) = variance within IV levels - error
      • F = MSm / MSr
        • F gender = variance due to manipulation of gender / variance due to error alone
          • F texture = variance due to manipulation of texture / variance due to error alone
            • F gender * texture = variance due to gender * texture / variance due to error alone
  3. logic

    • calculate F for each effect (2 main effects and interaction) - determine whether we can reject null for each
    • interaction - the combined effects of multiple IVs on the DV - significant interaction indicates that the effect of manipulating one IV depends on the level of the other IV
  4. post hoc

    • only relevant when IV has more than 2 levels and when main effect of IV is significant
    • eta2 = proportion of total variance attributable to the factor - partial eta 2 = proportion of total variance attributable to the factor, partialling out variance due to other factors
  5. Simple effects

    • the presence of an interaction suggests we need to consider differences at the level of cell means (simple effects) - simple effects = the effect of an IV at a single level of another IV
    • conduct t-tests between individual cell means
      • need to correct for multiple comparisons as SPSS doesn't do it for us - Boferroni correction - dividual require alpha level by no. of comparison e.g for 4 comparisons, 0.05/4 = 0.017

Information

Description
Degree Statistics Mind Map on stats week 5, created by Abbey Wrigley on 23/01/2019.
Created by
Abbey Wrigley
Views
0
Rating

More on Factorial ANOVA and Interactions

Feedback

What is it about?

What should we add next?