Presentation is loading. Please wait.

Presentation is loading. Please wait.

Effect Size Estimation in Fixed Factors Between- Groups Anova.

Similar presentations


Presentation on theme: "Effect Size Estimation in Fixed Factors Between- Groups Anova."— Presentation transcript:

1 Effect Size Estimation in Fixed Factors Between- Groups Anova

2 Contrast Review  Concerns design with a single factor A with a 2 levels (conditions) The omnibus comparison concerns all levels (i.e., df A > 2) A focused comparison or contrast concerns just two levels (i.e.,df = 1)  The omnibus effect is often relatively uninteresting compared with specific contrasts (e.g., treatment 1 vs. placebo control)  A large omnibus effect can also be misleading if due to a single discrepant mean that is not of substantive interest

3 Comparing Groups  Traditional approach is to analyze the omnibus effect followed by analysis of all possible pairwise contrasts (i.e. compare each condition to every other condition)  However, this approach is typically incorrect (Wilkinson & TFSI,1999)—for example, it is rare that all such contrasts are interesting Also, use of traditional methods for post hoc comparisons (e.g. Newman-Keuls) reduces power for every contrast, and power may already be low

4 Contrast specification and tests  A contrast is a directional effect that corresponds to a particular facet of the omnibus effect often represented with the symbol y for a population or yˆ for a sample a weighted sum of means  In a sample, a contrast is calculated as:  a 1, a 2,..., a j is the set of weights that specifies the contrast  As we have mentioned Contrast weights must sum to zero and weights for at least two different means should not equal zero Means assigned a weight of zero are excluded from the contrast Means with positive weights are compared with means given negative weights

5 Contrast specification and tests  For effect size estimation with the d family, we generally want a standard set of contrast weights  In a one-way design, the sum of the absolute values of the weights in a standard set equals two (i.e., ∑| a j | = 2.0)  Mean difference scaling permits the interpretation of a contrast as the difference between the averages of two subsets of means

6 Contrast specification and tests  An exception to the need for mean difference scaling is for trends (polynomials) specified for a quantitative factor (e.g., drug dosage)  There are default sets of weights that define trend components (e.g. linear, quadratic, etc.) that are not typically based on mean difference scaling  Not usually a problem because effect size for trends is generally estimated with the r family (measures of association)  Measures of association for contrasts of any kind generally correct for the scale of the contrast weights

7 Orthogonal Contrasts  Two contrasts are orthogonal if they each reflect an independent aspect of the omnibus effect  For balanced designs and unbalanced designs (latter)

8 Orthogonal Contrasts  The maximum number of orthogonal contrasts is df A = a − 1  For a set of all possible orthogonal pairwise contrasts, the SS A = the total SS from the contrasts, and their eta-squares will sum to the SS A eta-square  That is, the omnibus effect can be broken down into a − 1 independent directional effects  However, it is more important to analyze contrasts of substantive interest even if they are not orthogonal

9 Contrast specification and tests  t-test for a contrast against the nil hypothesis  The F is

10 Dependent Means  Test statistics for dependent mean contrasts usually have error terms based on only the two conditions compared— for example:  s 2 here refers to the variance of the contrast difference scores  This error terms do not assume sphericity, which we’ll talk about more with repeated measures design

11 Confidence Intervals  Approximate confidence intervals for contrasts are generally fine  The general form of an individual confidence interval for Ψ is:  df error is specific to that contrast

12 Contrast specification and tests  There are also corrected confidence intervals for contrasts that adjust for multiple comparisons (i.e., inflated Type I error) Known as simultaneous or joint confidence intervals  Their widths are generally wider compared with individual confidence intervals because they are based on a more conservative critical value  Program for correcting http://www.psy.unsw.edu. au/research/resources/psy program.html http://www.psy.unsw.edu. au/research/resources/psy program.html

13 Standardized contrasts  The general form for standardized contrasts (in terms of population parameters)

14 Standardized contrasts  There are three general ways to estimate σ (i.e., the standardizer) for contrasts between independent means:  1. Calculate d as Glass’s Δ i.e., use the standard deviation of the control group  2. Calculate d as Hedge’s g i.e., use the square root of the pooled within-conditions variance for just the two groups being compared  3. Calculate d as an extension of g Where the standardizer is the square root of MS W based on all groups Generally recommended

15 Standardized contrasts  Calculate from a t contrast for a paper not reporting effect size like they should  Recall the weights should sum to 2

16 CIs  Once the d is calculated one can easily obtain exact confidence intervals via the MBESS package in R or Steiger’s standalone program The latter will provide the interval for the noncentrality parameter which must then be converted to d

17 Cohen’s f  Cohen’s f provides what can interpreted as the average standardized mean difference across the groups in question  It has a direct relation to a measure of association  As with Cohen’s d, there are guidelines regarding Cohen’s f .10,.25,.40 for small, moderate and large effect sizes  These correspond to eta-square values of:.01,.06,.14  Again though, one should conduct the relevant literature for effect size estimation

18 Measures of Association  A measure of association describes the amount of the covariation between the independent and dependent variables  It is expressed in an unsquared metric or a squared metric—the former is usually a correlation, the latter a variance-accounted-for effect size  A squared multiple correlation (R 2 ) calculated in ANOVA is called the correlation ratio or estimated eta-squared ( 2 )

19 Eta-squared  A measure of the degree to which variability among observations can be attributed to conditions  Example:  2 =.50 50% of the variability seen in the scores is due to the independent variable.

20 More than One factor  It is a fairly common practice to calculate eta 2 (correlation ratio) for the omnibus effect but to calculate the partial correlation ratio for each contrast  As we have noted before SPSS calls everything partial eta-squared in it’s output, but for a one-way design you’d report it as eta-squared

21 Problem  Eta-squared (since it is R-squared) is an upwardly biased measure of association (just like R-squared was)  As such it is better used descriptively than inferentially

22 Omega-squared  Another effect size measure that is less biased and interpreted in the same way as eta-squared  So why do we not see omega-squared so much?  People don’t like small values  Stat packages don’t provide it by default

23 Omega-squared  Put differently

24 Omega-squared  Assumes a balanced design eta 2 does not assume a balanced design  Though the omega values are generally lower than those of the corresponding correlation ratios for the same data, their values converge as the sample size increases  Note that the values can be negative—if so, interpret as though the value were zero

25 Comparing effect size measures  Consider our previous example with item difficulty and arousal regarding performance

26 Comparing effect size measures 22 ω2ω2 Partial  2 f B/t groups.67.59.671.42 Difficulty.33.32.50.70 Arousal.17.14.33.45 Interactio n.17.14.33.45 Slight differences due to rounding, f based on eta-squared

27 No p-values  As before, programs are available to calculate confidence intervals for an effect size measure  Example using the MBESS package for the overall effect 95% CI on ω 2 : .20 to.69

28 No p-values  Ask yourself as we have before, if the null hypothesis is true, what would our effect size be (standardized mean difference or proportion of variance accounted for)?  0  Rather than do traditional hypothesis testing, one can simply see if our CI for the effect size contains the value of zero (or, in eta-squared case, gets really close)  If not, reject H 0  This is superior in that we can use the NHST approach, get a confidence interval reflecting the precision of our estimates, focus on effect size, and de-emphasize the p-value


Download ppt "Effect Size Estimation in Fixed Factors Between- Groups Anova."

Similar presentations


Ads by Google