Chapter 13 Two Groups Too Many? Try Analysis of Variance (ANOVA) Part IV Significantly Different: Using Inferential Statistics
Chapter 13 Two Groups Too Many?
Try Analysis of Variance (ANOVA)
Part IVSignificantly Different:
Using Inferential Statistics
What you will learn in Chapter 13
What Analysis of Variance (ANOVA) is and when it is appropriate to use
How to compute the F statistic
How to interpret the F statistic
How to use SPSS to conduct an ANOVAsingle factor design
Analysis of Variance (ANOVA)
Used when more than two group means are being tested simultaneouslyGroup means differ from one another on a
particular score / variableExample: DV = GRE Scores & IV = Ethnicity
Test statistic = F testR.A. Fisher, creator
Path to Wisdom & KnowledgeHow do I know if ANOVA is the right test?
Different Flavors of ANOVA
ANOVA examines the variance between groups and the variances within groupsThese variances are then compared against
each otherSimilar to the t Test…only in this case you
have more than two groupsOne-way ANOVA
Simple ANOVASingle factor (grouping variable)
More Complicated ANOVAFactorial Design
More than one treatment/factor examinedMultiple Independent Variables
One Dependent VariableExample – 3x2 factorial design
Number of Hours in Preschool
Gender
Male5 hours per week
10 hours per week
20 hours per week
Female 5 hours per week
10 hours per week
20 hours per week
Computing the F Statistic
Rationale…want the within group variance to be small and the between group variance to be large in order to find significance.
Hypotheses
Null hypothesis
Research hypothesis
Source Table
Source SS df MS F
Between 1,133.07 27 566.54 8.799
Within 1,738.40 29 64.39
Note: F value for two group is the same as t2
Degrees of Freedom (df)
NumeratorNumber of groups minus onek-13 groups --- 3 – 1 = 2
DenominatorTotal number of observations minus the number
of groupsN-1100 participants --- 30 – 3 = 97
Represented: F (2, 27)
How to Interpret
F (2,27) = 8.80, p < .05F = test statistic 2,27 = df between groups & df within groups
{Ah ha…3 groups and 30 total scores examined}8.80 = obtained value
Which we compared to the critical valuep < .05 = probability less than 5% that the
null hypothesis is trueMeaning the obtained value is GREATER than the
critical value
Omnibus Test
The F test is an “omnibus test” and only tells you that a difference exists
Must conduct follow-up t tests to find out where the difference is…BUT…Type I error increases with every follow-
up test / possible comparison made1 – (1 – alpha)k
Where k = number of possible comparisons
Using the ComputerSPSS and the One-Way ANOVA
What does it all mean?
SPSS Output
Post Hoc Comparison
Glossary Terms to Know
Analysis of varianceSimple ANOVAOne-way ANOVAFactorial design
Omnibus testPost Hoc comparisonsSource table
Chapter 17 What to Do When You’re Not Normal:
Chi-Square and Some Other Nonparametric Tests
Part IVSignificantly Different:
Using Inferential Statistics
What you will learn in Chapter 17
A brief survey of nonparametric statisticsWhen they should be usedHow they should be used
Introduction
Parametric statistics have certain assumptionsVariances of each group are similarSample is large enough to represent the
populationNonparametric statistics don’t require the
same assumptionsAllow data that comes in frequencies to be
analyzed…they are “distribution free”
One-Sample Chi-Square
Chi-square allows you to determine if what you observe in a distribution of frequencies is what you would expect to occur by chance.One-sample chi-square (goodness of fit test)
only has one dimensionTwo-sample chi-square has two dimensions
Computing Chi-Square
What do those symbols mean?
22 (O E)
Ex
More Hypotheses
Null hypothesis
H0: P1 = P2 = P3
Research hypothesis
H1: P1 P2 P3
Computing Chi Square
Category O E D (O-E)2 (O-E)2/2
For 23 30 7 49 1.63
Maybe 17 30 13 169 5.63
Against 50 30 20 400 13.33
Total 90 90 C2 = 20.6
So How Do I Interpret…
x2(2) = 20.6, p < .05
x2 represents the test statistic2 is the number of degrees of freedom20.6 is the obtained valuep < .05 is the probability
Using the ComputerOne-Sample Chi Square using SPSS
SPSS Output
What does it all mean?
Other Nonparametric Tests
Glossary Terms to Know
ParametricNonparametricOne-sample Chi Square