Professional Documents
Culture Documents
1. Number of variables
2. Measurement level of variables (type)
3. Check for normality
4. Formulate Ho and Ha
5. SPSS
6. Interpret the outcomes (statistic, significance, reject or accept Ho?)
7. Give an explanation of the results in normal English
Techniques
Nominal
Ordinal
Description of variables
Show distribution
Frequency table
Graphic representation
Bar Charts
Means
Mode
Median
Deviation
Range & IQR
Subgroup
Cross tabulation
1-sample
2-independent samples
2-dependent samples
(2 variables of the same
sample)
Multiple independent
samples
2 variables
Multiple variables
Comparing means
Binomial
Binomial
Chi-Square test
Wilcoxon Rank Sum test
(Mann-Whitney)
Wilcoxon signed-rank
test
Kruskal-Wallis H-test
Relationship between variables
Crosstabulation & ChiSpearman rank
square test
correlation coefficient
Chi-square test over
subgroups
Interval/Ratio
Histogram
Mean
Deviation
Statistical measures per
subgroup
One-Samples t-test
Independent Samples ttest
Paired-Samples t-test
Paired-samples t-test
Use: Compare means of dependent samples (2
variables of the same sample)
Requirements: (otherwise use Wilcoxon Signed
Rank Test)
Normality OR more than 30 observations
Paired-samples meaning two observations
for each respondent in the sample
SPSS: Analyze -> Compare Means
ANOVA
Use: Comparing means for more than 2 groups,
looking at Within Groups and/or between Groups.
Requirements:
Interval or ratio.
Independent, randomly selected samples.
Normality
The k population variances are equal (or
equal size groups)
SPSS: Analyze -> Compare Means -> One-Way
ANOVA
Test statistic: F test
Post-Hoc tests: Bonferroni, Tukey, Scheffe
Use: Determines which means are statistically
different.
SPSS: Analyze -> Compare Means -> One-Way
ANOVA -> Post Hoc
Bivariate
Use: Check whether variables are correlated.
Requirements:
Ratio or Interval variables
Normality
SPSS: Analyze -> Correlate -> Bivariate
Kruskal-Wallis H-test
Use: Compare the means of not normally distributed
variables.
Requirements:
The p samples are random and independent
5 or more measurements per sample
Probability distributions samples drawn
from are continuous
SPSS: Analyze -> Nonparametric tests -> K
independent samples
Linear Regression
Use: investigate the effect of one or more independent variables to a dependent variable OR find a linear
equation that explains the relationship between two variables.
Requirements:
Ratio & Interval variables (one can work around this by creating dummy variables from ordinal or
nominal variables)
Assume linear relationship
SPSS: Analyze -> Regression -> Linear
Quality measures:
R-square: % of variance of the dependent variable explained by independent variables.
F-statistic in the ANOVA table: check whether there is a relationship between dependent variable
and independent variables
B in coefficient table: positive or negative relationship
T test: whether the relationship is significant
Multivariate linear regression
Use: Find a linear equation that explains the relationship between multiple variables
Requirements:
Interval/Ratio scale variables
SPSS: Analyze -> Regression ->
Quality measures: similar to the linear regression