Categories: All - variables - distribution - relationship - samples

by nur aziera 9 years ago

2061

statistical test

Various statistical tests are utilized to analyze data depending on the nature of the samples and the distribution. Non-parametric tests like the Wilcoxon Signed Rank Test and Mann-Whitney Test are suitable for data that are not normally distributed and involve dependent and independent samples, respectively.

statistical test

STATISTICAL TESTS

WILCOXON SIGNED RANK TEST

2 dependent samples

KRUSKAL-WALLIS

Unpaired/independent samples

CHI SQUARE TEST

Characteristics
nominal data
displayed in contingency table
categorical variable
random sampling method
Only one sample

CORRELATION

Spearman's rank
not normally distributed
ordinal or interval variable
Pearson's product-moment
approximately normally distributed
interval variable
Relationship between variables
strength of relationship
positive or negative relationship

REGRESSION

Prediction
Equation
R2 - coefficient of determination
Y = a + bX
Models
one independent
one dependent

MANN-WHITNEY TEST

Non-parametric test
Not normally distributed
2 independent samples

TWO-WAY ANOVA

Same sample size
Samples normally distributed

ONE-WAY ANOVA

From normal distribution
Unpaired/independent sample
3 or more samples

DEPENDENT T-TEST

Same subjects
eg 2: marital status

single

after
before

INDEPENDENT T-TEST

Parametric test
Different subjects
eg 2: teeth location

maxillary

mandibular

eg 1: gender

female

male

Two conditions