12.01.2015 Views

RESEARCH METHOD COHEN ok

RESEARCH METHOD COHEN ok

RESEARCH METHOD COHEN ok

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

588 CHOOSING A STATISTICAL TEST<br />

Box 26.3<br />

Types of statistical tests for four scales of data<br />

Nominal Ordinal Interval and ratio<br />

Measures of<br />

association<br />

Tetrachoric<br />

correlation<br />

Point biserial<br />

correlation<br />

Phi coefficient<br />

Cramer’s V<br />

Spearman’s rho<br />

Kendall rank order<br />

correlation<br />

Kendall partial rank<br />

correlation<br />

Pearson product-moment<br />

correlation<br />

Measures of<br />

difference<br />

Chi-square Mann-Whitney U test t-test for two independent<br />

samples<br />

McNemar Kruskal-Wallis t-test for two related<br />

samples<br />

Cochran Q Wilcoxon matched pairs One-way ANOVA<br />

Binomial test<br />

Friedman two-way analysis of Two-way ANOVA for more<br />

variance<br />

Wald-Wolfowitz test Tukey hsd test<br />

Kolmogorov-Smirnov test Scheffé test<br />

Measures of linear<br />

relationship<br />

between<br />

independent and<br />

dependent<br />

variables<br />

Identifying<br />

underlying factors,<br />

data reduction<br />

Ordinal regression analysis<br />

Linear regression<br />

Multiple regression<br />

Factor analysis<br />

Elementary linkage analysis<br />

(e.g. parents, teachers, students and administrative<br />

staff in a school). Tests which can be applied<br />

to a single group include the binomial test, the<br />

chi-square one-sample test, and the Kolmogorov-<br />

Smirnov one-sample test; tests which can be<br />

applied to two groups include the chi-square test,<br />

Mann-Whitney U test, the t-test, the Spearman<br />

and Pearson tests of correlation; tests which can<br />

be applied to three or more samples include the<br />

chi-square test, analysis of variance and the Tukey<br />

test. We set out some of these tests in Box 26.2. It<br />

is essential to use the correct test for the correct<br />

number of groups.<br />

The statistical tests to be used also depend on<br />

the scales of data being treated (nominal – ratio)<br />

and the tasks which the researcher wishes to<br />

perform – the purpose of the analysis (e.g. to<br />

discover differences between groups, to lo<strong>ok</strong><br />

for degrees of association, to measure the<br />

effect of one or more independent variables<br />

on a dependent variable etc.). In preceding<br />

chapters we have described the different scales<br />

of data and the kinds of tests available<br />

for different purposes. In respect of these<br />

considerations, Box 26.3 summarizes some of the<br />

main tests here.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!