Scribd
Upload
22 views

Steps in Doing Analysis

The document provides guidance on selecting appropriate statistical tests based on the characteristics of data. It distinguishes between nominal, ordinal and scale data types. Parametric tests like t-tests and ANOVA are used for scale data that meets assumptions of normality and equal variances, while non-parametric tests like chi-square, median tests, and Mann-Whitney U tests are used when those assumptions are not met. Specific statistical tests are listed for comparing groups for nominal, ordinal and scale data.

Uploaded by

Yong Leong Chai
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
22 views

Steps in Doing Analysis

The document provides guidance on selecting appropriate statistical tests based on the characteristics of data. It distinguishes between nominal, ordinal and scale data types. Parametric tests like t-tests and ANOVA are used for scale data that meets assumptions of normality and equal variances, while non-parametric tests like chi-square, median tests, and Mann-Whitney U tests are used when those assumptions are not met. Specific statistical tests are listed for comparing groups for nominal, ordinal and scale data.

Uploaded by

Yong Leong Chai
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 4

Step 1: identify the types of the data and label it: nominal, ordinal or scale.

types of data (measures)

2 levels

3 or 4 levels

5 or more than 5 levels

normality: skewness between -1 and +1 nominal

ordinal

yes

no

normal: scale

not normal: ordinal

n< 50: Shapiro-wilk test Significance value: >0.05 n> 50: kolmogorov-smirnov significance value: >0.05

normally distributed

variance of the comparison group are equal or at least the variances ratio is known.

parametric test

interval or ratio (scale measure in SPSS)

selection of participants is independent

t-test parametric test ANOVA Pearson product correlation

chi-square test
significance test median test mann-whitney U test non-parametric test wilcoxon signed rank test kruskal-wallis test freidman test spearman rho

i.

Chi-square test: difference in proportions in two or more groups.

i. Median test: Difference of the medians of two independent groups. ii. Mann-whitney test: difference of the medians of two independent groups. iii. Wilcoxon signed rank test: difference in the ranks of two related groups iv. Kruskal-wallis test: difference in the ranks of three or more independent groups. v. Freidman test: difference in the ranks of three or more dependent groups.

Nominal (independent)

Ordinal (dependent)

i.

ii.

iii. iv.

T-test for independent samples: difference between means of two independent groups. T-test for dependent samples: difference between means of two dependent groups. Anova: difference among three or more independent groups Manova (factorial design): investigates two or more independent variables and the interactions between them.

i.

Spearman rho - Test whether a correlation is different from zero.

Scale (dependent)

i.

ii.

Pearson product correlation test whether a correlation is different from zero (a relationship exists) Multiple regression: - Combines variables that are known individually to predict (correlate with) the criterion into a multiple regression equation.

You might also like