Is it true that Nonparametric tests are always less powerful than parametric tests.?

Nonparametric tests can be more powerful than parametric tests when the assumptions of parametric tests are severely violated.

Is it true that You can only use nonparametric tests for ordinal data.?

Nonparametric tests are also used for nominal data and can be used for scale data if parametric assumptions are not met or sample size is very small.

1-DV Data Type Cheat Sheet

This cheat sheet outlines how to choose the appropriate statistical test for a single dependent variable (1-DV) based on the measurement scale of the IV and the research question (differences vs. associations).

Core Principles

Identify the measurement scale of your independent variable (IV): Nominal, Ordinal, or Scale (Interval/Ratio).
Determine if your research question aims to find differences between groups or associations between variables.
Parametric tests are generally used for Scale data, while Nonparametric tests are used for Ordinal and Nominal data.
Nonparametric tests can be used for Scale data if assumptions are violated or sample size is very small.
The number of samples/groups involved (one vs. two) is crucial for test selection.

Action Steps

Step 1: Identify your dependent variable (DV) and confirm it's a single DV.
Step 2: Determine the measurement scale of your independent variable (IV): Nominal, Ordinal, or Scale (Interval/Ratio).
Step 3: Identify your research question: Are you looking for differences between groups or associations between variables?
Step 4: Count the number of samples/groups involved in your IV (one sample or two samples/groups).
Step 5: Use the chart to select the appropriate statistical test based on steps 1-4.

Key Terms

1-DV: One Dependent Variable: The outcome variable being measured.
IV: Independent Variable: The variable manipulated or categorized to see its effect on the DV.
Parametric Test: Statistical tests that make assumptions about the distribution of the population (e.g., normality).
Nonparametric Test: Statistical tests that do not assume a specific population distribution; often used for nominal or ordinal data.
Scale Data: Interval or Ratio data, where differences between values are meaningful and there's a true zero point (for Ratio).
Ordinal Data: Data that can be ranked or ordered, but the differences between ranks are not necessarily equal.
Nominal Data: Categorical data where values have no inherent order (e.g., gender, color).
Between-Subjects Design: Different participants are in different groups (e.g., control vs. experimental).
Within-Subjects Design: The same participants are measured multiple times (e.g., before and after treatment).

Pro Tips

Always check the assumptions of parametric tests (e.g., normality, homogeneity of variance) before using them.
If assumptions are violated, switch to the corresponding nonparametric test.
Partial correlation is used when you want to examine the relationship between two variables while controlling for a third.
The Chi-Square test is specifically for nominal data when looking for associations.

Pitfalls to Avoid

Using a parametric test when data is nominal or ordinal.
Confusing 'differences' with 'associations'.
Applying tests for two samples when you only have one.
Ignoring the assumptions of parametric tests.
Using the wrong test for nominal data associations (e.g., t-test instead of Chi-Square).

Myth vs Reality

Nonparametric tests are always less powerful than parametric tests.: Nonparametric tests can be more powerful than parametric tests when the assumptions of parametric tests are severely violated.
You can only use nonparametric tests for ordinal data.: Nonparametric tests are also used for nominal data and can be used for scale data if parametric assumptions are not met or sample size is very small.

Real World Examples

Comparing anxiety levels between male and female students.: Independent-samples t-test (if anxiety is scale) or Mann-Whitney U test (if anxiety is ordinal).
Examining the relationship between hours studied and exam score.: Pearson's bivariate correlation (if both are scale).
Assessing if voting intention (Democrat, Republican, Independent) differs based on support for foreign aid (Yes, No).: Chi-square test of contingencies (Nominal x Nominal).
Measuring changes in blood pressure before and after taking a new medication.: Paired-samples t-test (if BP is scale) or Wilcoxon Signed Rank test (if BP is ordinal).