Statistics (Normality & Test of Difference)

59 Questions

What is the concept of 'psychological traits of states can be quantified and measured' referring to?

The assumption that psychological traits can be numerically evaluated

What does the term 'cut score' refer to in test scoring?

A reference point for classification

What is meant by psychometric soundness?

Technical quality of measurement techniques

Which assumption involves no theory and testing only based on observable data?

A posteriori

What does the assumption related to error sources imply about the assessment process?

Multiple sources of error are present in the assessment process

What does the discrimination index measure?

The item differentiates between different levels of performance

Which post hoc test should be used if ANOVA indicates significant results and normality is assumed?

Tukey's Test

Which of the following ensures the prevention of general familiarity with test content?

Controlling the use of tests

What is the appropriate test to use for comparing two groups with normal data and a Levene's test result of (Significant) <.05?

Welch's Test

If comparing three groups and Levene's test result is significant <.05 but the data is not normal, which test should be used?

Kruskal-Wallis Test

For three groups with normal data and a non-significant result from Levene's test, what post-hoc analysis should be performed?

Tukey's

Which test is appropriate for comparing two groups with non-normal data and a non-significant Levene's test?

Mann-Whitney U Test

If comparing three groups with non-normal data and Levene's test is non-significant, what post-hoc analysis is appropriate?

Dunn's

Match the following conditions for comparing 2 groups with the appropriate test:

YES and (S) <.05 = Welch's Test NOT and (NS) > .05 = Unpaired Independent t-test YES and (NS) >.05 = Mann-Whitney U Test

Match the following normal data conditions and Levene's test results with the appropriate test for comparing 3 groups:

YES and < .05 = Welch's Test NO and < .05 = Kruskal-Wallis Test YES and > .05 = ANOVA NO and > .05 = Kruskal-Wallis Test

Match the following appropriate test for comparing 3 groups with their corresponding normal data and Levene's test results:

Welch's Test = YES and < .05 Kruskal-Wallis Test = NO and > .05 ANOVA = YES and > .05

Match the following appropriate test for comparing 2 groups with their corresponding normal data and Levene's test results:

Mann-Whitney U Test = NOT and (NS) > .05 Unpaired Independent t-test = YES and (NS) > .05 Welch's Test = YES and (S) < .05

Match the following appropriate tests for comparing 3 groups to their respective post hoc analysis:

Welch's Test = Games-Howell Kruskal-Wallis Test = Dunn's ANOVA = Tukey's

Which test should be used if comparing three groups with normal data and Levene's test is non-significant?

Parametric (ANOVA)

Which test is used for comparing three groups when data is not normal and Levene's test is significant?

Kruskal-Wallis

If Levene's test is significant, which test should you consider for comparing three groups with normal data?

Welch's Test

Which test is used if comparing two groups with non-normal data and a significant Levene's test?

Mann-Whitney U

Which test is appropriate for comparing three groups with non-normal data and a non-significant Levene's test?

Kruskal-Wallis

If comparing three groups with non-normal data and Levene's test is significant, the appropriate test to use is the Kruskal-Wallis Test.

True

When comparing two groups with normal data and a non-significant Levene's test result, the appropriate test is Welch's Test.

False

If comparing three groups with normal data and Levene's test is significant, the appropriate test to use is ANOVA.

False

For comparing two groups with non-normal data and a non-significant Levene's test result, the appropriate test is the Mann-Whitney U Test.

True

If comparing three groups with normal data and a non-significant Levene's test result, the post-hoc analysis to use is Dunn's.

False

Which method uses a more objective-numerical approach to assess normality?

Shapiro Wilk

In Theory-Driven methods, what indicates that data is more normal in a QQ Plot?

More circle in line

According to the Shapiro Wilk test, what p-value indicates normality?

Greater than or equal to 0.05

Which descriptive-graphical method is classified as less objective?

Histogram

In a sample size of 60 participants, what approach is best to check for normality?

Shapiro Wilk

Which statistical test is appropriate for comparing two scores within the same group if the data is normally distributed?

Paired T-Test

For unpaired samples with two groups where the data is normally distributed, which test should be used?

Unpaired T-Test

If the data is not normally distributed and involves one group with two scores for each member, which of the following tests should be used?

Wilcoxon Signed Rank Test

When comparing two or more groups with continuous data that is not normally distributed, which test should be employed?

Mann-Whitney U Test

Which test is suited for data from a single group with two scores per member if the data is categorical but not normally distributed?

Wilcoxon Signed Rank Test

Which of the following is an appropriate test for comparing two independent groups with normally distributed data?

Unpaired T-Test

The Mann-Whitney U Test is used for unpaired samples with two groups if the data is not normally distributed.

True

A paired T-Test is used for one group with two scores per member if the data is not normally distributed.

False

For comparing two or more groups with normally distributed continuous data, an Unpaired T-Test should be employed.

True

Wilcoxon Signed Rank Test is used for paired samples if the data is normally distributed.

False

The appropriate test for comparing two scores within the same group, when the data is normally distributed, is the Paired T-Test.

True

Which test is appropriate for comparing two dependent samples with non-parametric data?

Wilcoxon Signed Rank Test

What non-parametric test should be used for comparing three dependent samples?

Friedman

If you are comparing the means of two independent groups with normally distributed data, which test should you use?

Unpaired T-Test

Which test is suitable for assessing differences between more than two independent groups with non-parametric data?

Kruskal-Wallis

Which parametric test should be used for comparing three or more dependent samples?

Repeated measures of ANOVA

Which test would you use to compare two dependent samples with normally distributed data?

Paired T-Test

When comparing the means of more than two independent groups with normally distributed data, which test is appropriate?

One Way ANOVA

Which of the following tests is used for comparing two unpaired samples when the data is not normally distributed?

Mann-Whitney U

Which test is ideal for comparing three or more independent groups with normally distributed data?

One Way ANOVA

For a non-parametric test involving more than two groups, which test would be suitable?

Kruskal-Wallis

Which parametric test is appropriate for comparing three dependent groups?

Repeated ANOVA

Which test should be used to compare three groups with non-parametric data?

Kruskal-Wallis test

For a single group with three different scores measured, which parametric test is appropriate?

Repeated ANOVA

If you have three groups and one score for each group with parametric data, which test should be used?

One-way ANOVA

Which test would be appropriate for comparing three dependent samples with non-parametric data?

Friedman test

Study Notes

Assumptions in Testing and Scoring

Psychological traits of states exist and can be quantified and measured
Test-related behavior predicts non-test related behavior
Test and other measurement techniques have strengths and weaknesses
Various sources of error are part of the assessment process
Testing and assessment can be conducted in a fair or unbiased manner

Test Scoring and Interpretation

Test scoring involves assigning evaluative codes to divide a set of data into two or more classifications
A cut score is a reference point, usually numerical, used to make classifications
Psychometric soundness refers to the technical quality of a test or assessment

Assumptions in Test Development

A posteriori approach: no theory-based test development
A priori approach: theory-based test development

Discrimination Index

Measures whether an item is biased or not

Assumptions in Testing and Scoring

Psychological traits of states exist and can be quantified and measured
Test-related behavior predicts non-test related behavior
Test and other measurement techniques have strengths and weaknesses
Various sources of error are part of the assessment process
Testing and assessment can be conducted in a fair or unbiased manner

Test Scoring and Interpretation

Test scoring involves assigning evaluative codes to divide a set of data into two or more classifications
A cut score is a reference point, usually numerical, used to make classifications
Psychometric soundness refers to the technical quality of a test or assessment

Assumptions in Test Development

A posteriori approach: no theory-based test development
A priori approach: theory-based test development

Discrimination Index

Measures whether an item is biased or not

Normality

When ANOVA assumptions are not met, alternative tests can be used, such as Welch's ANOVA.
Post-hoc tests for Welch's ANOVA include Tukey's Test and Games-Howell post hoc to compare groups.

Levene's Test

Used to test for homogeneity of variance among 2 or more groups.
Essential step before performing parametric tests like ANOVA.
If Levene's Test indicates non-normality, non-parametric alternatives like Kruskal Wallis or Welch's Test can be used.

Controlling the Use of Psych Tests

Ensuring tests are used by qualified professionals to maintain test integrity.
Preventing general familiarity with test content, which could invalidate test results.

Comparing 2 Groups

When comparing two groups with normal data, a significant Levene's Test (p < 0.05) indicates using Welch's Test.
If Levene's Test is not significant (p > 0.05), use the Unpaired Independent t-test if the data is not normal, or the Mann-Whitney U Test if the data is normal.

Comparing 3 Groups

With normal data and a significant Levene's Test (p < 0.05), use Welch's Test and Games-Howell for post-hoc analysis.
If the data is not normal and Levene's Test is significant (p < 0.05), use the Kruskal-Wallis Test and Dunn's for post-hoc analysis.
When the data is normal and Levene's Test is not significant (p > 0.05), use ANOVA and Tukey's for post-hoc analysis.
If the data is not normal and Levene's Test is not significant (p > 0.05), use the Kruskal-Wallis Test and Dunn's for post-hoc analysis.

Assumptions in Testing and Scoring

Psychological traits of states exist and can be quantified and measured
Test-related behavior predicts non-test related behavior
Test and other measurement techniques have strengths and weaknesses
Various sources of error are part of the assessment process
Testing and assessment can be conducted in a fair or unbiased manner

Test Scoring and Interpretation

Test scoring involves assigning evaluative codes to divide a set of data into two or more classifications
A cut score is a reference point, usually numerical, used to make classifications
Psychometric soundness refers to the technical quality of a test or assessment

Assumptions in Test Development

A posteriori approach: no theory-based test development
A priori approach: theory-based test development

Discrimination Index

Measures whether an item is biased or not

Comparing 2 Groups

When comparing two groups with normal data, a significant Levene's Test (p < 0.05) indicates using Welch's Test.
If Levene's Test is not significant (p > 0.05), use the Unpaired Independent t-test if the data is not normal, or the Mann-Whitney U Test if the data is normal.

Comparing 3 Groups

With normal data and a significant Levene's Test (p < 0.05), use Welch's Test and Games-Howell for post-hoc analysis.
If the data is not normal and Levene's Test is significant (p < 0.05), use the Kruskal-Wallis Test and Dunn's for post-hoc analysis.
When the data is normal and Levene's Test is not significant (p > 0.05), use ANOVA and Tukey's for post-hoc analysis.
If the data is not normal and Levene's Test is not significant (p > 0.05), use the Kruskal-Wallis Test and Dunn's for post-hoc analysis.

Normality and Statistical Tests

Levene's Test is used for one group of data.
Normality is a crucial assumption for certain statistical tests.
Parametric tests (e.g., ANOVA) require normality, while non-parametric tests (e.g., Kruskal-Wallis) do not.
When normality is not met, alternative tests can be used:
- Welch's Test can be used as an alternative to ANOVA.
- Kruskal-Wallis Test can be used as a non-parametric alternative to ANOVA.

Comparing 2 Groups

When comparing two groups with normal data, a significant Levene's Test (p < 0.05) indicates using Welch's Test.
If Levene's Test is not significant (p > 0.05), use the Unpaired Independent t-test if the data is not normal, or the Mann-Whitney U Test if the data is normal.

Comparing 3 Groups

With normal data and a significant Levene's Test (p < 0.05), use Welch's Test and Games-Howell for post-hoc analysis.
If the data is not normal and Levene's Test is significant (p < 0.05), use the Kruskal-Wallis Test and Dunn's for post-hoc analysis.
When the data is normal and Levene's Test is not significant (p > 0.05), use ANOVA and Tukey's for post-hoc analysis.
If the data is not normal and Levene's Test is not significant (p > 0.05), use the Kruskal-Wallis Test and Dunn's for post-hoc analysis.

Methods of Normality

Methods are categorized into two types: less objective and theory-driven

Less Objective Methods

Graphical method: Histogram is used to check for normality
Numerical method: Skewness and Kurtosis are used to check for normality

Theory-Driven Methods

Graphical method: QQ Plots are used to check for normality
In QQ Plots, the more circular the points are in a line, the more normal the data is
Numerical method: Shapiro-Wilk test is used to check for normality
In Shapiro-Wilk test, a significance level of 0.05 or less indicates normal data

Statistical Tests for Comparing Data

Paired-Samples Tests (Dependent Measures)

One group with two scores for each member
Data can be continuous or categorical
Appropriate tests: Paired T-Test or Wilcoxon Signed Rank Test
Paired T-Test: used for normally distributed data (interval/ratio)
Wilcoxon Signed Rank Test: used for non-normally distributed data (ordinal)

Unpaired-Samples Tests (Independent Measures)

Two or more groups with one score for each member of each group
Data can be continuous or categorical
Appropriate tests: Unpaired T-Test or Mann-Whitney U Test
Unpaired T-Test: used for normally distributed data (interval/ratio)
Mann-Whitney U Test: used for non-normally distributed data (ordinal)

Statistical Tests for Comparing Data

Paired-Samples Tests (Dependent Measures)

One group with two scores for each member
Data can be continuous or categorical
Appropriate tests: Paired T-Test or Wilcoxon Signed Rank Test
Paired T-Test: used for normally distributed data (interval/ratio)
Wilcoxon Signed Rank Test: used for non-normally distributed data (ordinal)

Unpaired-Samples Tests (Independent Measures)

Two or more groups with one score for each member of each group
Data can be continuous or categorical
Appropriate tests: Unpaired T-Test or Mann-Whitney U Test
Unpaired T-Test: used for normally distributed data (interval/ratio)
Mann-Whitney U Test: used for non-normally distributed data (ordinal)

Statistical Tests for Comparing Data

Paired-Samples Tests (Dependent Measures)

One group with two scores for each member
Data can be continuous or categorical
Appropriate tests: Paired T-Test or Wilcoxon Signed Rank Test
Paired T-Test: used for normally distributed data (interval/ratio)
Wilcoxon Signed Rank Test: used for non-normally distributed data (ordinal)

Unpaired-Samples Tests (Independent Measures)

Two or more groups with one score for each member of each group
Data can be continuous or categorical
Appropriate tests: Unpaired T-Test or Mann-Whitney U Test
Unpaired T-Test: used for normally distributed data (interval/ratio)
Mann-Whitney U Test: used for non-normally distributed data (ordinal)

Non-Parametric Tests

Wilcoxon Signed Rank Test is used for 2 dependent samples
Friedman test is used for 3 dependent samples
Rank Test (W-Test) is used for non-normal ordinal data
Mann-Whitney U test is a non-parametric test for groups
Kruskal-Wallis test is a non-parametric test for groups

Parametric Tests

PAIRED T-test is used for 2 dependent samples
Repeated measures of ANOVA is used for 3 dependent samples
Unpaired T-Test is used for groups
One Way / Two way ANOVA is used for groups

Test Selection

Dependent (paired) samples can be analyzed using non-parametric tests (Wilcoxon Signed Rank Test, Friedman) or parametric tests (PAIRED T-test, Repeated measures of ANOVA)
Independent (unpaired) samples can be analyzed using non-parametric tests (Mann-Whitney U, Kruskal-Wallis) or parametric tests (Unpaired T-Test, One Way / Two way ANOVA)
Non-normal ordinal data can be analyzed using non-parametric Rank Test (W-Test)

Parametric Tests

Repeated measures ANOVA is used for same group comparisons
One-way ANOVA and Two-way ANOVA are used for comparing three or more groups with one score

Non-Parametric Tests

Friedman test is used for comparing three or more scores in the same group
Kruskal-Wallis test is used for comparing three or more groups with one score

Make Your Own Quizzes and Flashcards

Convert your notes into interactive study material.