The document outlines the levels of measurement in statistics, including ratio, interval, ordinal, and nominal, and their importance in data analysis. It provides guidelines on selecting appropriate statistical analyses based on these measurement levels, such as t-tests for ratio/interval data and chi-square tests for nominal data. Additionally, it promotes Statistics Solutions as a dissertation consulting service offering editorial support to graduate students.

1. Levels of Measurement

2. Additional
Support
Statistics Solutions is a full-service dissertation consulting company
providing graduate students timely, editorial support for their
dissertations and scholarly projects
For information about our services, receive a complementary 30-min
consultation available Mon-Fri 9-5 ET:
https://app.hubspot.com/meetings/jeanine/dissertation-consultation
info@StatisticsSolutions.com
Phone: 877-437-8622

3. Overview
Why are levels of measurement important?
Basic levels of measurement
• Ratio and Interval
• Ordinal
• Nominal
How to determine appropriate analysis based on levels of
measurement
Questions
Need help with your dissertation? Call 727-442-4290

4. Why are levels
of
measurement
important?
Levels of measurement are the properties of
values/categories that appear in data
Levels of measurement help determine the most
appropriate ways to describe, analyze, and interpret data
Four basic levels: ratio, interval, ordinal, and nominal
Other classifications exist, and data may need
consideration beyond the four basic levels
Need help with your dissertation? Call 727-442-4290

5. Ratio and
Interval
Ratio and interval measures represent continuous numerical
values
• Ratio: Numeric values with a meaningful (non-arbitrary) zero
• Age (in years), income, height, weight, counts
• Interval: Numeric values with no meaningful (arbitrary) zero
• Temperature (in C/F), dates/years, scale scores and composite measurements
Degree of difference between values is known and consistent
Mathematical operations can be meaningfully applied
• Ratio:All operations (addition, subtraction, multiplication, division, etc.)
• Interval: Simple addition/subtraction; cannot meaningfully multiply or divide
Need help with your dissertation? Call 727-442-4290

6. Ordinal
Ordinal measures represent categories that can be
ranked or ordered (e.g., from lowest to highest)
• Education (high school, college, graduate school)
• Letter grade (A, B, C, D, F)
• Likert items (strongly disagree to strongly agree)
Degree of difference between categories is
unknown and/or inconsistent
Basic mathematical operations cannot be applied
Need help with your dissertation? Call 727-442-4290

7. Nominal
Nominal measures represent categories that CANNOT
be ranked or ordered
• Gender (male, female, other)
• Race/Ethnicity
• Location, region, or nationality
• Occupation or specialty
Categories cannot be meaningfully ordered, and degrees
of difference between categories cannot be determined
Basic mathematical operations cannot be applied
Need help with your dissertation? Call 727-442-4290

8. Instruments
with Likert
Responses
Likert items: Attitudes, behaviors, etc. are rated on a numeric scale:
Survey instruments may use multiple Likert items to measure a single
variable/construct (e.g., self-esteem, anxiety, self-efficacy)
A single Likert item is typically treated as ordinal
A sum/composite score of multiple Likert items is typically treated as interval
Need help with your dissertation? Call 727-442-4290

9. Choosing
Appropriate
Analyses
Descriptive Statistics
•Ratio/interval: mean, standard
deviation, median, mode, min,
max, skewness, kurtosis
•Ordinal: frequencies, median,
mode, min, max
•Nominal: frequencies, mode
Need help with your dissertation? Call 727-442-4290

10. Choosing
Appropriate
Analyses
When the dependent variable is
ratio/interval…
• t-tests and ANOVAs may be used to compare
groups (ordinal/nominal independent variable) or
repeated measurements
• Pearson correlations may be used to determine
the relationship between two ratio/interval
variables
• Linear regression may be used to predict
ratio/interval outcome variables (DVs) from
predictors (IVs) of any level of measurement
Need help with your dissertation? Call 727-442-4290

11. Choosing
Appropriate
Analyses
When the dependent variable is
ordinal…
• Non-parametric difference tests (Mann-Whitney,
Kruskal-Wallis, Wilcoxon, Friedman) may be used
to compare groups or repeated measurements
• Spearman correlations may be used to
determine the relationship between two ordinal
variables
• Ordinal logistic regression may be used to
predict ordinal outcome variables (DVs) from
predictors (IVs) of any level of measurement
Need help with your dissertation? Call 727-442-4290

12. Choosing
Appropriate
Analyses
When the dependent variable is
nominal…
• Chi-square tests may be used to
determine the relationship between two
nominal variables
• Binary/multinomial logistic regression
may be used to predict nominal outcome
variables (DVs) from predictors (IVs) of
any level of measurement
Need help with your dissertation? Call 727-442-4290

13. Additional
Support
Statistics Solutions is a full-service dissertation consulting company
providing graduate students timely, editorial support for their
dissertations and scholarly projects
For information about our services, receive a complementary 30-min
consultation available Mon-Fri 9-5 ET:
https://app.hubspot.com/meetings/jeanine/dissertation-consultation
info@StatisticsSolutions.com
Phone: 877-437-8622