Definitions: Qualitative and Quantitative Data
Qualitative data describes categories or characteristics that cannot be meaningfully measured with numbers (e.g., hair color, interview responses, blood type). Quantitative data consists of numerical values obtained through counting or measurement (e.g., height, temperature, test score). The core test: if you can perform meaningful arithmetic on the values, the data is quantitative. If the numbers are labels rather than measurements, the data is qualitative.
What Is Qualitative Data?
The word "qualitative" comes from the Latin qualitas, meaning "quality" or "nature." This data type captures the quality of something — what it is, not how much of it there is. A researcher collecting qualitative data records descriptions, labels, themes, or groupings. Analysis methods include counting frequencies of categories, identifying themes in text, or organizing responses into groups.
Qualitative data falls into two subtypes based on whether the categories have a natural order. When the categories have no inherent ranking — blood type, nationality, favorite color — the data is nominal. When the categories do follow an order but the gaps between them are not equal — satisfaction ratings, academic grades, pain intensity scales — the data is ordinal.
What Is Quantitative Data?
Quantitative data breaks further into two subtypes. Discrete data takes only whole-number values obtained through counting — the number of cars in a parking lot, the number of errors on an exam, or the number of goals scored in a match. Continuous data can take any value within a range and comes from measurement — body weight, air pressure, distance, or time elapsed. The distinction matters because discrete and continuous data are visualized and analyzed differently (histograms vs bar charts; different distributions). The full breakdown is in the discrete vs continuous data guide.
Key Differences: Qualitative vs Quantitative Data
The table below compares the two data types across 10 dimensions. Every row reflects a genuine difference in how the data is collected, stored, analyzed, or interpreted — not just terminology.
| Dimension | Qualitative Data | Quantitative Data |
|---|---|---|
| Nature | Descriptive — categories, labels, themes | Numerical — counts or measurements |
| Core question answered | What kind? Which category? What is it? | How much? How many? How often? |
| Data subtypes | Nominal, ordinal | Interval, ratio (discrete or continuous) |
| Arithmetic operations | Not meaningful (you cannot average "blood types") | Meaningful — addition, mean, standard deviation all valid |
| Central tendency measures | Mode (nominal); mode and median (ordinal) | Mean, median, mode all apply |
| Common collection methods | Interviews, open-ended surveys, observation, focus groups | Closed surveys, experiments, sensors, administrative records |
| Common analysis methods | Thematic analysis, content analysis, coding, chi-square test | Descriptive statistics, t-tests, ANOVA, regression, correlation |
| Visualization | Bar charts, pie charts, word clouds, thematic maps | Histograms, scatter plots, box plots, line charts |
| Sample size | Typically small (depth over breadth) | Typically large (statistical power requires adequate n) |
| Research goal | Exploration, theory generation, context, meaning | Testing hypotheses, measuring effects, predicting outcomes |
Examples: Qualitative and Quantitative Data
Qualitative Data Examples
The following dataset comes from a hypothetical patient intake survey at a hospital. Each row shows a variable, its value for a sample patient, and its qualitative subtype.
| Variable | Example Value | Subtype | Why Qualitative |
|---|---|---|---|
| Blood type | A+ | Nominal | Categories have no natural order; A is not "more" than B |
| Marital status | Married | Nominal | Labels with no inherent ranking |
| Pain intensity rating | Severe | Ordinal | Ordered (mild < moderate < severe) but gaps are not equal |
| Patient symptom description | "Sharp, burning pain in left knee" | Nominal (text) | Free-text — analyzed via thematic coding |
| Satisfaction with care | Very satisfied | Ordinal | Ordered scale; cannot say the gap between "satisfied" and "very satisfied" equals the gap between "neutral" and "satisfied" |
| Preferred communication | Phone | Nominal | Unordered categories — phone, email, text, in-person |
| Smoking status | Former smoker | Ordinal | Ordered: never < former < current |
Quantitative Data Examples
The same patient intake scenario produces quantitative variables. Note which are discrete (counted) and which are continuous (measured).
| Variable | Example Value | Subtype | Why Quantitative |
|---|---|---|---|
| Age | 47 years | Ratio (discrete) | Counted in whole years; true zero exists (0 years = newborn) |
| Body temperature | 38.6°C | Interval (continuous) | Measured; equal intervals. But 0°C is not "no temperature" — it's just the freezing point of water, so ratios are not meaningful |
| Body weight | 72.4 kg | Ratio (continuous) | Measured; true zero exists (0 kg = no mass). 144 kg is genuinely twice 72 kg |
| Number of prior visits | 3 | Ratio (discrete) | Counted; true zero exists; arithmetic fully meaningful |
| Systolic blood pressure | 128 mmHg | Ratio (continuous) | Measured; true zero (0 mmHg) = no pressure |
| Wait time in minutes | 23 min | Ratio (continuous) | Measured duration; 0 minutes = no wait; 46 min = twice as long |
| Number of medications prescribed | 2 | Ratio (discrete) | Whole-number count; true zero possible |
The Four Measurement Scales: Nominal, Ordinal, Interval, Ratio
Stanley Smith Stevens introduced the four-level hierarchy of measurement scales in 1946 in the journal Science. It remains the standard framework used in research methods courses, statistics textbooks, and data analysis software. Each scale level includes all the properties of the levels below it and adds one more.
Nominal Scale — Categories With No Order
Data exists in named categories with no inherent ranking. The only meaningful operation is checking equality (is category A the same as category B?). Appropriate statistics: frequency counts, mode, chi-square test. Examples: eye color, blood type, country of birth, favorite music genre, political party affiliation.
Ordinal Scale — Ordered Categories, Unequal Gaps
Categories follow a meaningful sequence, but the distance between consecutive categories is not uniform or known. You can rank responses, but you cannot say the difference between rank 1 and rank 2 equals the difference between rank 2 and rank 3. Appropriate statistics: mode, median, Spearman rank correlation, Mann-Whitney test. Examples: academic letter grades (A, B, C, D, F), Likert scales, military ranks, Olympic medal positions.
Interval Scale — Equal Gaps, No True Zero
The gap between any two adjacent values is the same throughout the scale, but there is no meaningful zero point. This means ratios are not interpretable — you cannot say that 80°F is "twice as hot" as 40°F. Appropriate statistics: mean, standard deviation, Pearson correlation, t-tests. Examples: temperature in Celsius or Fahrenheit, IQ scores, calendar years, SAT scores.
Ratio Scale — Equal Gaps and True Zero
Has all the properties of an interval scale plus a true, meaningful zero point. A value of zero means the complete absence of the measured property. Ratios between values are interpretable — a person earning $80,000 earns exactly twice as much as one earning $40,000. Appropriate statistics: all statistical operations including geometric mean and coefficient of variation. Examples: height, weight, distance, reaction time, income, count of items.
Temperature in Celsius and Fahrenheit is interval, not ratio, because 0°C does not mean "no temperature." Temperature in Kelvin is ratio because 0 K is absolute zero — the complete absence of thermal energy. This is a frequent multiple-choice question in research methods and statistics courses.
Summary: Which Scale Allows Which Operations?
| Scale | Type | Rank? | Equal Gaps? | True Zero? | Valid Statistics |
|---|---|---|---|---|---|
| Nominal | Qualitative | No | No | No | Mode, frequency, chi-square |
| Ordinal | Qualitative | Yes | No | No | Median, mode, Spearman ρ |
| Interval | Quantitative | Yes | Yes | No | Mean, SD, Pearson r, t-test |
| Ratio | Quantitative | Yes | Yes | Yes | All statistics, including geometric mean and CV |
Practice: Classify These 20 Variables
Each item below comes from a real research scenario. Work through the classification yourself, then check the answer. The pattern to check: can you perform meaningful arithmetic on the values? If yes, it is quantitative. If no, it is qualitative — and then ask whether the categories have an order (ordinal) or not (nominal).
A school district collects the following data on 500 students. Classify each variable.
Exam score (0–100): Ratio (quantitative) — Meaningful zero (0 = no correct answers); arithmetic valid; 90 is genuinely 1.5 times 60.
Letter grade (A, B, C, D, F): Ordinal (qualitative) — Ordered (A > B > C) but the "gap" between A and B may not equal the gap between B and C in terms of actual knowledge.
Preferred learning style (visual, auditory, kinesthetic): Nominal (qualitative) — Unordered categories; no learning style is inherently "more" than another.
Number of absences: Ratio (quantitative) — Whole-number count; 0 absences = perfect attendance; 10 absences is twice as many as 5.
Student's open-ended response to "What makes a good teacher?": Nominal (qualitative) — Free text; analyzed via thematic coding or content analysis.
✅ Notice: the same underlying concept (academic performance) produced both quantitative data (score) and qualitative data (grade, open-ended response). Research design choices determine the data type you collect.
An e-commerce company analyzes customer data after a product launch. Classify each variable.
Customer review text ("Fast shipping, great quality!"): Nominal (qualitative) — Descriptive text; analyzed via sentiment analysis or thematic coding.
Star rating (1–5 stars): Ordinal (qualitative) — Ordered but the gap between 1 and 2 stars does not necessarily equal the gap between 4 and 5 stars. Treating these as interval-level data is a common analysis error.
Total purchase amount ($142.50): Ratio (quantitative) — Measured in dollars; true zero ($0 = no purchase); $200 is exactly twice $100.
Number of items purchased: Ratio (quantitative) — Whole-number count; true zero; arithmetic valid.
Payment method (credit card, PayPal, bank transfer): Nominal (qualitative) — Unordered categories; no payment method is "more" than another.
✅ Mixed datasets like this are the norm in business analytics. Software like SPSS, R, or Python's pandas will require you to declare the correct data type for each variable before running any analysis.
Interactive Data Type Classifier
Data Type Identifier Tool
Type a variable or a few example values below. The tool will classify the data and explain why.
Try: age, gender, income, temperature, satisfaction rating, number of children, IQ score, blood type, wait time, education level
Qualitative vs Quantitative Research Methods
Data type and research method are closely linked but not identical. Qualitative research tends to collect qualitative data; quantitative research tends to collect quantitative data. The connection is logical: if you want to measure something numerically and test a hypothesis, you design a study that produces numbers. If you want to understand why people behave a certain way, you design a study that produces rich descriptions.
| Dimension | Qualitative Research | Quantitative Research |
|---|---|---|
| Purpose | Explore, describe, understand experiences, generate theory | Test hypotheses, measure effects, generalize findings |
| Research questions | How? Why? What does it mean? What is the experience of…? | How much? How often? Is there a significant difference? Does X cause Y? |
| Common methods | In-depth interviews, focus groups, ethnography, case studies, grounded theory | Surveys with closed questions, randomized controlled trials, cohort studies, experiments |
| Sample size | Small (8–30 typically); depth matters more than breadth | Large (30+ for basic tests; hundreds or thousands for surveys); statistical power depends on n |
| Analysis | Thematic analysis, content analysis, discourse analysis, grounded theory coding | Descriptive statistics, hypothesis testing, regression, ANOVA, structural equation modeling |
| Generalizability | Limited — findings are context-specific and transferable rather than statistically generalizable | High — statistical inference allows conclusions about the population from a random sample |
| Strengths | Rich, contextual, explores meaning; ideal when little prior theory exists | Replicable, objective, allows comparison across groups; strong external validity with random sampling |
| Limitations | Researcher subjectivity; results hard to replicate exactly; cannot make statistical inferences | May miss context and nuance; structured questions may not capture complex experiences |
| Key researchers | John W. Creswell, Kathy Charmaz, Norman Denzin | Ronald Fisher, Karl Pearson, Jerzy Neyman |
Mixed Methods Research: Combining Both Approaches
Mixed methods research uses both qualitative and quantitative data in a single study. John W. Creswell — widely regarded as the leading authority on mixed methods design — describes it as a methodology that "involves philosophical assumptions as well as methods of inquiry." The practical reason researchers combine the two: neither approach alone answers every question a study raises.
Case Study — Mixed Methods
Hospital Patient Experience Research
A hospital wants to improve patient satisfaction. The research team uses a quantitative survey of 800 patients, producing a mean satisfaction score of 6.8/10 and revealing that patients in the surgical ward scored significantly lower than others (t-test, p = 0.003). But the score alone does not explain why surgical patients are dissatisfied.
The team then conducts qualitative in-depth interviews with 20 surgical patients, coding responses into themes. The dominant theme — "staff didn't explain what was happening before procedures" — gives actionable insight the survey alone could not have captured. The two data types complement each other: numbers show what and how much; words show why.
Common Mixed Methods Designs
Convergent Design
Collect qualitative and quantitative data simultaneously, analyze separately, then compare. Use when both types of data are equally important.
Sequential Explanatory
Collect quantitative data first, then qualitative data to explain the quantitative results. Use when you need context for statistical findings.
Sequential Exploratory
Collect qualitative data first to explore the topic, then design a quantitative phase informed by those findings. Common in instrument development.
Embedded Design
One method is nested within the other (e.g., qualitative interviews embedded within a randomized trial). The primary design is one type; the secondary adds depth.
When to Use Qualitative, Quantitative, or Mixed Methods
The choice of method follows from the research question, not the other way around. Write your research question first — then ask which approach will answer it.
Method Selection Decision Tree
The DATA Framework for Variable Classification
Students often misclassify variables when facing unfamiliar examples in exams or research projects. The DATA Framework provides a four-question checklist that works for any variable, from simple survey questions to complex experimental measurements.
Determine the research goal
Are you trying to describe a category (qualitative goal) or measure a quantity (quantitative goal)? The research purpose shapes the variable type you need.
Assess the type of information captured
Does the variable record a label, name, or description? Or does it record a count or measurement with numeric meaning?
Test whether arithmetic on the values is meaningful
Can you calculate a meaningful average of the values? If "blood type" values averaged to 1.7, would that mean anything? No → qualitative. If height values averaged to 175.3 cm, does that mean something? Yes → quantitative.
Apply the appropriate scale and method
Once classified, identify the measurement scale (nominal, ordinal, interval, ratio) and select the correct statistical summary or test for that scale level.
Case Studies: Real-World Applications
Healthcare: Clinical Trial Design
Case Study — Healthcare
COVID-19 Vaccine Efficacy Research
A Phase III clinical trial of a vaccine collects both data types. Quantitative data includes the number of confirmed infections per 1,000 person-months of follow-up in vaccine vs placebo groups — this produces the efficacy percentage via a ratio calculation. Qualitative data includes participant descriptions of adverse events ("mild arm soreness for two days"), which cannot be meaningfully averaged but are coded and counted to assess safety profiles.
The primary endpoint (efficacy) is quantitative and analyzed using statistical methods like the relative risk ratio. Secondary safety analysis uses both qualitative descriptions and quantitative counts of adverse event severity grades.
Social Science: Poverty Research
Case Study — Social Science
Understanding Long-Term Unemployment
A government research institute studying long-term unemployment uses quantitative data (unemployment duration in weeks, income in dollars, number of job applications submitted) to establish the scale of the problem and compare demographic groups using ANOVA. Qualitative data from in-depth interviews captures what applicants experience during the job search — feelings of stigma, gaps in digital literacy, network access constraints — that the numbers alone cannot explain. Together, they produce a policy recommendation with both statistical evidence and lived experience context.
Business: Customer Satisfaction Research
Case Study — Business
Net Promoter Score Analysis
A software company runs a customer satisfaction survey. The Net Promoter Score (NPS) question — "How likely are you to recommend this product? (0–10)" — produces ordinal data that is conventionally treated as ratio-level for calculation purposes (a common analytical trade-off). The follow-up open question — "What is the main reason for your score?" — produces qualitative data analyzed via thematic coding. The NPS number tells the company how many customers are promoters vs detractors. The open responses tell them why — revealing that the most common theme among detractors is "slow customer support response times," enabling a targeted operational change.
Analysis Methods by Data Type
Choosing the wrong analysis method for your data type produces misleading or invalid results. The table below maps data types to their appropriate descriptive statistics and inferential tests, drawn from the statistical method guidance at NIST's Engineering Statistics Handbook and UCLA's statistical consulting resources.
| Data Type & Scale | Descriptive Statistics | Common Inferential Tests | Visualization |
|---|---|---|---|
| Nominal (Qualitative) | Mode, frequency count, proportion | Chi-square test of independence or goodness-of-fit | Bar chart, pie chart |
| Ordinal (Qualitative) | Median, mode, frequency count | Mann-Whitney U, Wilcoxon signed-rank, Kruskal-Wallis, Spearman ρ | Bar chart, stacked bar, ordered frequency table |
| Interval (Quantitative) | Mean, median, standard deviation, variance, range | t-tests, ANOVA, Pearson r, linear regression | Histogram, box plot, scatter plot, line chart |
| Ratio (Quantitative) | All of the above, plus geometric mean, coefficient of variation | All of the above, plus log-transformed regression, relative risk | All of the above |
Need to select a statistical test based on your data type? The statistical test selector at Statistics Fundamentals walks through the decision based on your number of groups, data type, and whether you have paired or independent samples.
Software Tools for Qualitative and Quantitative Analysis
SPSS (IBM)
Industry standard for quantitative social science research. Handles descriptive stats, t-tests, ANOVA, regression, and chi-square with a point-and-click interface.
Python (pandas, scipy)
Free and powerful for both data types. pandas handles data classification and manipulation; scipy.stats provides hypothesis testing; matplotlib produces visualizations.
R
The statistical computing standard in academia. Factor variables handle qualitative data; numeric vectors handle quantitative data. Packages like ggplot2 visualize both.
NVivo
Dedicated qualitative analysis software. Codes themes in interview transcripts, documents, and multimedia. Also supports mixed methods by importing survey data alongside qualitative materials.
Atlas.ti
Qualitative and mixed methods research tool for coding, theming, and annotating text, audio, video, and images. Common in social science and healthcare research.
Qualtrics / Google Forms
Survey platforms that produce both data types simultaneously: closed Likert questions → ordinal/quantitative; open-ended questions → qualitative text data.
Common Misconceptions and Exam Traps
| Misconception | Why It's Wrong | Correct Classification |
|---|---|---|
| "A Likert scale (1–5) is quantitative because it uses numbers" | The numbers are labels for ordered categories. The gap between '2 — disagree' and '3 — neutral' is not guaranteed to equal the gap between '4 — agree' and '5 — strongly agree'. | Ordinal qualitative data. Use median, not mean, unless the interval-level assumption is explicitly justified and acknowledged. |
| "Temperature in Celsius is ratio data because it measures real temperature" | 0°C is not the absence of temperature — it's the freezing point of water, an arbitrary reference. Ratios are therefore meaningless: 40°C is not "twice as hot" as 20°C. | Interval quantitative data. Temperature in Kelvin is ratio (0 K = absolute zero = true absence of thermal energy). |
| "Qualitative data means the study used interviews" | Data type and collection method are different things. A Likert survey produces ordinal data, which is qualitative. An interview can also produce counts or ratings, which are quantitative. | Qualitative data = non-numeric or categorically coded values. The collection method can vary. |
| "Zip codes are quantitative because they are numbers" | Zip code 90210 is not "greater than" 10001. Arithmetic on zip codes is meaningless. They are identification labels, not measurements. | Nominal qualitative data — numeric codes used as category labels. |
| "Qualitative data is less scientific or rigorous than quantitative data" | Rigor in qualitative research comes from different criteria: credibility, transferability, dependability, and confirmability (Lincoln & Guba, 1985). Neither type is inherently more rigorous — quality depends on how the method matches the research question. | Both types have scientific standards; they are evaluated differently. See APA's Psychological Methods for published qualitative research standards. |
| "You can convert qualitative data to quantitative data by assigning numbers to categories" | Assigning 1 = male, 2 = female does not make gender quantitative. The numbers are codes, not measurements. Computing "average gender" of 1.4 carries no meaning. | Nominal data remains nominal regardless of coding. Dummy variables in regression encode categories but do not change their fundamental nature. |
Quick Reference: Glossary of Key Terms
The table below defines every key term used in the study of qualitative vs quantitative data. Definitions follow the standard used by the American Statistical Association (ASA) and major research methods textbooks.
| Term | Definition | Example |
|---|---|---|
| Qualitative data | Non-numeric data describing categories, labels, or characteristics | Blood type: A, B, AB, O |
| Quantitative data | Numeric data obtained through counting or measurement | Height: 175 cm |
| Nominal data | Qualitative data in unordered categories | Eye color: blue, green, brown |
| Ordinal data | Qualitative data in ordered categories with unequal gaps | Education: high school < bachelor's < master's < doctorate |
| Interval data | Quantitative data with equal gaps but no true zero | Temperature in Celsius: 20°C, 25°C, 30°C |
| Ratio data | Quantitative data with equal gaps and a true zero | Weight in kg: 0 kg (no mass), 50 kg, 100 kg |
| Discrete data | Quantitative data taking only whole-number values; obtained by counting | Number of errors: 0, 1, 2, 3… |
| Continuous data | Quantitative data that can take any value within a range; obtained by measuring | Body temperature: 37.1°C, 37.15°C… |
| Variable | A characteristic that can take different values across subjects or observations | Height, gender, test score, blood type |
| Thematic analysis | A qualitative method for identifying, analyzing, and reporting patterns (themes) within qualitative data | Coding interview transcripts to find recurring themes about job satisfaction |
| Content analysis | Systematic categorization of qualitative content; can produce frequency counts (quantitative output) | Counting how often certain words appear in 200 newspaper articles |
| Mixed methods | Research integrating both qualitative and quantitative data and analysis in a single study | Survey (quantitative) followed by interviews (qualitative) to explain survey results |
| Sampling | The process of selecting a subset of individuals from a population for study | Random sampling of 200 students from a school of 2,000 |
| Descriptive statistics | Numerical summaries of data (mean, median, mode, standard deviation, frequencies) | Mean exam score = 74.2, SD = 11.3 |
| Inferential statistics | Statistical methods for drawing conclusions about a population from sample data | Hypothesis testing, confidence intervals, regression |
| Hypothesis testing | A procedure for deciding whether sample data provides sufficient evidence to reject a default claim about a population | Testing whether mean exam scores differ between two teaching methods |
Frequently Asked Questions
Qualitative data describes categories or characteristics using words or labels — for example, blood type (A, B, AB, O), marital status, or open-ended survey responses. Quantitative data consists of numbers obtained through counting or measurement — for example, age in years, weight in kilograms, or number of sales per month. The key diagnostic: if you can perform meaningful arithmetic (calculate an average that makes sense), the data is quantitative. If the values are labels or categories where arithmetic is meaningless, the data is qualitative.
Age is quantitative data. It is measured numerically (25 years, 40 years) and arithmetic operations on it are meaningful — you can calculate an average age, and saying one person is twice as old as another makes mathematical sense. Age has a true zero (0 years = newborn), which makes it ratio-level quantitative data. The one exception: if age is recorded in categories such as "child," "adult," or "senior," it becomes ordinal qualitative data because those categories follow an order but the gaps are not uniform.
Gender is qualitative (categorical) data. It describes a category — male, female, non-binary — rather than a numerical measurement. Even when gender categories are assigned numeric codes for data entry (e.g., 1 = male, 2 = female), performing arithmetic on those codes is meaningless. Gender is nominal qualitative data because the categories have no natural order — male is not "more than" or "less than" female in any measurable sense.
Income measured in dollars or any currency is ratio quantitative data. It has a true zero ($0 = no income), equal intervals between values, and arithmetic is fully meaningful — $80,000 is exactly twice $40,000. However, if income is recorded in brackets ("under $30,000," "$30,000–$60,000," "over $60,000"), it becomes ordinal qualitative data. The data type changes depending on how the variable is measured and recorded.
The four measurement scales, introduced by Stanley Smith Stevens (1946), are: (1) Nominal — unordered categories (blood type, nationality); (2) Ordinal — ordered categories with unequal gaps (satisfaction ratings, education level); (3) Interval — equal gaps but no true zero (temperature in Celsius, IQ scores); (4) Ratio — equal gaps and a true zero (height, weight, income, age). Each higher level includes all the properties of the levels below it. The scale level determines which descriptive statistics and inferential tests are valid for that variable.
A Likert scale is technically ordinal qualitative data — the categories are ordered but the gaps between them are not guaranteed to be equal. However, in practice, many researchers in psychology, education, and social science treat aggregated Likert items as interval-level (quantitative) when certain conditions hold: the scale has five or more response options, responses are approximately normally distributed, and the scale is composed of multiple items rather than a single question. This practice is debated in the statistical literature. The strictly correct approach for a single Likert item is to use nonparametric tests (median, Mann-Whitney U, Spearman correlation) rather than the mean and Pearson correlation.
Qualitative research explores the depth of human experience, meaning, and context — asking "why" and "how." It uses methods like in-depth interviews, focus groups, and ethnographic observation. Sample sizes are small; analysis involves coding and identifying themes. Quantitative research tests hypotheses and measures relationships between variables — asking "how much" and "is there a significant difference." It uses structured surveys, experiments, and administrative data. Sample sizes are large; analysis involves descriptive statistics, hypothesis testing, and regression. Mixed methods research combines both to answer questions that neither approach can address alone.
Discrete quantitative data comes from counting and can only take whole-number values with gaps between them — for example, the number of students in a class (you cannot have 24.7 students), the number of goals scored in a match, or the number of errors on an exam. Continuous quantitative data comes from measurement and can take any value within a range, including fractions — for example, height (175.3 cm), time (12.47 seconds), or body temperature (37.2°C). The boundary between discrete and continuous matters for choosing between certain distributions and visualization methods. For a full explanation, see the discrete vs continuous data guide.
Sources and References
This guide draws on primary statistical theory and research methods literature. All definitions, scale classifications, and methodological descriptions are cross-referenced against peer-reviewed sources and government statistical standards.
- Stevens, S.S. (1946) — "On the theory of scales of measurement." Science, 103(2684), 677–680. Foundation text establishing the four measurement scale levels. sciencemag.org
- Creswell, J.W. & Creswell, J.D. (2018) — Research Design: Qualitative, Quantitative, and Mixed Methods Approaches (5th ed.). SAGE Publications. Standard graduate-level research methods textbook.
- NIST Engineering Statistics Handbook — Data Analysis. National Institute of Standards and Technology. itl.nist.gov
- OpenStax Introductory Statistics — Ch. 1: Sampling and Data. Rice University. openstax.org
- Penn State STAT 200 — Elementary Statistics. 1.1 Variables and Data. Penn State Eberly College of Science. online.stat.psu.edu
- American Statistical Association (ASA) — Statistical guidelines and definitions. amstat.org
- Lincoln, Y.S. & Guba, E.G. (1985) — Naturalistic Inquiry. SAGE Publications. Standard reference for qualitative research rigor criteria (credibility, transferability, dependability, confirmability).