Unit 4: Types of Data in Research

I. Orientation: Understanding Research Data

Research data are systematically collected observations, measurements, accounts, documents, or records used to answer research questions and evaluate claims. The form of data and the method of obtaining it must fit the research objective: numerical questions generally require quantitative evidence, while questions about meaning, experience, or social processes often require qualitative evidence.

• Empirical foundation: Data connect research conclusions to observable evidence, such as survey responses, interview transcripts, laboratory measurements, or official statistics.
• Source: Data may be primary, collected directly for the current study, or secondary, originally collected for another purpose.
• Form: Data may consist of numbers, words, images, sounds, documents, behaviours, or physical traces.
• Methodological orientation: Quantitative research emphasizes measurement and numerical analysis; qualitative research emphasizes meaning, context, and interpretation.
• Unit of analysis: The entity being studied may be an individual, household, organization, event, document, community, or country.
• Variable or category: Quantitative studies commonly organize observations into measurable variables, whereas qualitative studies often develop concepts, categories, and themes from detailed material.
• Quality: Useful data should be relevant, sufficiently accurate, ethically obtained, and appropriate to the research question.
• Ethics: Collection, storage, analysis, and reporting must protect consent, privacy, confidentiality, and the integrity of the evidence.

II. Primary Data — Evidence Collected for the Present Study

A. Primary data and sources

Primary data are original data collected firsthand by the researcher specifically to address the objectives of the current investigation.

• Direct relationship: The researcher determines what will be collected, from whom, under what conditions, and by which instrument.
• Surveys and questionnaires: Structured questions generate responses about attributes, opinions, intentions, or behaviour; for example, a five-point satisfaction scale produces data coded from 1 to 5.
• Interviews: Structured, semi-structured, or unstructured conversations provide direct accounts from participants. Audio recordings are commonly transcribed before analysis.
• Observation: Researchers record events or behaviour in natural or controlled settings, such as counting customer visits or documenting classroom interaction.
• Experiments: The researcher manipulates an independent variable and observes its effect on a dependent variable while controlling other conditions.
• Focus groups: Guided group discussions reveal shared beliefs, disagreements, language, and interaction among participants.
• Measurements and tests: Instruments generate data such as temperature in degrees Celsius, examination scores, reaction time in milliseconds, or blood pressure in millimetres of mercury.
• Field materials: Photographs, diaries, field notes, maps, videos, and physical samples may serve as primary evidence when created or gathered for the study.
• Advantages:
  • Relevance: Collection is designed around the exact research question.
  • Control: The researcher can define sampling, measurement, and quality procedures.
  • Timeliness: Data reflect the period in which the current study is conducted.
• Limitations:
  • Cost and time: Recruiting participants and administering instruments can require substantial resources.
  • Researcher influence: Question wording, observation, or interviewer behaviour may affect responses.
  • Access: Some populations, locations, or sensitive experiences may be difficult to study directly.

B. Applications and quality considerations

The value of primary data depends on whether collection procedures produce evidence that is credible and fit for purpose.

• Sampling: The selected participants should appropriately represent, or meaningfully illuminate, the population or phenomenon under study.
• Instrument quality: A questionnaire or test should measure the intended concept consistently and accurately.
• Pilot testing: Trial administration can expose ambiguous questions, unsuitable response options, or practical difficulties.
• Standardization: Using the same instructions and procedures across cases reduces avoidable variation.
• Documentation: Dates, settings, instruments, consent procedures, and data-processing decisions should be recorded to support transparency.

III. Secondary Data — Reuse of Existing Evidence

A. Secondary data and sources

Secondary data are data previously collected, recorded, or compiled by another person or organization and later used for a different research purpose.

• Official statistics: Population censuses, labour-force surveys, health records, crime statistics, and education databases provide large-scale numerical evidence.
• Institutional records: Schools, hospitals, companies, and non-governmental organizations maintain administrative data such as attendance, admissions, sales, or service use.
• Published research: Journal articles, reports, theses, and archived datasets may supply findings or data suitable for reanalysis.
• Documentary sources: Newspapers, letters, policy papers, court decisions, meeting minutes, and historical archives can provide qualitative or quantitative evidence.
• Digital sources: Websites, online repositories, social-media archives, and platform records may contain text, images, networks, or behavioural traces.
• Advantages:
  • Economy: Existing evidence reduces the time and expense of new data collection.
  • Scale: National surveys and administrative databases may include samples beyond the capacity of an individual researcher.
  • Historical reach: Archived records allow changes to be examined across years or generations.
  • Non-reactivity: Previously produced records are not created in response to the current researcher’s presence.
• Limitations:
  • Mismatch: Definitions, units, categories, or collection periods may not fit the present question.
  • Unknown quality: Errors in sampling, measurement, coding, or record-keeping may be difficult to evaluate.
  • Incomplete coverage: Important groups or variables may be absent.
  • Restricted access: Confidential or proprietary datasets may require permission.

B. Evaluation and use

Secondary evidence must be critically assessed before it is incorporated into analysis.

• Authority: The competence, reputation, and purpose of the collecting organization should be examined.
• Method: Researchers should check the original population, sample, instrument, response rate, and collection procedure.
• Definitions: A term such as “unemployed” must have the same operational meaning across datasets before figures are compared.
• Currency: Older data may be unsuitable when studying rapidly changing conditions.
• Consistency: Units, coding schemes, time intervals, and geographical boundaries should be comparable.
• Triangulation: Comparing census data, institutional records, and interviews can reveal convergence or inconsistency between sources.

IV. Qualitative Research — Meaning, Experience, and Context

A. Nature of qualitative research

Qualitative research investigates how people interpret experiences and how social meanings, practices, identities, or processes develop within particular contexts.

• Interpretive orientation: Reality is examined through participants’ perspectives rather than reduced only to predetermined measurements.
• Natural settings: Data are frequently gathered where activities ordinarily occur, such as homes, workplaces, communities, or classrooms.
• Flexible design: Questions and sampling may evolve as new concepts emerge from fieldwork.
• Depth: Small, purposively selected samples can provide detailed accounts of complex experiences.
• Researcher involvement: The researcher participates in data collection and interpretation, making reflexivity essential.
• Inductive analysis: Patterns and explanations are often developed from the data, although existing theory may guide interpretation.
• Context sensitivity: Statements and actions are interpreted in relation to culture, history, setting, and social relationships.
• Outputs: Findings are usually presented as themes, categories, narratives, case descriptions, or conceptual models rather than statistical estimates.

B. Data and variables used in qualitative methods

Qualitative methods primarily use non-numerical data and analytical concepts rather than variables fixed in advance.

• Forms of data: Interview transcripts, focus-group discussions, observation notes, diaries, photographs, videos, documents, and online interactions are common materials.
• Cases: The main unit may be a person, organization, community, event, conversation, or text.
• Codes: Short labels identify meaningful segments; the statement “I avoid the clinic because staff dismiss me” might receive codes such as avoidance and perceived disrespect.
• Categories: Related codes are grouped into broader concepts, such as “barriers to healthcare.”
• Themes: Themes express recurring patterns of meaning across cases, such as “institutional mistrust.”
• Attributes: Age, role, location, or experience level may be recorded to compare perspectives, but these do not necessarily function as statistical variables.
• Relationships: Analysis explores how categories connect, including sequence, contradiction, cause, context, or consequence.
• Saturation: Data collection may continue until additional cases provide little new information relevant to the developing analysis.
• Quality criteria: Credibility, dependability, confirmability, and transferability are supported through techniques such as triangulation, reflexive notes, detailed description, and an audit trail.

C. Applications and limitations

Qualitative research is especially suitable for examining poorly understood phenomena, lived experience, organizational processes, and the meanings attached to behaviour.

• Applications: It can explain why a policy is resisted, how patients experience treatment, or how workplace norms shape decisions.
• Strength: Detailed evidence can reveal mechanisms and perspectives that fixed-response instruments overlook.
• Limited generalization: Findings from purposive, context-specific samples do not automatically represent a whole population.
• Interpretive risk: Researcher assumptions may shape coding and conclusions unless analysis is reflexive and transparent.
• Resource demands: Transcription, coding, comparison, and interpretation can be time-intensive.

V. Quantitative Research — Measurement and Numerical Explanation

A. Nature of quantitative research

Quantitative research represents phenomena numerically and uses statistical procedures to describe patterns, test hypotheses, estimate relationships, or evaluate causal effects.

• Deductive orientation: A theory commonly generates a hypothesis that is tested against observations.
• Standardized measurement: Participants receive consistent questions, tests, or experimental conditions.
• Structured design: Variables, instruments, sampling plans, and analytical procedures are usually specified before collection.
• Large samples: Probability sampling can support statistical generalization from a sample to a population.
• Objectivity: Explicit measurement and analysis rules aim to reduce subjective influence.
• Statistical analysis: Frequencies, percentages, means, correlations, confidence intervals, and regression models summarize or explain data.
• Replication: Clearly documented procedures allow another researcher to repeat the study.
• Causal testing: Randomized experiments provide strong evidence of causation by balancing alternative influences across groups.

B. Data and variables used in quantitative methods

Quantitative data consist of numerical values assigned to observations according to defined measurement rules.

• Independent variable: The presumed cause or predictor, such as hours of instruction.
• Dependent variable: The measured outcome, such as examination score.
• Control variable: A factor such as age or prior attainment included to isolate the relationship of interest.
• Confounding variable: An uncontrolled factor related to both predictor and outcome that may create a misleading association.
• Categorical variables:
  • Nominal: Categories have no inherent order, such as blood group.
  • Ordinal: Categories have a meaningful order, such as low, medium, and high satisfaction.
• Numerical variables:
  • Interval: Equal differences are meaningful, but zero is not absolute, as with temperature in degrees Celsius.
  • Ratio: Equal intervals and a true zero permit ratio comparisons, as with age, income, or mass.
• Discrete data: Countable values, such as the number of children in a household.
• Continuous data: Values within a range, such as height measured in centimetres.
• Operationalization: An abstract concept is converted into a measurable indicator; “academic achievement” may be operationalized as a standardized test score.
• Descriptive model:

Mean = Σx / n

• Σx is the sum of all observed values.
• n is the number of observations.
• For scores 60, 70, 80, the mean is 210 / 3 = 70.

C. Applications and limitations

Quantitative methods are useful when researchers need measurable comparisons, population estimates, hypothesis tests, or estimates of relationships.

• Applications: Surveys estimate prevalence, experiments compare treatments, and longitudinal datasets measure change over time.
• Strength: Standardized data enable concise comparison across many observations.
• Measurement limitation: A numerical indicator may capture only part of a complex concept.
• Statistical limitation: Correlation alone does not establish causation.
• Data-quality risk: Biased samples, missing values, unreliable instruments, and incorrect models can produce misleading precision.

VI. Reporting Qualitative Evidence

A. Writing up qualitative research

Writing up qualitative research presents an evidence-based interpretation that connects participants’ accounts, contextual description, analytical themes, and the research question.

• Research context: The report identifies the setting, participants, sampling rationale, and circumstances of data collection.
• Methodological transparency: It explains interview, observation, transcription, coding, and theme-development procedures.
• Reflexivity: The researcher states how background, role, relationships, and assumptions may have influenced access and interpretation.
• Thematic organization: Findings are arranged by analytical themes or processes rather than merely reproducing questions in sequence.
• Evidence integration: Each interpretation is supported by concise quotations, field-note extracts, or document examples.
• Quotation use: Quotations illustrate a claim but do not replace analysis; the writer explains what the words reveal and why they matter.
• Participant protection: Pseudonyms and removal of identifying details preserve confidentiality unless explicit attribution is ethically authorized.
• Variation and contradiction: Reports include minority views, negative cases, and tensions that qualify the dominant pattern.
• Contextual detail: Sufficient description enables readers to judge whether findings may transfer to another setting.
• Discussion: Themes are connected to the research question and relevant concepts without claiming statistical representativeness.
• Coherent conclusion: The report states the interpretation supported by the data, its boundaries, and its implications.
• Auditability: Clear links among raw material, codes, categories, themes, and conclusions allow readers to assess the analytical reasoning.