Data: Measurement, Analysis and Statistical Techniques | Notes | UGC NET Education

The term data is derived from the Latin word datum, meaning “something given.” In educational research, data refers to the facts, figures, observations, measurements, responses or other information collected systematically to investigate a research problem, test a hypothesis, identify relationships and draw meaningful conclusions.

Data are the raw materials of research. They may not have a clear meaning when considered in isolation, but when they are organised, processed and analysed, they help researchers develop information and knowledge. For example, the marks obtained by students in an achievement test constitute data. When these marks are analysed to calculate the class average, compare student performance or identify learning difficulties, they provide meaningful information for educational decision-making.

According to Kerlinger (1986), data can be understood as the scores or measurements obtained from subjects on one or more variables. A variable is a characteristic that can take different values, such as age, intelligence, academic achievement, motivation or teaching effectiveness. Data collected about these variables enable researchers to examine patterns, differences and relationships.

Examples of educational research data include students’ achievement test scores, responses to questionnaires, interview transcripts, classroom observation records, attendance percentages, teacher reflections and school enrolment statistics.

Characteristics of Data

1. Systematic collection: Data should be collected according to a planned procedure using appropriate research methods and tools. Systematic collection reduces errors and improves the consistency of research findings.
2. Relevance: Data must be directly related to the objectives and research questions of the study. Irrelevant data consume time and resources without contributing meaningfully to the investigation.
3. Quantitative or qualitative nature: Data may be numerical, such as examination marks and attendance percentages, or descriptive, such as students’ opinions, teachers’ experiences and classroom interactions.
4. Accuracy: Data should represent the actual observations, responses or measurements as correctly as possible. Errors in recording, measurement or data entry can affect the findings.
5. Validity: Data should appropriately represent the characteristic or phenomenon that the researcher intends to study. For example, a test designed to measure mathematical ability should assess relevant mathematical skills.
6. Reliability and consistency: Data collected through a reliable instrument or procedure should produce reasonably consistent results under comparable conditions. Reliability is particularly important when using achievement tests, rating scales and questionnaires.
7. Verifiability: Data should be recorded and documented in a manner that allows checking, examination and verification by the researcher or other authorised investigators.
8. Objectivity: Data collection and interpretation should minimise personal bias. Clearly defined procedures and standardised instruments help researchers maintain objectivity.
9. Analysability: Data should be organised in a suitable form for interpretation. Numerical data can be analysed statistically, whereas descriptive data may be examined through coding, categorisation and thematic analysis.
10. Empirical basis: Research data are generally grounded in observation, experience, measurement, documents or recorded responses rather than unsupported assumptions.

Importance of Data in Educational Research

Data play a central role in scientific investigation because research findings depend on the quality and relevance of the information collected.

• Basis of analysis and interpretation: Data enable researchers to examine educational problems systematically and identify meaningful patterns.

• Testing hypotheses: Empirical data help researchers determine whether the evidence supports or contradicts a proposed hypothesis.

• Comparison: Data allow comparisons between students, schools, teaching methods, educational programmes and assessment techniques.

• Prediction: Statistical relationships and historical patterns in data can help researchers estimate future outcomes, such as student achievement or school enrolment.

• Generalisation: Properly collected data from an appropriate sample may support conclusions about a larger population, subject to the sampling method and limitations of the study.

• Policy formulation: Data on enrolment, attendance, dropout rates, learning outcomes and teacher availability help educational authorities formulate and evaluate policies.

• Decision-making: Teachers and administrators use data to identify learning gaps, allocate resources and improve teaching strategies.

• Scientific objectivity: Evidence-based investigation reduces dependence on personal opinions and unsupported beliefs.

• Evaluation of programmes: Data help assess the effectiveness of educational interventions, curriculum changes, teacher training and student-support programmes.

Data in Quantitative and Qualitative Research

Data are broadly classified as quantitative and qualitative according to their nature and the type of information they provide.

Quantitative research is particularly useful when the researcher wants to measure variables, compare groups or test hypotheses statistically. Qualitative research is useful when the objective is to understand why people behave in a particular way, how they experience an educational situation or what meanings they attach to an event. The two approaches can also be combined in mixed-methods research, in which numerical evidence and descriptive information are used together to investigate a problem more comprehensively.

Types of Data

Data can be classified on several bases, including their source of collection, nature, measurement scale, continuity and mode of collection or time dimension. These classifications serve different purposes, and a single dataset may belong to several categories simultaneously.

1. Based on the Source of Collection:

Data are classified as primary or secondary according to whether the researcher collects them directly for the present investigation or obtains them from existing sources.

Primary data provide direct information related to the research problem and allow the researcher greater control over the collection procedure. However, collecting them may require considerable time, money and effort.

Secondary data are often economical and readily available. Their usefulness depends on their accuracy, relevance, date of collection, original purpose and suitability for the present study. For example, a researcher investigating school dropout rates may collect primary data through interviews with students or use secondary data from government education reports.

2. Based on the Nature or Form of Data:

According to their form, data are classified into quantitative and qualitative data.

Quantitative data can be further classified into discrete and continuous data. Qualitative data may include nominal categories, such as subject stream, as well as descriptive information collected through interviews and observations. The distinction between qualitative and quantitative data depends on the form and analytical purpose of the information.

3. Based on the Measurement Scale:

The classification of measurement scales was developed by S. S. Stevens (1946). He identified four major scales: nominal, ordinal, interval and ratio. These scales differ in the mathematical properties of the information they represent and determine which statistical operations are appropriate.

• The nominal scale is the simplest measurement scale because it identifies categories without ranking them. The ordinal scale adds ranking, but the difference between two ranks cannot automatically be treated as equal to the difference between two other ranks.
• The interval scale permits meaningful comparisons of differences. For example, the difference between 20°C and 30°C is equal to the difference between 30°C and 40°C, but 40°C is not twice as hot as 20°C in the ratio-scale sense.
• The ratio scale permits comparisons involving both differences and ratios because it possesses a true zero. For example, a class with 40 students has twice as many students as a class with 20 students.
• Important distinction: Examination marks are often treated as numerical or interval-like data in educational analysis, but their precise measurement properties depend on the assessment design. A score of zero does not necessarily mean the complete absence of the underlying ability.

4. Based on Continuity:

According to the possible values they can take, numerical data are classified as discrete and continuous.

• Discrete data arise mainly from counting, whereas continuous data arise mainly from measurement. For example, the number of students in a class may be 40 or 41, but it cannot ordinarily be 40.5. A student's height, however, may be recorded as 160 cm, 160.5 cm or 160.55 cm, depending on the measuring instrument.
• Achievement scores require careful classification. The number of marks awarded in a test may be discrete when only whole marks are permitted. A calculated or standardised achievement score may contain decimal values, but the presence of decimals alone does not make a variable genuinely continuous.

5. Based on the Mode of Collection and Time Dimension:

Data can also be classified according to when they are collected and whether observations are repeated over time. Cross-sectional and longitudinal data describe the temporal design of a study, whereas time-series data are organised as observations recorded at successive time points.

• Cross-sectional research is useful for examining the existing characteristics of a population and comparing different groups at a particular time. For example, a researcher may compare the digital learning skills of teachers from rural and urban schools in the same year.
• Longitudinal research examines change over time and may use repeated observations of the same individuals, schools or groups. It is useful for studying student development, the long-term effects of educational interventions and changes in learning outcomes. However, longitudinal studies may face difficulties such as participant dropout, changing measurement procedures and higher costs.
• Time-series data are particularly useful for identifying long-term trends, seasonal variations and changes following a policy intervention. For example, annual enrolment figures can reveal whether participation in school education is increasing or decreasing over a decade. Unlike a simple cross-sectional dataset, a time series explicitly preserves the chronological order of observations.
• A distinction is important: longitudinal data commonly follow the same subjects or units over time, whereas time-series data focus on the successive values of a variable. Longitudinal data may be analysed as time series when the observations have an appropriate temporal structure.