Collecting and Classifying Data: Types, Sources, and Collection Methodologies
Classification of Data
Numerical Data (Quantitative Data):
Defined as data that consists of numbers and represents measurable quantities or counts.
Supports arithmetic operations such as addition, subtraction, multiplication, and calculating summary statistics such as the mean () and standard deviation ().
Divided fundamentally into two distinct subcategories: discrete numerical data and continuous numerical data.
Examples include height, weight, test scores, population counts, and temperatures.
Discrete Numerical Data:
Represents numerical values obtained through counting and is restricted to specific, distinct values along a numerical scale (typically whole numbers or integers, ).
Values cannot be subdivided into meaningfully smaller increments between two consecutive points; there are gaps between possible values.
Examples include:
Number of students present in a classroom (e.g., , ; a class cannot have students).
Number of cars passing through an intersection in an hour.
Number of goals scored in a soccer match.
Standard shoe sizes (e.g., sizes , , ).
Continuous Numerical Data:
Represents numerical values obtained through measurement and can theoretically take any real numerical value within a given interval or range ().
The precision of the value is constrained solely by the precision of the measuring instrument used.
Examples include:
Height of an individual (e.g., , ).
Time elapsed during a race (e.g., , ).
Ambient air temperature (e.g., ).
Mass of a chemical substance (e.g., ).
Categorical Data (Qualitative Data):
Defined as data that consists of non-numerical values or labels that place observations into descriptive categories, groups, or qualities.
Mathematical operations such as calculating a meaningful average are not applicable to categorical data.
Divided into two primary types:
Nominal Data: Categories that possess no natural, inherent, or meaningful order or ranking (e.g., eye color, blood type, gender, country of birth).
Ordinal Data: Categories that have a logical, intrinsic hierarchy or ranking, though the mathematical differences between intervals are not necessarily uniform or measurable (e.g., customer satisfaction ratings such as poor, neutral, good; education levels; survey responses on a Likert scale).
Sources of Data
Primary Source (Primary Data):
Refers to firsthand raw data collected directly by the researcher, investigator, or organization specifically for the original purpose or investigation at hand.
Methods of collection include conducting experiments, issuing questionnaires, interviewing subjects, direct field measurements, or running observational studies.
Key Advantages:
Highly tailored to the specific hypotheses and questions of the study.
Complete oversight over data quality, accuracy, operational definitions, and collection methodology.
Current and up to date.
Key Disadvantages:
Often expensive to plan, organize, and administer.
Time-consuming and resource-intensive relative to secondary data acquisition.
Secondary Source (Secondary Data):
Refers to existing data that was collected, compiled, and published by an individual, organization, or agency other than the primary investigator, originally for a different purpose.
Typical sources include government censuses, academic journal publications, public datasets, organizational financial records, and historical archives.
Key Advantages:
Fast, readily accessible, and cost-effective.
Enables access to large-scale, longitudinal, or nationwide data that an individual researcher could not feasible generate independently.
Key Disadvantages:
May not perfectly align with the specific variables, definitions, or needs of the research question.
Potential inaccuracies, outdated information, or methodological biases of the original collectors that cannot be controlled or verified.
Data Collection Scope and Methodologies
Census:
A comprehensive data collection process that gathers information from every single member, item, or unit comprising the entire population under investigation.
Key Characteristics:
Provides complete parameter values (, ) for the population rather than sample estimates (, ).
Possesses zero sampling error because the entire population is measured ().
Key Disadvantages:
Enormously expensive, complex, and time-consuming to execute.
Often practically unfeasible or impossible if the population is extremely large, dispersed, dynamic, or if the testing process is destructive (e.g., testing the lifespan of light bulbs until failure).
Sample:
A representative subset, slice, or portion of individuals selected from the larger overall population used to make inferences, estimates, and generalizations about the entire population.
Key Characteristics:
Much faster, cheaper, and more feasible than a full census.
Introduces sampling variability and sampling error, which must be quantified through inferential statistics (e.g., confidence intervals, standard error).
Requires sound probabilistic sampling techniques (such as simple random sampling, stratified sampling, or systematic sampling) to avoid selection bias and ensure the sample accurately reflects the demographics and traits of the target population.
Observation (Observational Study):
A data collection method in which an investigator systematically watches, records, monitors, and measures subjects, events, or behaviors as they naturally occur in their environment without direct manipulation, intervention, or experimental control.
Key Characteristics:
Subjects are not assigned to treatment groups or exposed to artificial variables.
Eliminates response biases that can occur in self-reported surveys (e.g., recall bias, social desirability bias).
Useful in ecological studies, consumer behavior tracking, and sociological field research.
Key Limitations:
Cannot establish definitive cause-and-effect relationships due to the presence of uncontrolled confounding variables.
Potential for observer bias, where the observer's subjective interpretations unintentionally influence the recorded data.
May alter natural behavior if subjects realize they are being monitored (the Hawthorne effect).