DATA COLLECTION IN STATISTICS
Objectives and Scope:
Clearly define objectives and scope of the enquiry.
Determines data to be collected and techniques.
Eliminates irrelevant data and omissions.
Matches with available resources.
Statistical Units:
Well-defined object or group for measurements.
Should be unambiguous, specific, stable over time/place.
Appropriate and uniform for the enquiry.
Avoid irrelevant data or omissions.
Sources of Data:
Primary data from surveys, experiments.
Secondary data from published sources.
Balance cost, time, accuracy.
Method of Collection:
Census vs sampling methods.
Mail surveys, personal interviews, schedules.
Skill and training of investigators.
Accuracy:
Desired confidence interval for estimates.
Balance cost and accuracy.
Type of Enquiry:
Exploratory, descriptive, analytical.
Influences data collection and analysis techniques.
In summary, clearly defined objectives, appropriate statistical units,
relevant sources, systematic collection methods, desired accuracy level,
and matching the type of enquiry to techniques are critical prerequisites
before embarking on data collection and analysis in a statistical
investigation.
TYPES OF STATISTICAL UNITS
Units of collection are of two types:
Units of Enumeration,- The basic unit on which observations are
made, like a person, household, farm, etc. Should be clearly defined,
e.g. what constitutes a "household".
Units of Recording,- The units in which data are recorded, like
kilograms, meters, rupees, etc. Can be simple units like kilogram or
meter, or composite units like skilled worker, ton-kilometer, etc.
Composite units need clear definitions.
Units of Analysis and Interpretation - The units used to analyze and
interpret data, like rates, ratios, percentages and coefficients. Used to
make comparisons between different sets of data. Relative figures like
coefficients allow comparison independent of units of measurement.
Sources of Information
Primary Data - Collected firsthand by the investigator for that specific
inquiry.
Secondary Data - Data collected by others and published, like
government publications, magazines, etc. Must be carefully scrutinized for
reliability and suitability before use.
Method of Data Collection
Census - 100% enumeration of the population.
Sample - Study a representative subset or fraction of the population.
Used when census is impractical due to large population size, time, cost
constraints, or destructive testing needed. Requires determining sample
size and sampling technique.
Degree of Accuracy
Perfect accuracy is impossible and unnecessary. Reasonable
precision is enough for valid inferences.
Precision needed depends on the objectives and scope of the
inquiry.
Information from previous similar studies can guide expected
precision.
Types of Inquiries
Official, Semi-Official, Unofficial,- Affects facilities available and
compulsions to provide information.
Initial, Repetitive,- Repetitive builds on previous, initial is new.
Confidential, Non-Confidential,- Confidential results not made
public.
Direct, Indirect,- Direct measures quantitative data, indirect
makes qualitative data quantitative.
Regular, Ad-hoc,- Regular at fixed intervals, ad-hoc as needed.
METHODS OF COLLECTING PRIMARY DATA
Direct Personal Investigation
Investigator collects data personally through inquiries and
interviews.
Suited for intensive local studies, not extensive surveys.
Provides original, reliable data but limited in scope due to time, cost,
subjectivity.
Indirect Oral Investigation
Data collected through interviews by appointed enumerators.
Less expensive than direct investigation.
Reliability depends on enumerators' skills and integrity.
Information from Local Agencies
Local correspondents collect and submit information periodically.
Inexpensive but reliability is low due to correspondents' biases.
Useful for rough estimates where high accuracy not needed.
Mailed Questionnaire Method
Questionnaires mailed to respondents to fill and return.
Economical for extensive surveys but suffers from low response
rate.
No way to verify accuracy of responses.
Schedules through Enumerators
Enumerators visit respondents with schedule of questions.
Gives better response rate and reliability than questionnaires.
Expensive but can use for illiterates. Success depends on
enumerators.
The choice depends on the survey scope, resources available, and
accuracy needed. Drafting the questionnaire/schedule is critical for
success.
