DATA COLLECTION

By:
1-Mohamed Khater
Abdalla Elhendawy-2
Rodayna Mohsen-3
 WHAT IS THE DIFFERENCE BETWEEN
DATA AND INFORMATION?

• Data: Raw numbers or facts

Example: 78, 64, 36, 70, 52
• Information: Processed data providing useful knowledge
Example: Average exam mark of five students is 60%
• Principle: Data consists of raw numbers, while information provides useful
knowledge.
 IMPORTANCE OF DATA
COLLECTION

• Managers need relevant information for decision-making.

• Three main steps in preparing information:
1-Data collection
2-Processing to give information
3-Presentation
• Figure 4.1: Highlights the process from data to information presentation.
• Without proper data collection, managers can't make informed decisions.
Figure 4.1: Highlights the process from data to information presentation.
Without proper data collection, managers can't make informed decisions.
 STEPS IN DATA COLLECTION

• Define Purpose:
Purpose of the data and its usage.
Examples: Weather forecasts, company information, city’s history, etc.
• Decide Needed Data:
Determine which data is necessary to achieve the purpose.
• Design & Collect Data:
Plan and execute the data collection process.
Figure 4.2: Illustrates the steps in data collection.
 AMOUNT OF DATA & COLLECTION

• Amount of Data: Balance between enough data for good decisions and not
being swamped by irrelevant details.
• Marginal Cost: Extra cost of collecting one more piece of data.
• Marginal Benefit: Value from the last bit of collected data.
• Optimal Amount: Point where marginal cost equals marginal benefit (as
shown in Figure 4.3).
• Time consideration in data collection: Fast vs. thorough methods.
• Importance of accurate data and careful planning.
Figure 4.3 Finding the amount of data to collect
 CLASSIFYING DATA
• Nominal Data
Categories with no meaningful units.
Examples: Occupation, type of economy, color.
Analysis: Count observations in distinct categories.
• Ordinal Data
Categories with meaningful order.
Examples: Small, medium, large; exam results.
Analysis: Ranking categories in order of importance.
 CLASSIFYING DATA
• Cardinal Data
Measurable attributes.
Discrete: Integer values (e.g., number of children).
Continuous: Any value (e.g., weight of a biscuit bag).
• Mismatch in Data Types
Feet lengths vs. shoe sizes.
Heights vs. described heights (nearest cm or inch).
 PRIMARY VS. SECONDARY DATA
• Primary Data
New data collected for a specific purpose.
Benefits: Exact fit, up-to-date, reliable.
• Secondary Data
Existing data collected for other purposes.
Benefits: Cheaper, faster, easier.
• Advantages of Secondary Data
Impartial sources: Governments, universities, industry bodies.
 CHOOSING BETWEEN PRIMARY
AND SECONDARY DATA
• Primary Data
Worth collecting for major decisions (e.g., new product launch).
• Secondary Data
Suitable for broader issues (e.g., future economic conditions).
• Combination of Primary and Secondary Data
Secondary data for the overall picture.
Primary data for detailed insights.
 USING SAMPLES TO COLLECT DATA

• Primary Data Collection

When no suitable secondary data is available.
• Population
Statistical sense: All items/people with a common characteristic.
Examples: Post Office First-class letters, Whirlpool dishwashers, potential
customers for a new toy game.
 IDENTIFYING THE RIGHT
POPULATION
• Importance
Mistake in defining the population renders data collection and analysis pointless.
• Challenges
Ambiguities in defining the population.
Example: Student opinion survey - Full-time vs. part-time, distance-learning,
work experience, etc.
Example: 'Imported cars' - Components vs. assembly, almost-finished vs.
finished cars.
 SAMPLING FRAMES
• Definition: Complete list of a population.
• Common Sources
Electoral registers, memberships (e.g., AA), employee lists, customer loyalty cards, website
addresses, credit rating agencies.
• Difficulties
Obtaining lists for specific populations.
Example: People who bought an imported TV in the last 5 years, supermarket shoppers,
people who caught a cold last winter.
• Limitations with non-human populations.
Example: Basking sharks visiting the Cornish coast.
 CHALLENGES WITH SAMPLING
• Size of Population
Example: 'Households using electricity in Germany' has over 60 million entries.
• Two Alternatives
Census: Collecting data from every member of a population.
Sample: Collecting data from a subset (sample) of the population or Data from a
representative sample .
Census: More accurate but time-consuming and expensive.
Sampling: Easier, cheaper, faster but less accurate.
TYPES OF SAMPLES
 RANDOM SAMPLING
• Definition: Every member has an equal chance of being selected.
• Example: Choosing one football player out of 11.
• Characteristics
Not haphazard but systematic.
• Bias: When a sample does not reflect the population.
 SYSTEMATIC SAMPLING
• Definition: Data collected at regular intervals.
• Example: Interview every tenth person.
• Challenges
Not truly random.
Regularity introduces bias.
 STRATIFIED SAMPLING
• Definition: Population divided into distinct groups (strata).
• Objective: Ensure each stratum is represented in the sample.
• Example: 60% women and 40% men in a company.
• Challenges with Stratified Sampling
Small Groups: Small strata might be overlooked.
Solution: Include all strata, even if they are small.
Result: Biased but inclusive sample.
 QUOTA SAMPLING
• Definition: Specifying characteristics to match population.
• Objective: Exact proportions of population characteristics in the sample.
• Process: Interviewers fill quotas for specific characteristics.
• Characteristics: Not completely random due to quotas.
 MULTI-STAGE SAMPLING
• Objective: Sample confined to smaller geographical areas.
• Process
Dividing country into regions.
Selecting random subdivisions within regions.
Identifying sample individuals from smaller areas.
• Example: Sampling magazine subscribers geographically.
 CLUSTER SAMPLING
• Definition: Sampling clusters or groups rather than individuals.
• Benefits
Cost-effective.
Easy for certain populations or industries.
• Example: Surveying people working in a specific industry using companies as
clusters.
 METHODS OF DATA COLLECTION
• Observation: Directly watching and recording events.
Can be human or automatic.
More reliable but subject to observer errors and machine faults.
• Questionnaires: Asking people for data.
Can gauge awareness, feelings, views, reasons, and intensity.
Problematic due to potential bias and emotional responses.
 METHODS OF ADMINISTERING
QUESTIONNAIRES
• Personal Interviews
Detailed and reliable but costly.
Response rates are generally high.
• Telephone Interviews
Cost-effective but can introduce bias.
Convenient but lacks visual cues.
 METHODS OF ADMINISTERING
QUESTIONNAIRES
• Postal Surveys
Cheap but low response rates.
Limited opportunity for observation or clarification.
• Email Surveys
Wide reach but prone to bias.
Affected by spam filters and selective participation.
 SURVEY METHODS
• Panel Surveys: Monitoring opinions over time.
Can become biased with changing panel views.
Expensive and difficult to administer.
• Longitudinal Surveys: Monitoring over long periods.
Expensive and resource-intensive.
Studies often focused on sociological and health changes.
 EFFECTIVE QUESTIONNAIRES
• Logical sequence: Follow a logical flow of related questions.
• Keep it short: Avoid long and poorly presented questionnaires.
• Clarity: Use simple and unambiguous language.
• Avoid Bias: Carefully phrase questions to avoid leading or loaded wording.
 CHALLENGES WITH NON-RESPONSES
• Reasons: Language barriers, absence, refusal, etc.
• Implications: Non-responses can introduce bias.
• Follow-up: It's crucial to follow up with non-respondents to reduce bias.
 CONCLUSION

• Data collection methods vary in cost, reliability, and reach.

• Designing a questionnaire requires careful consideration to avoid bias and
gather accurate data.
• Handling non-responses is essential to ensure the sample accurately represents
the population.