Population: The entire group of individuals about which we want to gather information
Sample: The subset of the population from which we actually collect data
- Ask: "Who is the researcher interested in studying?" → Population
- Ask: "From whom did the researcher actually collect data?" → Sample
- Identify the numerical measure being calculated
Researcher wants to study "average height of adult males in a city" → All adult males in the city
Researcher collected data from "500 randomly selected men from the city's phone directory" → 500 men from phone directory
The true average height of ALL adult males in the city (unknown value)
The average height of the 500 men measured (calculated from sample data)
Sample: 500 men from phone directory
Parameter: True average height of all adult males (μ)
Statistic: Average height of 500 measured men (x̄)
Population: All adult males in the city
Sample: 500 men randomly selected from the city's phone directory
Parameter: The true average height of all adult males in the city (μ)
Statistic: The average height calculated from the 500 measured men (x̄)
• Parameter: Numerical characteristic of the population (usually unknown)
• Statistic: Numerical characteristic of the sample (known from data)
• Notation: Greek letters for parameters (μ, σ), Roman letters for statistics (x̄, s)
Researchers want to predict election outcome → All eligible voters in the election
Contacting "registered voters" → Database of registered voters
1,200 registered voters who were successfully contacted and responded
Parameter: True proportion of ALL eligible voters supporting Candidate A; Statistic: 65% from sample
Sample: 1,200 registered voters
Parameter: True proportion voting for Candidate A
Statistic: 65% in sample
Population: All eligible voters in the election
Sample: 1,200 registered voters contacted
Parameter: True proportion of all eligible voters who will vote for Candidate A
Statistic: 65% of the sample who said they will vote for Candidate A
Potential Issues: Undercoverage (unregistered voters), non-response bias, timing of poll
• Undercoverage: Excludes unregistered voters
• Non-response bias: Responders may differ from non-responders
• Timing: Opinions may change before election
• Wording: Question phrasing can influence responses
Population Value
Fixed but unknown
Symbol: p (true proportion)
Sample Value
Known but varies
Symbol: p̂ (sample proportion)
This example demonstrates multi-stage sampling where the population is selected in stages: first schools, then students within schools.
"Reading comprehension among grade 8 students" → All grade 8 students in the district
Stage 1: 5 schools randomly selected; Stage 2: 30 students per school randomly selected
5 schools × 30 students = 150 grade 8 students
Parameter: True average reading score for ALL grade 8 students; Statistic: 72.5 from sample
in District
(5 schools × 30)
Sample: 150 students (5 schools × 30 students each)
Parameter: True mean reading score of all grade 8 students
Statistic: 72.5 (mean from sample)
Population: All grade 8 students in the school district
Sample: 150 students (30 randomly selected from each of 5 randomly selected schools)
Parameter: The true mean reading comprehension score of all grade 8 students in the district
Statistic: 72.5, the mean reading comprehension score of the 150 students in the sample
• Cost-effective: Reduces travel and administrative costs
• Practical: More feasible than simple random sampling of all students
• Representative: Can maintain good representation if properly executed
Sampling Frame: The list or source from which we draw our sample
Parameter: A numerical characteristic of the population (μ, σ, p)
Statistic: A numerical characteristic of the sample (x̄, s, p̂)
Sample Size: Number of individuals in the sample (n)
Statistics: Calculated values that describe samples (known from data)
Relationship: Statistics are used to estimate parameters
Medical research often involves complex sampling strategies and ethical considerations that affect population definition.
"Patients who underwent a specific surgery" → All such patients in the region
"3 hospitals in the region" → Hospital records from these 3 hospitals
200 randomly selected patients from the 3 hospitals
Parameter: True average recovery time for all patients; Statistic: 14.2 days from sample
Sample: 200 patients from 3 hospitals
Parameter: True mean recovery time (μ)
Statistic: 14.2 days (x̄)
Population: All patients in the region who underwent the specific surgery in the past year
Sample: 200 randomly selected patients from 3 hospitals in the region
Parameter: The true mean recovery time for all such patients in the region
Statistic: 14.2 days, the mean recovery time for the 200 patients in the sample
Potential Limitations: Undercoverage (other hospitals), selection bias, temporal factors
• Undercoverage: Patients treated at other facilities not included
• Temporal bias: Surgical techniques may have improved over the year
• Health system variation: Different hospitals may have different protocols
• Selection criteria: Hospitals may have different patient populations
Market research often faces significant response bias challenges that affect the validity of conclusions.
"Customer satisfaction" → All customers who purchased the product
"Email to 10,000 customers" → Company's customer database
1,500 customers who responded to the email survey
Parameter: True satisfaction of all customers; Statistic: 4.2 from responders
Sample: 1,500 survey responders
Parameter: True mean satisfaction (μ)
Statistic: 4.2 (x̄)
Response Rate: 15%
Population: All customers who purchased the product in the last month
Sample: 1,500 customers who responded to the email survey (15% response rate)
Parameter: The true mean satisfaction rating of all customers who purchased the product
Statistic: 4.2, the mean satisfaction rating of the 1,500 respondents
Biases Present: Non-response bias (low response rate), voluntary response bias
• Non-response bias: Only 15% responded, responders may differ from non-responders
• Voluntary response bias: Dissatisfied customers may be less likely to respond
• Selection bias: Only customers with email addresses reached
• Recency bias: Recent purchasers may have different perspectives
🔍 Definitions of Key Concepts
⚖️ Core Rules and Principles
⚙️ Step-by-Step Methods for Identification
- Read the entire problem carefully to understand the research objective
- Ask: "About whom/what does the researcher want to learn?" → This identifies the population
- Ask: "From whom/what was data actually collected?" → This identifies the sample
- Look for numerical values: Those calculated from the sample are statistics; those describing the population are parameters
- Consider the sampling method: How was the sample selected from the population?
- Evaluate representativeness: Does the sample accurately reflect the population?
🧪 Examples: From Simple to Advanced
💡 Tips, Tricks, and Common Pitfalls
📝 Key Notes for Memorization
📋 Common Parameter-Statistic Pairs
Mean: μ (mu)
Standard Deviation: σ (sigma)
Proportion: p
Correlation: ρ (rho)
Mean: x̄ (x-bar)
Standard Deviation: s
Proportion: p̂ (p-hat)
Correlation: r
Questions & Answers
Question: I'm having trouble distinguishing between the sampling frame and the population. Can you explain the difference?
Answer: Great question! The distinction is crucial:
- Target Population: The entire group you want to make conclusions about (e.g., all registered voters in a state)
- Sampling Frame: The list or source from which you actually select your sample (e.g., voter registration database)
Analogy: If you want to survey all fish in a lake (population), your sampling frame might be the areas of the lake where fishing is allowed (not all areas of the lake).
Problem: When the sampling frame doesn't match the target population, you get undercoverage bias. For example, using a phone book to survey households excludes those without listed numbers.
Question: How do I know if a number I see in a problem is a parameter or a statistic?
Answer: Use this decision tree:
- Does the number come from data that was actually collected? If yes, it's a statistic.
- Does the number describe the entire population being studied? If yes, it's likely a parameter (though sometimes hypothetical).
- Is it presented as a fact about the population? Parameters are usually unknown and estimated.
Examples: "The average height of the 50 students measured was 68 inches" → statistic (actual data). "The true average height of all students in the school" → parameter (unknown value).
Notation clue: Parameters use Greek letters (μ, σ, π), statistics use Roman letters (x̄, s, p̂).
Question: Can a sample be larger than a population? What happens if it is?
Answer: No, by definition, a sample cannot be larger than the population it's drawn from:
- Definition: A sample is a SUBSET of the population
- Mathematically: Sample size n ≤ Population size N
- Practically: You can't sample more individuals than exist in the group
What if n > N?: This indicates a calculation error or misunderstanding of the population definition. Re-examine the problem to ensure you've correctly identified the population.
Special case: If you somehow managed to measure every individual in the population, that would be a census, not a sample, and the distinction between parameter and statistic would disappear.