Week 1 Note's: Data Collection & Sampling (1.1–1.3)

1.1 Introduction to Data Collection

• Course Context (MATH 1401, Week 1): Focuses on Orientation, Practice Assignment, and Academic Dishonesty Quiz in iCollege.

• Goals: Identify individuals and variables, classify variables (categorical/quantitative), identify population/sample, distinguish observational study/experiment.

• What is statistics? It's a scientific discipline focused on working with data: collection, organization, analysis, interpretation, and presentation.

  • Purpose: Make informed decisions based on data.

• Statistical Problem-Solving Process:

  1. Ask Questions

  2. Collect/Consider Data

  3. Analyze Data

  4. Interpret Results

• Core Concepts:

  • Individuals: Entities described in a dataset (e.g., a student, a tree, a car).

  • Variable: An attribute that takes different values for different individuals (e.g., height, major, color).

  • Types of Variables:

  • Categorical: Labels or groups (e.g., a student's major: "Math," "Biology"; eye color: "blue," "brown").

  • Quantitative: Numeric values (quantities/measurements) (e.g., height: $65$ inches; test score: $85$ points).

  • Data Structure: Rows are individuals, columns are variables.

1.2 Populations, Samples, and Study Designs

• Goal: Make inferences about a population.

• Population: The entire group of interest (e.g., all students at a university).

• Census: Data from every individual in the population.

• Sample: A subset of individuals from the population (e.g., $100$ students interviewed from the university).

• Purpose of Sampling: To infer about the population when a census is impractical (due to cost, time).

• Observational Studies vs. Experiments:

  • Observational Study: Observe and measure variables without influencing responses.

  • Examples: Surveying opinions on a new policy; tracking medical histories of patients.

  • Experiment: Deliberately impose treatments to measure a response.

  • Examples: Applying different fertilizers to plants to measure growth; testing different drug dosages.

  • Important: Observational studies describe, but typically can't establish causation. Experiments aim to determine causal effects.

1.2 Sampling: Good and Bad (Bias, Randomness, and Sampling Methods)

• Goals: Describe how convenience and voluntary response sampling lead to bias, explain how random sampling avoids bias, describe other bias sources.

• Bias: Design of a study systematically overestimates or underestimates a quantity of interest.

• Sampling Methods:

  • Volunteer Sampling: Individuals self-select.

  • Example: An online poll where people choose to vote on a housing preference. Leads to bias as motivated individuals respond.

  • Convenience Sampling: Individuals chosen because they are easy to reach.

  • Example: Asking students passing by a dining hall about housing preference. May not represent all students.

• Avoiding Bias: Use random sampling.

  • Random Sample: Individuals selected by a chance process (e.g., drawing names from a hat).

• Types of Bias:

  • Undercoverage: Some population members are less likely to be chosen or excluded (e.g., a phone survey that only calls landlines, ignoring cell-only users).

  • Nonresponse: Chosen individuals cannot be contacted or refuse to participate (e.g., mail-in surveys with low return rates).

  • Response Bias: Inaccurate responses due to survey design or factors (e.g., poorly worded questions, social desirability bias in sensitive topics).

1.3 Simple Random Sampling (SRS) and Sampling Variability

• Simple Random Sample (SRS): Each sample of size $n$ has an equal chance of being selected from the population. Sampling is typically done without replacement.

  • Example 1: Drawing $50$ names randomly from a hat containing all population names.

  • Example 2: Numbering everyone $1$ to $N$, then randomly selecting $n$ numbers.

• Sampling Variability (Sampling Error): Different random samples of the same size yield different estimates due to chance.

  • Relationship with Sample Size ($n$): As $n$ increases, sampling variability decreases, leading to more precise and consistent estimates closer to the population parameter.

  • Intuition: Larger samples give more consistent results.

  • Mathematical Intuition (for mean): Standard Error ($SE( ar{X})$) decreases as $n$ increases; often expressed as $SE( ar{X}) = \frac{\sigma}{\sqrt{n}}$.

1.4 Quick Connections to Practice and Real-World Relevance

• Summary: Statistics helps identify who to study (population), how to collect data (sampling design), and how to interpret results (inference).

• Causality: Observational studies describe but usually can't establish causation. Experiments can establish causation by applying treatments.

• Real-World Impact: Concepts of sampling, bias, and sample size are vital in polls, surveys, market research, and clinical trials.

• Key Formulas/Notations:

  • Population size: $N$

  • Sample size: $n$

  • SRS Principle: Each subset of size $n$ from the population has equal probability.

  • Sampling without replacement: Individual chosen once.

  • Conceptual SE for simple cases: $SE( \bar{X}) = \frac{\text{population standard deviation } \sigma}{\text{(effective) } \sqrt{n}}$

A dot plot is a simple type of data display that shows data points as dots above a number line. It's used to visualize the distribution of a small dataset, especially for quantitative variables. Each dot represents a single observation from the dataset.

In statistics, a bimodal distribution is a distribution with two distinct peaks, indicating two different modes or common values in a dataset.