What is the primary goal of statistics?

To understand variability and draw conclusions from data

Introduction to the Practice of Statistics

Statistics is the science of collecting, organizing, summarizing, and analyzing information to draw conclusions or answer questions, with a focus on understanding variability.

Core Principles

Statistics involves collecting, organizing, summarizing, and analyzing data to draw conclusions.
Data describes characteristics of individuals and inherently possesses variability.
The goal of statistics is to describe and understand sources of variability.
The process of statistics includes identifying the research objective, collecting data, describing data, and performing inference.
A population is the entire group of individuals studied, while a sample is a subset of that population.
Descriptive statistics summarize data using numerical summaries, tables, and graphs, while inferential statistics use sample results to make conclusions about the population.

Key Terms

Statistics: The science of collecting, organizing, summarizing, and analyzing information to draw conclusions or answer questions.
Data: A fact or proposition used to draw a conclusion or make a decision.
Variability: The characteristic of data that causes it to differ among individuals or over time.
Population: The entire group of individuals to be studied.
Individual: A person or object that is a member of the population being studied.
Sample: A subset of the population that is being studied.
Descriptive Statistics: Organizing and summarizing data through numerical summaries, tables, and graphs.
Statistic: A numerical summary based on a sample.
Inferential Statistics: Methods that use sample results to extend conclusions to the population and measure reliability.
Parameter: A numerical summary of a population.
Qualitative Variable: Allows for classification based on an attribute or characteristic.
Quantitative Variable: Provides numerical measures that can be added or subtracted for meaningful results.
Discrete Variable: Has a finite or countable number of possible values; values result from counting.
Continuous Variable: Has an infinite number of possible values that are not countable; values can take on any value within an interval.
Nominal Level: Values are names, labels, or categories; no inherent order.
Ordinal Level: Values can be ranked or ordered, but the differences between values may not be meaningful.
Interval Level: Ordered values with meaningful differences; zero does not mean absence of quantity.
Ratio Level: Ordered values with meaningful differences; zero means absence of quantity; ratios are meaningful.
Observational Study: Measures variables without attempting to influence them.
Designed Experiment: Randomly assigns individuals to groups and manipulates an explanatory variable.
Confounding Variable: An explanatory variable whose effect cannot be separated from another explanatory variable.
Lurking Variable: An unconsidered explanatory variable that affects the response variable.
Cross-sectional Study: Collects data at a specific point in time or over a very short period.
Case-control Study: Retrospective study matching individuals with a characteristic to those without.
Cohort Study: Observes a group over a long period, collecting data over time (prospective).
Random Sampling: Process of using chance to select individuals for a sample.
Simple Random Sample: Every possible sample of size n has an equally likely chance of occurring.
Stratified Sample: Population divided into strata; simple random sample taken from each stratum.
Systematic Sample: Selecting every kth individual from a population.
Cluster Sample: Randomly selects groups (clusters) and surveys all individuals within selected clusters.
Convenience Sample: Individuals are easily obtained; not based on randomness.
Sampling Bias: Sampling technique favors one part of the population over another.
Undercoverage: A segment of the population is lower in the sample than in the population.
Nonresponse Bias: Respondents have different opinions than non-respondents.
Response Bias: Answers do not reflect true feelings (interviewer error, wording, etc.).
Nonsampling Errors: Errors from sampling bias, nonresponse, response bias, or data-entry error.
Sampling Error: Error from using a sample to estimate population information.
Experiment: Controlled study to determine the effect of explanatory variables on a response variable.
Experimental Unit: Person, object, or item upon which a treatment is applied.
Treatment: Combination of factor values applied to experimental units.
Control Group: Baseline treatment for comparison.
Placebo: Innocuous treatment resembling the experimental treatment.
Placebo Effect: Improvement due to belief in the treatment, not the treatment itself.
Blinding: Nondisclosure of treatment to experimental unit or researcher.
Single-Blind: Experimental unit does not know the treatment.
Double-Blind: Neither experimental unit nor researcher knows the treatment.
Completely Randomized Design: Each experimental unit is randomly assigned to a treatment.
Matched-Pairs Design: Experimental units are paired; two levels of treatment.
Null Hypothesis (H0): A statement to be tested, assumed true until evidence indicates otherwise.
Alternative Hypothesis (H1): A statement evidence is sought to support.
Type I Error: Rejecting the null hypothesis when it is true.
Type II Error: Not rejecting the null hypothesis when it is false.
Level of Significance (α): Probability of making a Type I error.
P-value: Probability of observing a statistic as extreme or more extreme than the one observed, assuming H0 is true.
Statistically Significant: Observed results are unlikely under the null hypothesis.
Simulation: Recreating a random event to measure how often a goal is observed.
Law of Large Numbers: As repetitions increase, observed proportion approaches probability.
Random Process: Scenario with uncertain outcomes, but predictable long-term proportions.
Probability: Measure of the likelihood of a random phenomenon occurring.
Sample Space (S): Collection of all possible outcomes of a probability experiment.
Event (E): Any collection of outcomes from a probability experiment.
Disjoint Events: Events with no outcomes in common (mutually exclusive).
Independent Events: Occurrence of one event does not affect the probability of another.
Conditional Probability (P(E|F)): Probability of event E occurring given event F has occurred.
General Multiplication Rule: P(E and F) = P(E) * P(F|E) or P(F) * P(E|F).
Complement Rule: P(E') = 1 - P(E).
Multiplication Rule for Independent Events: P(E and F) = P(E) * P(F).
Counting Principle: Multiplying the number of selections for each step to find total possibilities.
Permutation (nPr): Ordered arrangement, no repetition, r objects from n.
Combination (nCr): Unordered selection, no repetition, r objects from n.
Factorial (n!): Product of an integer and all positive integers below it.
Normal Distribution: Bell-shaped, symmetric probability distribution.
Z-score: Distance of a data value from the mean in standard deviations.
Percentile: Value below which a certain percentage of observations fall.
Quartiles: Values dividing data into four equal parts (Q1, Median, Q3).
Interquartile Range (IQR): Range of the middle 50% of data (Q3 - Q1).
Outlier: Extreme observation, typically outside fences (Q1 - 1.5*IQR) and (Q3 + 1.5*IQR).
Five-Number Summary: Minimum, Q1, Median, Q3, Maximum.
Boxplot: Graphical representation of the five-number summary.
Mean: Average of data values.
Median: Middle value when data is ordered.
Mode: Most frequent data value.
Dispersion: Degree to which data are spread out.
Range: Largest data value minus smallest data value.
Standard Deviation: Average distance of data values from the mean.
Variance: Square of the standard deviation.
Empirical Rule: Describes data within 1, 2, or 3 standard deviations of the mean for bell-shaped distributions.
Chebyshev's Inequality: Provides a lower bound for the proportion of data within k standard deviations of the mean for any distribution.
Weighted Mean: Mean calculated with values having different weights.
Grouped Data: Data summarized in frequency distributions (classes/intervals).
Contingency Table: Two-way table showing frequencies for two categorical variables.
Marginal Distribution: Frequency or relative frequency distribution of a single variable in a contingency table.
Conditional Distribution: Relative frequency of a response variable category given a specific explanatory variable category.
Simpson's Paradox: Association between variables inverts or disappears when a third variable is introduced.
Time-Series Plot: Graph plotting data over time, showing trends.
Misleading Graph: Graph that deceives due to scale, inconsistent increments, or misplaced origin.

Pitfalls to Avoid

Convenience samples can lead to meaningless results.
Observational studies can only show association, not causation.
Confounding variables can distort the relationship between explanatory and response variables.
Lurking variables can influence the outcome of a study.
Data-entry errors can lead to non-representative results.

Real World Examples

Study of high school student sleeping patterns: Investigating the association between school start times and sleep duration.
Flu shots benefit seniors study: Determining if flu shots reduce hospitalization and mortality rates.
Xylitol effectiveness in preventing dental caries: Measuring children's preference for milk flavored with xylitol.
Testosterone in Athletes study: Investigating if moral reasoning among journalists has declined.
Drive-thru ticket price analysis: Assessing if a new process increases average bill per car.
Quality control in bottle filling: Checking if the standard deviation of fill volumes exceeds the target.
Valentine's Day gift spending: Estimating the average amount spent on gifts.
Employee travel times: Calculating mean, median, and standard deviation of commute times.
Waist circumference of 2-year-old males: Using the Empirical Rule and Chebyshev's Inequality to describe data spread.
Used Toyota Prius prices: Approximating the mean and standard deviation from grouped price data.
Unemployment rates by state: Organizing continuous data into frequency and relative frequency distributions.
Death rates by ethnicity: Using time-series plots to identify trends in mortality.
Birth rates over time: Constructing a time-series plot to visualize trends in birth rates.
Misleading graphs: Identifying and correcting graphical misrepresentations.

Statistics

Sample size for study on sleep patterns: 383 adolescents
Mean age of adolescents in sleep study: 15.5 years
Standard deviation of age in sleep study: 0.6 year
Longer sleep duration for later start times: 27.5 minutes
Additional sleep per hour delay in start time: 21 minutes

Quiz

What is the primary goal of statistics?: To understand variability and draw conclusions from data
Which of the following is NOT a step in the process of statistics?: Eliminate all variability
A variable that provides numerical measures and can be added or subtracted meaningfully is called:: Quantitative