AP Statistics Comprehensive Multiple Choice and Review Notes

Descriptive Statistics and Data Representation

• Range and Interquartile Range (IQR):     * The range of a sample data set is defined as the difference between the maximum and minimum values. It can never be greater than the range of the population from which it is drawn.     * The Interquartile Range ($IQR$) is the distance between the first quartile ($Q_1$) and the third quartile ($Q_3$). Specifically, $IQR = Q_3 - Q_1$. It represents the spread of the middle 50% of the data.     * Effect of Outliers: Outliers significantly affect the range but do not affect the $IQR$, making the $IQR$ a more robust measure of spread for skewed distributions.

• Distribution Shape and Measures of Center:     * Symmetry: A distribution is symmetrical if the left and right sides of its histogram are mirror images. In perfectly symmetrical distributions, the mean equals the median.     * Skewed Right (Positive Skew): The distribution has a long tail extending toward the higher values on the right. In this case, the mean is typically greater than the median ( ext{Mean} > ext{Median}).     * Skewed Left (Negative Skew): The distribution has a long tail extending toward the lower values on the left. Here, the mean is typically less than the median ( ext{Mean} < ext{Median}).     * Unimodality: Symmetrical distributions are not always unimodal; they can be bimodal or uniform.

• Normal Curves and Bell-Shaped Distributions:     * All normal curves are bell-shaped and symmetrical, but not all bell-shaped, symmetrical curves are normal.

• Graphical Displays for Univariate Data:     * Stemplots (Stem-and-Leaf Plots): These are used for quantitative univariate data. They show actual data values and are effective for identifying symmetry, gaps, clusters, and outliers.     * Histograms: These display individual frequencies or relative frequencies but do not show actual data values.     * Boxplots: These summarize data using the five-number summary: Minimum, $Q_1$, Median, $Q_3$, and Maximum. Boxplots are excellent for comparing distributions and identifying outliers (using the $1.5 imes IQR$ rule).

• Outliers:     * An outlier is an observation that is significantly different from the other observations in the data set.

• Mathematical Operations on Data Sets:     * Addition/Subtraction: Adding a constant $c$ to every value in a data set increases the mean, median, and quartiles by $c$, but the standard deviation, range, and $IQR$ remain unchanged.     * Multiplication/Division: Multiplying every value by a constant $k$ multiplies the mean, median, standard deviation, range, and $IQR$ by $k$. The variance is multiplied by $k^2$.

Probability Rules and Calculations

• Foundational Rules:     * The probability of any event $A$ must satisfy 0 gle P(A) gle 1.     * The sum of probabilities for all outcomes in a sample space must equal 1.     * Complement Rule: The probability that an event does not occur is $P(A^c) = 1 - P(A)$.

• Addition Rule:     * For any two events $A$ and $B$, the probability that $A$ or $B$ occurs is P(A gcup B) = P(A) + P(B) - P(A gcap B).     * If events are mutually exclusive (disjoint), then P(A gcap B) = 0, so P(A gcup B) = P(A) + P(B).

• Multiplication Rule and Independence:     * Two events $A$ and $B$ are independent if the occurrence of one does not affect the probability of the other.     * If independent, P(A gcap B) = P(A) gtimes P(B).     * To check for independence, verify if $P(A|B) = P(A)$ or if P(A gcap B) = P(A) gtimes P(B).

• Conditional Probability:     * The probability of event $A$ given that event $B$ has occurred is P(A|B) = gfrac{P(A gcap B)}{P(B)}. In simpler terms, this is gfrac{P( ext{and})}{P( ext{given})}.

• Special Cases:     * At least one: $P( ext{at least one}) = 1 - P( ext{none})$.     * Mutually Exclusive vs. Independent: Mutually exclusive events can never be independent (except if one has a probability of 0) because if one occurs, the probability of the other occurring becomes zero.

Experimental Design Principles

• Four Principles of Experimental Design:     1. Control: Managing extraneous variables to prevent confounding and to provide a baseline for comparison (often using a control group or placebo).     2. Randomization: Using chance to assign experimental units to treatments to reduce bias and balance the effects of lurking variables.     3. Replication: Repeating the experiment on many subjects or experimental units to reduce the role of chance variation and increase the reliability of results.     4. Blocking: Grouping similar experimental units together to reduce variability within the treatment groups. Blocking is to experimental design what stratification is to sampling design.

• Key Terminology:     * Explanatory Variable (Factor): The variable manipulated by the researcher.     * Response Variable: The outcome measured after treatment.     * Confounding: Occurs when the effects of two variables on a response variable cannot be distinguished from each other.     * Placebo Effect: The phenomenon where subjects show improvement simply because they believe they are receiving a treatment.     * Blinding: Single-blind is when the subjects do not know which treatment they receive; Double-blind is when neither the subjects nor the researchers evaluating them know who received which treatment.     * Matched Pairs Design: A special type of randomized block design where subjects are either paired based on similar characteristics or serve as their own control (before/after measurements).

Sampling Design and Bias

• Reliable Sampling Methods:     * Simple Random Sample (SRS): A sample of size $n$ where every possible group of size $n$ has an equal chance of being selected.     * Stratified Random Sample: The population is divided into homogeneous groups (strata), and an SRS is taken from each stratum.     * Cluster Sample: The population is divided into heterogeneous groups (clusters). Entire clusters are randomly selected, and every member of the chosen clusters is surveyed.     * Systematic Random Sample: Participants are selected according to a random starting point and a fixed, periodic interval (e.g., every $k$-th person).

• Sources of Bias:     * Undercoverage: Some groups in the population are left out of the process of choosing the sample.     * Nonresponse Bias: Occurs when a selected individual cannot be contacted or refuses to participate.     * Response Bias: Occurs when participants give inaccurate answers, often due to the wording of the question, interviewer behavior, or sensitive subject matter.     * Voluntary Response Bias: Occurs when people choose themselves by responding to a general appeal (e.g., call-in polls).     * Convenience Sampling: Choosing individuals who are easiest to reach; this almost always produces biased data.

Random Variables and Probability Distributions

• Expected Value (Mean of a Random Variable):     * Calculated as the weighted average: E(X) = gmu_X = gsum x_i gtimes P(x_i).     * Variance of a Random Variable: gsigma^2_X = gsum (x_i - gmu_X)^2 gtimes P(x_i).

• Binomial Distribution:     * Used when there is a fixed number of trials ($n$), only two possible outcomes (success/failure), trials are independent, and the probability of success ($p$) is constant.     * Mean: gmu = n gtimes p     * Standard Deviation: gsigma = gsqrt{n gtimes p gtimes (1 - p)}

• Geometric Distribution:     * Used when trials are conducted until the first success occurs.     * Probability of success on the $k$-th trial: P(X = k) = (1 - p)^{k-1} gtimes p.     * Mean: gmu = gfrac{1}{p}

• Uniform Distribution:     * A continuous distribution where all outcomes in a range from $a$ to $b$ are equally likely. The total area under the density curve is 1.

Normal Distributions and Sampling Distributions

• The Normal Distribution:     * Notation: X gsim N( gmu, gsigma).     * Empirical Rule (68-95-99.7): Approximately 64% of data falls within 1 gsigma, 95% within 2 gsigma, and 99.7% within 3 gsigma of the mean.     * Z-Scores: Standardizing a value to represent how many standard deviations it is from the mean: z = gfrac{x - gmu}{ gsigma}.

• Sampling Distributions:     * The distribution of a statistic (like gbar{x} or ghat{p}) across all possible samples of the same size from the same population.     * Central Limit Theorem (CLT): For a large sample size (usually n gge 30), the sampling distribution of the sample mean gbar{x} is approximately normal, regardless of the shape of the population distribution.     * Parameters for means: gmu_{ gbar{x}} = gmu; gsigma_{ gbar{x}} = gfrac{ gsigma}{ gsqrt{n}}.     * Parameters for proportions: gmu_{ ghat{p}} = p; gsigma_{ ghat{p}} = gsqrt{ gfrac{p(1-p)}{n}}.

Statistical Inference: Confidence Intervals and Hypothesis Testing

• Confidence Intervals:     * Standard form: ext{Estimate} gpm ext{Margin of Error}, where ext{Margin of Error} = ext{Critical Value} gtimes ext{Standard Error}.     * Increasing sample size decreases the width of the interval. Increasing the confidence level increases the width.

• Hypothesis Testing:     * Null Hypothesis ($H_0$): The claim being tested, usually a statement of "no effect."     * Alternative Hypothesis ($H_a$): The claim we hope to find evidence for (greater than, less than, or different from).     * P-value: The probability of obtaining results as extreme as, or more extreme than, the observed results, assuming $H_0$ is true.     * Significance Level ( galpha): The threshold for rejecting $H_0$. If P gle galpha, reject $H_0$.

• Errors in Inference:     * Type I Error: Rejecting a true null hypothesis (P( ext{Type I}) = galpha).     * Type II Error: Failing to reject a false null hypothesis (P( ext{Type II}) = gbeta).     * Power: The probability of correctly rejecting a false null hypothesis ( ext{Power} = 1 - gbeta). Power increases with larger sample size, larger distance between null and true values, or a higher galpha.

Linear Regression

• Least Squares Regression Line (LSRL):     * Equation: ghat{y} = b_0 + b_1x.     * Slope ($b_1$): The estimated change in the response variable $y$ for every one-unit increase in the explanatory variable $x$.     * Intercept ($b_0$): The estimated value of $y$ when $x = 0$.

• Correlation ($r$):     * Measures the strength and direction of a linear relationship between two quantitative variables. -1 gle r gle 1.

• Coefficient of Determination ($R^2$):     * The percentage of variation in the response variable $y$ that can be explained by the linear relationship with the explanatory variable $x$.

• Residuals:     * ext{Residual} = ext{Observed } y - ext{Predicted } ghat{y}. A good linear fit has a residual plot with no discernible pattern and points scattered randomly around the horizontal axis ($y=0$).