Fifty Fathoms Contents logo

Fifty Fathoms is a set of fifty statistics demonstrations. This is a list of them, with a description of the topics each demo addresses. You can also see the names of the files each demo uses. (The files are on the CD that comes with the book.)

Back to the Fifty Fathoms page.


Measures of Center and Spread
 1 The Meaning of Mean The mean • How individual values affect the mean
Meaning of Mean.ftm
 2 Mean and Median Measures of center: mean, median, and midrange • Resistance: what happens to the measures when you move one point
Mean and Median.ftm
 3 What Do Normal Data Look Like? Normally-distributed data • The effect of changing the mean and standard deviation
Normal Data.ftm
 4 Transforming the Mean and Standard Deviation What happens to mean and standard deviation when you add a constant to every value or multiply every value by a constant
Transform Mean and SD.ftm
 5 The Mean is Least Squares, Too Defining the mean as the place where the sum of squares is a minimum (just like the least-squares line) • The median, and what it minimizes
Mean is Least Squares.ftm
Mean is Least Squares 2.ftm

Regression and Correlation
 6 Least Squares Linear Regression Explore the squares in least squares • Minimizing the areas of the squares built on residuals
Least Squares.ftm
 7 Standard Scores Using standard scores to compare unlike scales • Making a scale in terms of standard deviations
Standard Scores.ftm
 8 Devising the Correlation Coefficient How the correlation coefficient measures what it does
Correlation.ftm
 9 Correlation Coefficients of Samples How samples from a correlated population yield different values for the correlation • How sample size affects that sampling distribution
Correlation in Samples.ftm
Correlation in Samples 2.ftm
10 Regression Towards the Mean Regression towards the mean • The meaning—and asymmetry—of the least-squares line
Regression Towards the Mean.ftm

Random Walks and the Binomial Distribution
11 Flipping Coins—the Law of Large Numbers How the proportion of heads approaches 0.5 as sample size increases • How the number of heads does not approach half the sample size
Law of Large Numbers.ftm
Law of Large Numbers 2.ftm
12 How Random Walks Go as Root N How the distance from the origin increases with the number of steps
Random Walk Root n.ftm
Random Walk Root n part 2.ftm
13 Building the Binomial Distribution Constructing the binomial distribution by resampling • How the distribution depends on the population proportion
Building Binomial.ftm
Building Binomial part 2.ftm
Building Binomial part 3.ftm
14 More Binomial How the binomial distribution depends on sample size for small N • The relationship between the distribution of sample proportions and the distribution of sample counts
Binomial Small N.ftm
15 Two-Dimensional Random Walks Unexpected behavior in 2D random walks • How the 2D walk eventually looks like a 2D normal distribution
2D Random Walk.ftm

Standard Deviation, Standard Error, and Student's t
16 Standard Error and Standard Deviation Getting a feel for the difference between standard deviation and standard error
SD and SE.ftm
17 What Is Standard Error, Really? The connection between standard error and the sampling distribution of the mean • How the sample size connects standard deviation and standard error
What is SE.ftm
18 The Road to Student's t Using standard error as the scale for measuring how far a sample mean is from the true mean • How these quantities are not normally distributed; in fact they follow a t distribution
Road to t.ftm
19 A Close Look at the t Statistic How sample mean, standard deviation, t, and P interrelate • How they depend on the values of individual points in a sample
Close Look at Student's t.ftm

Sampling Distributions
20 The Distribution of Sample Proportions How sample size and population proportion affect the distribution
Dist of Sample Props.ftm
21 Adding Uniform Random Variables What happens when you add two uniform random variables • How that corresponds to adding two dice
Adding Uniform.ftm
22 How Errors Add Basic error analysis • If two quantities each have some measurement error, finding the error in their sum
Adding Errors.ftm
23 Sampling Distributions and Sample Size How sampling distributions (of the mean) get narrower as you increase sample size
Sampling Dists Sample Size.ftm
24 How the Width of the Sampling Distribution Depends on N How the width (as measured by IQR) of a sampling distribution of the mean is inversely proportional to the square root of the sample size
Width of Dist Depends on N.ftm
25 Does n – 1 Really Work in the SD? Unbiased estimators • How the familiar formula for sample standard deviation is not unbiased • Why we should care about variance
Estimating SD from a Sample.ftm
26 German Tanks Unbiased estimators • Evaluating estimators from their sampling distributions • Even among unbiased estimators, some are better than others
Tanks.ftm
Tanks2.ftm
27 The Central Limit Theorem A demo of the CLT • How sampling distributions usually look normal • Cases where they do not
Central Limit Theorem.ftm

Confidence Intervals
28 The Confidence Interval of a Proportion Defining the confidence interval • Looking at sample results in terms of plausibility
CI of a Proportion.ftm
29 Capturing with Confidence Intervals How confidence intervals of a proportion do not always capture the population value
Capturing Props with CIs.ftm
CI Stairway.ftm
CI Stairway 2.ftm
30 Where Does That Root(p(1 – p)) Come From? The standard deviation of a variable that's only zero or one • Connecting the “proportion” situation to the “mean” situation
SD of a Bernoulli Variable.ftm
31 Why np>10 is a Good Rule of Thumb Explaining the np > 10 rule for using the normal approximation in the CI of a proportion
Why np is greater than 10.ftm
Binomial v Normal.ftm
32 How the Width of the CI Depends on N How the width of a confidence interval is inversely proportional to the square root of the sample size
Width of CI depends on N.ftm
33 Using the Bootstrap to Estimate a Parameter The bootstrap • Using resampling (with replacement) to create an interval for a parameter
Bootstrap.ftm
34 Exploring the Confidence Interval of the Mean How the CI depends on individual values
CI of the Mean.ftm
35 Capturing the Mean with Confidence Intervals How confidence intervals of a mean do not always capture the population value • What repeated CIs look like
Capturing the Mean with CIs.ftm

Hypothesis Tests
36 Fair and Unfair Dice Creating a measure of “fairness” • Sampling distributions • Testing hypotheses empirically • The chi-square statistic
Fair and Unfair Dice.ftm
37 Scrambling to Compare Means Randomization test • Using scrambling to simulate the null hypothesis • Generating a sampling distribution
Scrambling.ftm
Scrambling 2.ftm
38 Using a t-Test to Compare Means Comparing means with Student's t
Compare Means Using t.ftm
39 Another Look at a t-Test Repeated t-tests on samples from the same distribution • How t, P, mean, and standard deviation interrelate
Exploring t.ftm
40 On the Equivalence of Tests and Estimates How a hypothesis test and a confidence interval are really the same
Tests and CIs.ftm
41 Paired Versus Unpaired How a paired test gives a significant result more easily than its unpaired counterpart
Paired Versus Unpaired.ftm
42 Analysis of Variance Assessing whether means are different in different groups • Introduction to ANOVA
Within and Between.ftm

Power in Tests
43 The Distribution of P-values How the distribution of P is flat if the null hypothesis is true • How it changes if the null hypothesis is false
Distribution of P-values.ftm
44 Power How power—the chance that you reject the null hypothesis— changes with the population parameters
Power.ftm
45 Power and Sample Size How power—the chance that you reject the null hypothesis— changes with sample size
Power and Sample Size.ftm
46 Heteroscedasticity and its Consequences Homoscedasticity is an assumption behind many statistical calculations. What happens when that assumption is not met?
Heteroscedasticity.ftm
Heteroscedasticity 2.ftm

Distributions
47 Wait Time and the Geometric Distribution The distribution of times until something happens • How this is the geometric distribution
Wait Time.ftm
48 The Exponential and Poisson Distributions The continuous analog to the geometric distribution • How many events happen in a different time: a Poisson distribution
Exponential and Poisson.ftm
49 Sampling Without Replacement and the Hypergeometric Distribution How distributions change when the sample is large compared to the population
Hypergeometric.ftm
50 The Bizarre Cauchy Distribution The Cauchy distribution • The meaning of mean and standard deviation; how it's possible for a distribution to have neither
Cauchy.ftm

eeps media © 2002.