IBM SPSS Statistics 19 Guide to Data Analysis

Marija Norusis earned a PhD in biostatistics from the University of Michigan. She was SPSS's first professional statistician. During this time, she wrote her first book, The SPSS Introductory Guide. Since then she has written numerous volumes of highly acclaimed SPSS documentation, and textbooks that demystify statistics and SPSS. Dr. Norusis has been on the faculties of the University of Chicago and Rush Medical College, teaching statistics to diverse audiences. When not working on SPSS guides, Marija analyzes real data as a statistical consultant. For more detailed information about Dr. Norusis and her SPSS guides, visit her website at www.norusis.com.

PART 1. GETTING STARTED WITH IBM SPSS STATISTICS

1. Introduction

About This Book

Getting Started with IBM SPSS Statistics

Describing Data

Testing Hypotheses

Examining Relationships

Lets Get Started



2. An Introductory Tour of IBM SPSS Statistics

Starting IBM SPSS Statistics

Help Is Always at Hand

Copying the Data Files

Opening a Data File

Statistical Procedures

The Viewer Window

Viewer Objects

The Data Editor Window

Entering Non-Numeric Data

Clearing the Data Editor without Saving Changes

The IBM SPSS Statistics Online Tutorial

The IBM SPSS Statistics Toolbar

The IBM SPSS Statistics Help System

Contextual Help

What's Next?



3. Sources of Data

Know Your Data

Survey Data

Asking the Question

Measuring Time

Selecting Participants

Selecting a Sample

General Social Survey

Random-Digit Dialing

Internet Surveys

Designing Experiments

Random Assignment

Minimizing Bias

Summary

What's Next?

Exercises



PART 2. DESCRIBING DATA

4. Counting Responses

Describing Variables

A Simple Frequency Table

Sorting Frequency Tables

Pie Charts

Bar Charts

Summarizing Internet Time

Histograms

Mode and Median

Percentiles

Summary

What's Next?

How to Obtain a Frequency Table

Format: Appearance of the Frequency Table

Statistics: Univariate Statistics

Charts: Bar Charts, Pie Charts, and Histograms

Exercises



5. Computing Descriptive Statistics

Summarizing Data

Scales of Measurement

Mode, Median, and Arithmetic Average

Comparing Mean and Median

Summarizing Time Spent Online

Measures of Variability

Range

Variance and Standard Deviation

The Coefficient of Variation

Standard Scores

Summary

What's Next?

How to Obtain Univariate Descriptive Statistics

Options: Choosing Statistics and Sorting Variables

Exercises



6. Comparing Groups

Age, Education, and Internet Use

Plotting Means

Layers: Defining Subgroups by More than One Variable

Summary

What's Next?

How to Obtain Subgroup Means

Layers: Defining Subgroups by More than One Variable

Options: Additional Statistics and Display of Labels

Exercises



7. Looking at Distributions

Marathon Completion Times

Age and Gender

Marathon Times for Mature Runners

Summary

What's Next?

How to Explore Distributions

Explore Statistics

Graphical Displays

Options

Exercises



8. Counting Responses for Combinations of Variables

Library Use and Education

Row and Column Percentages

Bar Charts

Adding Control Variables

Library Use and the Internet

Summary

What's Next?

How to Obtain a Crosstabulation

Layers: Three or More Variables at Once

Cells: Percentages, Expected Counts, and Residuals

Bivariate Statistics

Format: Adjusting the Table Format

Exercises



9. Plotting Data

Examining Population Indicators

Simple Scatterplots

Scatterplot Matrices

Overlay Plots

Three-Dimensional Plots

Identifying Unusual Points

Rotating 3-D Scatterplots

Summary

What's Next?

How to Obtain a Scatterplot

Obtaining a Simple Scatterplot

Obtaining an Overlay Scatterplot

Obtaining a Scatterplot Matrix

Obtaining a 3-D Scatterplot

Editing a Scatterplot

Exercises



PART 3. TESTING HYPOTHESES

10. Evaluating Results from Samples

From Sample to Population

A Computer Model

The Effect of Sample Size

The Binomial Test

Summary

What's Next?

Exercises



11. The Normal Distribution

The Normal Distribution

Samples from a Normal Distribution

Means from a Normal Population

Are the Sample Results Unlikely?

Testing a Hypothesis

Means from Non-Normal Distributions

Means from a Uniform Distribution

Summary

What's Next?

Exercises



12. Testing a Hypothesis about a Single Mean

Examining the Data

The T Distribution

Calculating the T Statistic

Confidence Intervals

Other Confidence Levels

Confidence Interval for a Difference

Confidence Intervals and Hypothesis Tests

Null Hypotheses and Alternative Hypotheses

Rejecting the Null Hypothesis

Summary

What's Next?

How to Obtain a One-Sample T Test

Options: Confidence Level and Missing Data

Exercises



13. Testing a Hypothesis about Two Related Means

Marathon Runners in Paired Designs

Looking at Differences

Is the Mean Difference Zero?

Two Approaches

The Paired-Samples T Test

Are You Positive?

Some Possible Problems

Examining Normality

Summary

What's Next?

How to Obtain a Paired-Samples T Test

Options: Confidence Level and Missing Data

Exercises



14. Testing a Hypothesis about Two Independent Means

Examining Television Viewing

Distribution of Differences

Standard Error of the Mean Difference

Computing the T Statistic

Output from the Two-Independent-Samples T Test

Confidence Intervals for the Mean Difference

Testing the Equality of Variances

Effect of Outliers

Introducing Education

Can You Prove the Null Hypothesis?

Interpreting the Observed Significance Level

Power

Monitoring Death Rates

Does Significant Mean Important?

Summary

What's Next?

How to Obtain an Independent-Samples T Test

Define Groups: Specifying the Subgroups

Options: Confidence Level and Missing Data

Exercises



15. One-Way Analysis of Variance

Hours in a Work Week

Describing the Data

Confidence Intervals for the Group Means

Testing the Null Hypothesis

Assumptions Needed for Analysis of Variance

Analyzing the Variability

Comparing the Two Estimates of Variability

The Analysis-of-Variance Table

Multiple Comparison Procedures

Television Viewing, Education, and Internet Use

Summary

What's Next?

How to Obtain a One-Way Analysis of Variance

Post Hoc Multiple Comparisons: Finding the Difference

Options: Statistics and Missing Data

Exercises



16. Two-Way Analysis of Variance

The Design

Examining the Data

Testing Hypotheses

Degree and Gender Interaction

Necessary Assumptions

Analysis-of-Variance Table

Testing the Degree-by-Gender Interaction

Testing the Main Effects

Removing the Interaction Effect

Where Are the Differences?

Multiple Comparison Results

Checking Assumptions

A Look at Television

Extensions

Summary

What's Next?

How to Obtain a GLM Univariate Analysis

GLM Univariate: Model

GLM Univariate: Plots

GLM Univariate: Post Hoc

GLM Univariate: Options

GLM Univariate: Save

Exercises



17. Comparing Observed and Expected Counts

Freedom or Manners?

Observed and Expected Counts

The Chi-Square Statistic

A Larger Table

Does College Open Doors?

A One-Sample Chi-Square Test

Power Concerns

Summary

What's Next?

Exercises



18. Nonparametric Tests

Nonparametric Tests for Paired Data

Sign Test

Wilcoxon Test

Who's Sending E-mail?

Mann-Whitney Test

Kruskal-Wallis Test

Friedman Test

Summary

How to Obtain Nonparametric Tests

Chi-Square Test

Binomial Test

Two-Independent-Samples Tests

Several-Independent-Samples Tests

Two-Related-Samples Tests

Several-Related-Samples Tests

Options: Descriptive Statistics and Missing Values

Exercises



PART 4. EXAMINING RELATIONSHIPS

19. Measuring Association

Components of the Justice System

Proportional Reduction in Error

Measures of Association for Ordinal Variables

Concordant and Discordant Pairs

Measures Based on Concordant and Discordant Pairs

Evaluating the Components

Measuring Agreement

Correlation-Based Measures

Measures Based on the Chi-Square Statistic

Summary

What's Next?

Exercises



20. Linear Regression and Correlation

Life Expectancy and Birthrate

Choosing the Best Line

Calculating the Least-Squares Line

Calculating Predicted Values and Residuals

Determining How Well the Line Fits

Explaining Variability

Some Warnings

Summary

What's Next?

How to Obtain a Linear Regression

Statistics: Further Information on the Model

Residual Plots: Basic Residual Analysis

Linear Regression Save: Creating New Variables

Linear Regression Options

Exercises



21. Testing Regression Hypotheses

The Population Regression Line

Assumptions Needed for Testing Hypotheses

Testing Hypotheses

Testing that the Slope Is Zero

Confidence Intervals for the Slope and Intercept

Predicting Life Expectancy

Predicting Means and Individual Observations

Standard Error of the Predicted Mean

Confidence Intervals for the Predicted Means

Prediction Intervals for Individual Cases

Summary

What's Next?

How to Obtain a Bivariate Correlation

Options: Additional Statistics and Missing Data

How to Obtain a Partial Correlation

Options: Additional Statistics and Missing Data

Exercises



22. Analyzing Residuals

Residuals

Standardized Residuals

Studentized Residuals

Checking for Normality

Checking for Constant Variance

Checking Linearity

Checking Independence

A Final Comment on Assumptions

Looking for Influential Points

Studentized Deleted Residuals

Summary

What's Next?

Exercises



23. Building Multiple Regression Models

Predicting Life Expectancy

The Model

Assumptions for Multiple Regression

Examining the Variables

Looking at How Well the Model Fits

Examining the Coefficients

Interpreting the Partial Regression Coefficients

Changing the Model

Partial Correlation Coefficients

Tolerance and Multicollinearity

Beta Coefficients

Building a Regression Model

Methods for Selecting Variables

Summary

What's Next?

How to Obtain a Multiple Linear Regression

Options: Variable Selection Criteria

Exercises



24. Multiple Regression Diagnostics

Examining Normality

Scatterplots of Residuals

Leverage

Changes in the Coefficients

Cook's Distance

Plots against Independent Variables

Partial Regression Plot

Why Bother?

Summary

Exercises



Appendices

A. Obtaining Charts in IBM SPSS Statistics

Overview

Creating Bar Charts

Creating a Bar Chart for Single Variable

Creating a Clustered Bar Chart

Creating a Chart with Multiple Variables

Modifying Charts

Collapsing Pie Chart Slices

Changing the Scale of Histogram

Saving Chart Files

B. Transforming and Selecting Data

Data Transformations

Transformations at a Glance

Saving Changes

Delaying Processing of Transformations

Recoding Values

Computing Variables

The Calculator Pad

Automatic Recoding

Conditional Transformations

Case Selection

Temporary or Permanent Selection

Other Selection Methods

C. The T Distribution

D. Areas under the Normal Curve

E. Descriptions of Data Files

F. Answers to Selected Exercises