Muutke küpsiste eelistusi

Comparing Groups: Randomization and Bootstrap Methods Using R [Kõva köide]

3.50/5 (4 hinnangut Goodreads-ist)
(University of Iowa), (University of Maryland), (University of Minnesota)
  • Formaat: Hardback, 332 pages, kõrgus x laius x paksus: 239x158x25 mm, kaal: 590 g, Graphs: 50 B&W, 0 Color
  • Ilmumisaeg: 01-Jul-2011
  • Kirjastus: John Wiley & Sons Inc
  • ISBN-10: 0470621699
  • ISBN-13: 9780470621691
Teised raamatud teemal:
Comparing Groups: Randomization and Bootstrap Methods Using RSuurem pilt
  • Formaat: Hardback, 332 pages, kõrgus x laius x paksus: 239x158x25 mm, kaal: 590 g, Graphs: 50 B&W, 0 Color
  • Ilmumisaeg: 01-Jul-2011
  • Kirjastus: John Wiley & Sons Inc
  • ISBN-10: 0470621699
  • ISBN-13: 9780470621691
Teised raamatud teemal:
Püsilink: https://www.kriso.ee/db/9780470621691.html
Märksõnad:
"This book, written by three behavioral scientists for other behavioral scientists, addresses common issues in statistical analysis for the behavioral and educational sciences. Modern Statistical & Computing Methods for the Behavioral and Educational Sciences using R emphasizes the direct link between scientific research questions and data analysis. Purposeful attention is paid to the integration of design, statistical methodology, and computation to propose answers to specific research questions. Furthermore, practical suggestions for the analysis and presentation of results, in prose, tables and/or figures, are included. Optional sections for each chapter include methodological extensions for readers desiring additional technical details. Rather than focus on mathematical calculations like so many other introductory texts in the behavioral sciences, the authors focus on conceptual explanations and the use of statistical computing. Statistical computing is an integral part of statistical work, and to support student learning in this area, examples using the R computer program are provided throughout the book. Rather than relegate examples to the end of chapters, the authors interweave computer examples with the narrative of the book. Topical coverage includes an introduction to R, data exploration of one variable, data exploration of multivariate data - comparing two groups and many groups, permutation and randomization tests, the independent samples t-Test, the Bootstrap test, interval estimates and effect sizes, power, and dependent samples"--

Provided by publisher.

This book, written by three behavioral scientists for other behavioral scientists, addresses common issues in statistical analysis for the behavioral and educational sciences. Modern Statistical & Computing Methods for the Behavioral and Educational Sciences using R emphasizes the direct link between scientific research questions and data analysis. Purposeful attention is paid to the integration of design, statistical methodology, and computation to propose answers to specific research questions. Furthermore, practical suggestions for the analysis and presentation of results, in prose, tables and/or figures, are included. Optional sections for each chapter include methodological extensions for readers desiring additional technical details. Rather than focus on mathematical calculations like so many other introductory texts in the behavioral sciences, the authors focus on conceptual explanations and the use of statistical computing.  Statistical computing is an integral part of statistical work, and to support student learning in this area, examples using the R computer program are provided throughout the book. Rather than relegate examples to the end of chapters, the authors interweave computer examples with the narrative of the book. Topical coverage includes an introduction to R, data exploration of one variable, data exploration of multivariate data - comparing two groups and many groups, permutation and randomization tests, the independent samples t-Test, the Bootstrap test, interval estimates and effect sizes, power, and dependent samples.

Arvustused

The book can be used from upper-undergraduate and graduate level courses on statistical methods, particularly in the educational and behavioral sciences. The book also serves as a valuable resource for researchers who need a practical guide to modern data analytic and computational methods.  (Zentralblatt Math, 1 August 2013) The three authors of this book have a deep understanding of research methods and statistics and provide great value in this book for students of this subject and readers interested in it.  (Biz India, 8 May 2012)

 

List of Figures xiii
List of Tables xxi
Foreword xxiii
Preface xxv
Acknowledgments xxxi
1 An Introduction to R 1(20)
1.1 Getting Started
2(2)
1.1.1 Windows OS
2(1)
1.1.2 Mac OS
2(1)
1.1.3 Add-On Packages
2(2)
1.2 Arithmetic: R as a Calculator
4(1)
1.3 Computations in R: Functions
4(3)
1.4 Connecting Computations
7(2)
1.4.1 Naming Conventions
8(1)
1.5 Data Structures: Vectors
9(4)
1.5.1 Creating Vectors in R
9(2)
1.5.2 Computation with Vectors
11(1)
1.5.3 Character and Logical Vectors
12(1)
1.6 Getting Help
13(1)
1.7 Alternative Ways to Run R
14(1)
1.8 Extension: Matrices and Matrix Operations
14(4)
1.8.1 Computation with Matrices
15(3)
1.9 Further Reading
18(1)
Problems
19(2)
2 Data Representation and Preparation 21(28)
2.1 Tabular Data
23(1)
2.1.1 External Formats for Storing Tabular Data
23(1)
2.2 Data Entry
24(1)
2.2.1 Data Codebooks
25(1)
2.3 Reading Delimited Data into R
25(4)
2.3.1 Identifying the Location of a File
26(2)
2.3.2 Examining the Data in a Text Editor
28(1)
2.3.3 Reading Delimited Separated Data: An Example
28(1)
2.4 Data Structure: Data Frames
29(4)
2.4.1 Examining the Data Read into R
29(4)
2.5 Recording Syntax using Script Files
33(1)
2.5.1 Documentation File
34(1)
2.6 Simple Graphing in R
34(3)
2.6.1 Saving Graphics to Insert into a Word-Processing File
35(2)
2.7 Extension: Logical Expressions and Graphs for Categorical Variables
37(8)
2.7.1 Logical Operators
38(2)
2.7.2 Measurement Level and Analysis
40(2)
2.7.3 Categorical Data
42(2)
2.7.4 Plotting Categorical Data
44(1)
2.8 Further Reading
45(1)
Problems
46(3)
3 Data Exploration: One Variable 49(16)
3.1 Reading In the Data
50(2)
3.2 Nonparametric Density Estimation
52(6)
3.2.1 Graphically Summarizing the Distribution
52(1)
3.2.2 Histograms
52(1)
3.2.3 Kernel Density Estimators
53(1)
3.2.4 Controlling the Density Estimation
53(2)
3.2.5 Plotting the Estimated Density
55(3)
3.3 Summarizing the Findings
58(4)
3.3.1 Creating a Plot for Publication
59(2)
3.3.2 Writing Up the Results for Publication
61(1)
3.4 Extension: Variability Bands for Kernel Densities
62(1)
3.5 Further Reading
62(1)
Problems
63(2)
4 Exploration of Multivariate Data: Comparing Two Groups 65(28)
4.1 Graphically Summarizing the Marginal Distribution
66(1)
4.2 Graphically Summarizing Conditional Distributions
66(6)
4.2.1 Indexing: Accessing Individuals or Subsets
68(1)
4.2.2 Indexing Using a Logical Expression
69(1)
4.2.3 Density Plots of the Conditional Distributions
70(1)
4.2.4 Side-by-Side Box-and-Whiskers Plots
70(2)
4.3 Numerical Summaries of Data: Estimates of the Population Parameters
72(8)
4.3.1 Measuring Central Tendency
73(1)
4.3.2 Measuring Variation
74(2)
4.3.3 Measuring Skewness
76(2)
4.3.4 Kurtosis
78(2)
4.4 Summarizing the Findings
80(7)
4.4.1 Creating a Plot for Publication
80(1)
4.4.2 Using Color
81(4)
4.4.3 Selecting a Color Palette
85(2)
4.5 Extension: Robust Estimation
87(4)
4.5.1 Robust Estimate of Location: The Trimmed Mean
87(2)
4.5.2 Robust Estimate of Variation: The Winsorized Variance
89(2)
4.6 Further Reading
91(1)
Problems
91(2)
5 Exploration of Multivariate Data: Comparing Many Groups 93(22)
5.1 Graphing Many Conditional Distributions
94(6)
5.1.1 Panel Plots
96(1)
5.1.2 Side-by-Side Box-and-Whiskers Plots
97(3)
5.2 Numerically Summarizing the Data
100(1)
5.3 Summarizing the Findings
101(2)
5.3.1 Writing Up the Results for Publication
102(1)
5.3.2 Enhancing a Plot with a Line
102(1)
5.4 Examining Distributions Conditional on Multiple Variables
103(4)
5.5 Extension: Conditioning on Continuous Variables
107(5)
5.5.1 Scatterplots of the Conditional Distributions
110(2)
5.6 Further Reading
112(1)
Problems
113(2)
6 Randomization and Permutation Tests 115(22)
6.1 Randomized Experimental Research
118(1)
6.2 Introduction to the Randomization Test
119(3)
6.3 Randomization Tests with Large Samples: Monte Carlo Simulation
122(8)
6.3.1 Rerandomization of the Data
124(1)
6.3.2 Repeating the Randomization Process
125(1)
6.3.3 Generalizing Processes: Functions
126(1)
6.3.4 Repeated Operations on Matrix Rows or Columns
127(1)
6.3.5 Examining the Monte Carlo Distribution and Obtaining the p-Value
127(3)
6.4 Validity of the Inferences and Conclusions Drawn from a Randomization Test
130(2)
6.4.1 Exchangeability
130(1)
6.4.2 Nonexperimental Research: Permutation Tests
131(1)
6.4.3 Nonexperimental, Nongeneralizable Research
131(1)
6.5 Generalization from the Randomization Results
132(1)
6.6 Summarizing the Results for Publication
133(1)
6.7 Extension: Tests of the Variance
133(1)
6.8 Further Reading
134(1)
Problems
135(2)
7 Bootstrap Tests 137(34)
7.1 Educational Achievement of Latino Immigrants
138(2)
7.2 Probability Models: An Interlude
140(1)
7.3 Theoretical Probability Models in R
141(2)
7.4 Parametric Bootstrap Tests
143(3)
7.4.1 Choosing a Probability Model
144(1)
7.4.2 Standardizing the Distribution of Achievement Scores
144(2)
7.5 The Parametric Bootstrap
146(2)
7.5.1 The Parametric Bootstrap: Approximating the Distribution of the Mean Difference
146(2)
7.6 Implementing the Parametric Bootstrap in R
148(6)
7.6.1 Writing a Function to Randomly Generate Data for the boot() Function
148(2)
7.6.2 Writing a Function to Compute a Test Statistic Using the Randomly Generated Data
150(1)
7.6.3 The Bootstrap Distribution of the Mean Difference
151(3)
7.7 Summarizing the Results of the Parametric Bootstrap Test
154(1)
7.8 Nonparametric Bootstrap Tests
154(6)
7.8.1 Using the Nonparametric Bootstrap to Approximate the Distribution of the Mean Difference
157(1)
7.8.2 Implementing the Nonparametric Bootstrap in R
158(2)
7.9 Summarizing the Results for the Nonparametric Bootstrap Test
160(1)
7.10 Bootstrapping Using a Pivot Statistic
161(3)
7.10.1 Student's t-Statistic
161(3)
7.11 Independence Assumption for the Bootstrap Methods
164(2)
7.12 Extension: Testing Functions
166(2)
7.12.1 Ordering a Data Frame
166(2)
7.13 Further Reading
168(1)
Problems
168(3)
8 Philosophical Considerations 171(8)
8.1 The Randomization Test vs. the Bootstrap Test
172(1)
8.2 Philosophical Frameworks of Classical Inference
173(6)
8.2.1 Fisher's Significance Testing
174(1)
8.2.2 Neyman-Pearson Hypothesis Testing
175(1)
8.2.3 p-Values
176(3)
9 Bootstrap Intervals and Effect Sizes 179(26)
9.1 Educational Achievement Among Latino Immigrants: Example Revisited
180(1)
9.2 Plausible Models to Reproduce the Observed Result
180(5)
9.2.1 Computing the Likelihood of Reproducing the Observed Result
181(4)
9.3 Bootstrapping Using an Alternative Model
185(6)
9.3.1 Using R to Bootstrap under the Alternative Model
187(3)
9.3.2 Using the Bootstrap Distribution to Compute the Interval Limits
190(1)
9.3.3 Historical Interlude: Student's Approximation for the Interval Estimate
190(1)
9.3.4 Studentized Bootstrap Interval
191(1)
9.4 Interpretation of the Interval Estimate
191(1)
9.5 Adjusted Bootstrap Intervals
192(1)
9.6 Standardized Effect Size: Quantifying the Group Differences in a Common Metric
192(5)
9.6.1 Effect Size as Distance-Cohen's δ
193(2)
9.6.2 Robust Distance Measure of Effect
195(2)
9.7 Summarizing the Results
197(1)
9.8 Extension: Bootstrapping the Confidence Envelope for a Q-Q Plot
197(1)
9.9 Confidence Envelopes
198(4)
9.10 Further Reading
202(2)
Problems
204(1)
10 Dependent Samples 205(22)
10.1 Matching: Reducing the Likelihood of Nonequivalent Groups
206(1)
10.2 Mathematics Achievement Study Design
206(5)
10.2.1 Exploratory Analysis
209(2)
10.3 Randomization/Permutation Test for Dependent Samples
211(5)
10.3.1 Reshaping the Data
212(2)
10.3.2 Randomization Test Using the Reshaped Data
214(2)
10.4 Effect Size
216(1)
10.5 Summarizing the Results of a Dependent Samples Test for Publication
217(1)
10.6 To Match or Not to Match...That is the Question
218(2)
10.7 Extension: Block Bootstrap
220(3)
10.8 Further Reading
223(1)
Problems
224(3)
11 Planned Contrasts 227(26)
11.1 Planned Comparisons
228(1)
11.2 Examination of Weight Loss Conditioned on Diet
228(4)
11.2.1 Exploration of Research Question 1
229(1)
11.2.2 Exploration of Research Question 2
230(1)
11.2.3 Exploration of Research Question 3
231(1)
11.3 From Research Questions to Hypotheses
232(1)
11.4 Statistical Contrasts
233(4)
11.4.1 Complex Contrasts
236(1)
11.5 Computing the Estimated Contrasts Using the Observed Data
237(2)
11.6 Testing Contrasts: Randomization Test
239(1)
11.7 Strength of Association: A Measure of Effect
240(3)
11.7.1 Total Sum of Squares
241(2)
11.8 Contrast Sum of Squares
243(1)
11.9 Eta-Squared for Contrasts
243(1)
11.10 Bootstrap Interval for Eta-Squared
244(1)
11.11 Summarizing the Results of a Planned Contrast Test Analysis
245(1)
11.12 Extension: Orthogonal Contrasts
245(6)
11.13 Further Reading
251(1)
Problems
251(2)
12 Unplanned Contrasts 253(32)
12.1 Unplanned Comparisons
254(1)
12.2 Examination of Weight Loss Conditioned on Diet
254(3)
12.3 Omnibus Test
257(12)
12.3.1 Statistical Models
257(1)
12.3.2 Postulating a Statistical Model to Fit the Data
258(2)
12.3.3 Fitting a Statistical Model to the Data
260(2)
12.3.4 Partitioning Variation in the Observed Scores
262(6)
12.3.5 Randomization Test for the Omnibus Hypothesis
268(1)
12.4 Group Comparisons After the Omnibus Test
269(1)
12.5 Ensemble-Adjusted p-values
270(3)
12.5.1 False Discovery Rate
272(1)
12.6 Strengths and Limitations of the Four Approaches
273(3)
12.6.1 Planned Comparisons
273(1)
12.6.2 Omnibus Test Followed by Unadjusted Group Comparisons
274(1)
12.6.3 Omnibus Test Followed by Adjusted Group Comparisons
274(1)
12.6.4 Adjusted Group Comparisons without the Omnibus Test
275(1)
12.6.5 Final Thoughts
276(1)
12.7 Summarizing the Results of Unplanned Contrast Tests for Publication
276(1)
12.8 Extension: Plots of the Unplanned Contrasts
276(6)
12.8.1 Simultaneous Intervals
280(2)
12.9 Further Reading
282(1)
Problems
283(2)
References 285
Andrew S. Zieffler, PhD, is Lecturer in the Department of Educational Psychology at the University of Minnesota. Dr. Zieffler has published numerous articles in his areas of research interest, which include the measurement and assessment in statistics education research and statistical computing. Jeffrey R. Harring, PhD, is Assistant Professor in the Department of Measurement, Statistics, and Evaluation at the University of Maryland. Dr. Harring currently focuses his research on statistical models for repeated measures data and nonlinear structural equation models.

Jeffrey D. Long, PhD, is Professor of Psychiatry in the Carver College of Medicine at The University of Iowa and Head Statistician for Neurobiological Predictors of Huntington's Disease (PREDICT-HD), a longitudinal NIH-funded study of early detection of Huntington's disease. His interests include the analysis of longitudinal and time-to-event data and ordinal data.