Muutke küpsiste eelistusi

Best Practices in Logistic Regression [Pehme köide]

4.80/5 (9 hinnangut Goodreads-ist)
  • Formaat: Paperback / softback, 488 pages, kõrgus x laius: 231x187 mm, kaal: 720 g
  • Ilmumisaeg: 24-Apr-2014
  • Kirjastus: SAGE Publications Inc
  • ISBN-10: 1452244790
  • ISBN-13: 9781452244792
Teised raamatud teemal:
Best Practices in Logistic RegressionSuurem pilt
  • Formaat: Paperback / softback, 488 pages, kõrgus x laius: 231x187 mm, kaal: 720 g
  • Ilmumisaeg: 24-Apr-2014
  • Kirjastus: SAGE Publications Inc
  • ISBN-10: 1452244790
  • ISBN-13: 9781452244792
Teised raamatud teemal:
Püsilink: https://www.kriso.ee/db/9781452244792.html
Märksõnad:
Osborne explains logistic regression to advanced statistics students, seasoned researchers, and anyone else interested in using what he calls a quirky technique. He presents it conceptually, with minimal formulae or mathematics, but with pedagogical examples that carry throughout the book. He proposes some best practices along the way. Among his topics are a conceptual introduction to bivariate logistic regression, using unordered categorical independent variables in logistic regression, curvilinear effects in logistic regression, modern and effective methods for dealing with missing data, and multi-level modeling with logistic regression. Annotation ©2014 Ringgold, Inc., Portland, OR (protoview.com)

This book explores the fun things researchers can do with logistic regression, explicates and simplifies the confounding complexities of understanding what logistic regression is, and provides evidence-based guidance as to what the best practices in performing logistic regression are.

Arvustused

 "This text is extremely student-friendly . . . it is a nearly perfect balance of conceptual explanation and application using example data sets" -- Denna L. Wheeler, Oklahoma State University Center for Health Sciences "This is an absolutely stellar approach to a very difficult and under-used analysis. The use of humor, practical examples, the use of real data, and the inclusion of both basic and advanced concepts without being overly concerned with the derivation of the analysis, foster a better understanding of logistic regression." -- Frank B. Underwood, University of Evansville "The text will serve well to widely expand the usage of the logistic regression in social science research. The not-too-technical explanation of core concepts, with numerous computer outputs for illustrations, makes it a perfect text for the senior undergraduate and graduate-level course, as well as a reference for the analytical practitioner." -- Professor David Han, University of Texas, San Antonio "I appreciate the emphasis on application and the coverage of topics that are useful in research but neglected by other books on this method." -- Dr. Chuck W. Peek, University of Florida "This is a very impressive book. The topic is timely." -- Shanta Pandey, Washington University, St. Louis  "It is a very good text and covers topics, such as the need to clean data, inefficiency/volatility of estimates, and missing data effects, that are not generally dealt with." -- P. Neal Ritchey, University of Cincinnati "The book includes detailed explanations of various logistic regression models using a range of data and analysis results. It is very suitable for social science students." -- Daoqin Tong, University of Arizona "This book is concise, accessible, and reader-friendly, particularly for those in education research. The value of this book lies not only in laying out certain "best practices," but more importantly in pointing out common pitfalls and showing newcomers the way around." -- Yang Cao, University of North Carolina, Charlotte

Preface xvii
Acknowledgments xxi
About the Author xxiii
1 A Conceptual Introduction To Bivariate Logistic Regression
1(18)
What Is Ordinary Least Squares Regression and How Is Logistic Regression Different?
3(7)
OLS Regression---A Deeper Conceptual Look
7(1)
Maximum Likelihood Estimation---A Gentle but Deeper Look
8(2)
Differences and Similarities in Assumptions Between OLS and Logistic Regression
10(5)
Distributional Assumptions
10(1)
Linearity of the Relationship
10(4)
Perfect Measurement
14(1)
Homoscedasticity (or Constant Variance of the Residuals)
14(1)
Independence of Observations
15(1)
Similarities Between OLS Regression and Logistic Regression
15(1)
Summarizing the Overall A Model
15(1)
What Is Discriminant Function Analysis and How Is Logistic Regression Superior/Different?
16(1)
Summary
17(1)
References
17(2)
2 How Does Logistic Regression Handle A Binary Dependent Variable?
19(26)
Probabilities, Conditional Probabilities, and Odds
21(8)
A Brief Thought Experiment on the Logistic Curve
24(2)
The Benefits of Odds Over Simple OLS Regression for Binary Outcomes
26(1)
The Odds Ratio
26(1)
The Logit
27(2)
Summarizing So Far
29(1)
Still More Fun With Logits, Odds, and Probabilities
30(2)
The Logit Revisited
31(1)
Where Is the Logit of the Other Group?
32(1)
Converting From Logits Directly Back to Conditional Probabilities
33(4)
Some Benefits of Conditional Probabilities
34(3)
Confidence Intervals for Logits, Odds Ratios, and Predicted Probabilities
37(2)
One Concrete Reason Why You Should Care About This Stuff---Clarity of Communication!
39(1)
Enrichment
39(2)
Answer Key
41(2)
Syntax Examples
43(1)
References
44(1)
3 Performing Simple Logistic Regression
45(40)
Simple Binary Logistic Regression With One Independent Variable
45(11)
Example 1 The Relationship Between Obesity and Diagnosis of Diabetes
46(10)
An Example Summary of These Results
56(1)
Logistic Regression With a Continuous Independent Variable
56(11)
Example Summary of This Analysis
64(1)
Example 2 The Relationship Between Family Poverty and Dropping Out of School Prior to Graduation
64(2)
Example Summary of This Analysis
66(1)
Predicting Dropout From a Continuous Family SES Variable
67(4)
Example Summary of This Analysis
70(1)
Are Classification Tables Ever Useful?
71(4)
How Should We Interpret Odds Ratios That Are Less Than 1.0?
75(3)
Summary
78(1)
Enrichment
78(2)
Answer Key
80(2)
Syntax Examples
82(1)
References
83(2)
4 A Practical Guide To Testing-Assumptions And Cleaning Data For Logistic Regression
85(46)
Independence of Observations
86(1)
Collinearity of Independent Variables
87(1)
Fully Represented Data Matrix
88(1)
Perfect Measurement
89(2)
Model Is Correctly Specified---No Important Variables Omitted, No Extraneous Variables Included
91(9)
The Logit Link Function Is Appropriate
91(1)
The Relationships Are "Linear on the Logit"
92(3)
Effects of Independent Variables Are Additive in Nature
95(1)
All Important Variables Are Included in the Analysis
96(1)
No Extraneous Variables Are Included in the Analysis
97(3)
Normality of Errors, Distributional Assumptions
100(2)
Exploring Data for Inappropriately Influential Cases
102(8)
Examining Residuals
104(6)
Leverage and Influence in Logistic Regression
110(4)
Cook's Distance (Influence)
112(2)
DfBetas
114(2)
The Relationship Between DfBetas and Other Measures of Influence/Leverage
116(1)
Summary
116(1)
Enrichment
116(1)
Answer Key
117(11)
Syntax Examples
128(2)
References
130(1)
5 Continuous Predictors: Why Splitting Continuous Variables Into Categories Is Undesirable
131(40)
What Is Categorization and Why Does It Exist?
132(2)
How Widespread Is This Practice?
134(2)
Why Do Researchers Use Dichotomization and Similar Techniques?
136(3)
The Evils of Cutoff Scores
137(2)
Are Analyses More Easily Interpreted With Dichotomous Variables?
139(1)
Are Analyses With Dichotomous Variables Easier to Compute?
140(1)
Are Dichotomous Variables More Reliable?
140(10)
Dichotomization Improves Reliability of Measurement
141(9)
Does Dichotomization Effectively Deal With Non-Normality or Outliers?
150(2)
Extreme Groups Analysis: Generally Ill-Advised and Often Dishonest
152(1)
Some Other Drawbacks of Dichotomization
153(1)
Curvilinearity and Interactions Will Be Masked or Undetectable
153(1)
The Variable Is by Nature Categorical
153(1)
Spurious Statistical Significance
154(1)
Dichotomization of Variables and Logistic Regression
154(3)
Inhibiting Accurate Meta-Analysis of Results
156(1)
What About Other k-Groups Categorization Schemes?
157(1)
So What Are Best Practices for Logistic Regression With Continuous Variables?
157(3)
Best Practices in Reporting Results from Continuous Variables: Conditional Probabilities
160(3)
Summary
163(1)
Enrichment
163(2)
Answer Key
165(2)
References
167(4)
6 Using Unordered Categorical Independent Variables In Logistic Regression
171(30)
Ordinality
171(1)
Equal Intervals
172(1)
True Zero Point
172(1)
Different Classifications of Measurement
173(2)
Dummy Coding
175(4)
Define the Reference Group
175(1)
Set Up the Dummy Coded Variables
175(4)
Alternatives to Dummy Coding
179(6)
Difference (Reverse Helmert) Contrasts
182(2)
Deviation Contrasts
184(1)
SAS Contrasts
185(2)
Dummy Coding in SAS
187(2)
Summary
189(1)
Enrichment
189(1)
Answer Key
190(8)
Syntax Examples
198(1)
References
199(2)
7 Curvilinear Effects In Logistic-Regression
201(42)
A Brief Review of the Assumption of Linearity
202(1)
Illegitimate Causes of Curvilinearity
203(4)
Model Misspecification: Omission of Important Variables
203(1)
Violating Equal Intervals in Coding Continuous Variables
203(2)
Poor Data Cleaning
205(2)
Detection of Nonlinear Effects
207(2)
Theory
207(1)
Ad Hoc Testing
208(1)
Box-Tidwell Transformations
208(1)
Curvilinear Logistic Regression Example: Diabetes and Age
209(4)
Adding Quadratic and Cubic Terms to the Logistic Regression Analysis
209(4)
An Example Summary of This Analysis
213(1)
Estimating Curvilinear Relationships Using Box-Tidwell Transformations
213(1)
Data Cleaning and Curvilinear Effects
214(6)
SAS Analyses Using DIFCHISQ
220(2)
Advanced Topics in Curvilinear Regression: Estimating Minima and Maxima as Well as Slope at Any Point on the Curve
222(6)
Summary
228(1)
Enrichment
229(1)
Answer Key
229(11)
Syntax Examples
240(2)
References
242(1)
8 Logistic Regression With Multiple Independent Variables: Opportunities And Pitfalls
243(54)
The Basics of Multiple Predictors
244(1)
What Are the Implications of This Act?
245(3)
Example Summary of Previous Analysis
248(1)
Different Methods of Entry
249(5)
User-Controlled Methods of Entry
249(1)
Hierarchical Entry
249(1)
Blockwise Entry
250(1)
Software-Controlled Entry
251(3)
Collinearity Issues
254(2)
Assessing the Overall Model---Why There Is No R2 for Logistic Regression
256(2)
Interactions
258(9)
What Is an Interaction?
259(1)
Procedural Issues in Testing for Interactions Between Continuous Variables
259(4)
Procedural Issues With Graphing
263(4)
Example Summary of Interaction Analysis
267(1)
Interactions Between Categorical and Continuous Variables
267(4)
Interactions and Data Cleaning
271(3)
Curvilinear Interactions
274(7)
Step 1 Create the Terms Prior to Analysis
276(1)
Step 2 Build Your Equation Slowly
276(5)
Curvilinear Interactions With Categorical Variables
281(3)
Summary
284(1)
Enrichment
285(1)
Answer Key
286(4)
Curvilinear Effects With BMI and Smoking
290(4)
References
294(3)
9 A Brief Overview Of Probit Regression
297(16)
What Is a Probit?
298(2)
The Probit Link
300(2)
Why Are There Two Different Procedures If They Produce the Same Results?
302(4)
The Value of Having a Sensible Intercept
306(3)
Some Nice Features of Probit
309(1)
Assumptions of Probit Regression
310(1)
Summary and Conclusion
310(1)
Enrichment
310(1)
Syntax Examples
311(1)
References
311(2)
10 Replication And Generalizability In Logistic Regression
313(44)
Sample Size, Power, and Volatility in Logistic Regression
314(1)
What Is Statistical Power and Why Should You Care About It?
314(1)
How Null Hypothesis Statistical Testing Relates to Power
315(1)
What Do Statistical Tests Tell Us?
316(2)
So What Does Failure to Reject the Null Hypothesis Mean?
318(1)
So What Is Power and How Does It Relate to Error Rates?
319(3)
Power in Logistic Regression
320(2)
Summary of Points Thus Far
322(1)
Who Cares as Long as p < .05? Volatility in Logistic Regression Analyses
323(14)
SES Analyses
324(11)
Random Variable Analysis
335(2)
Cross-Validation and Replication of Logistic Regression Analyses
337(5)
Prediction and Explanation Using Regression
337(4)
Comparing Internal Validation Models in Logistic Regression
341(1)
In-Depth With BOOTSTRAP
342(7)
Biased Sample #1 (N = 100, b = 0.361, OR = 1.435)
342(2)
Biased Sample #2 (N = 100, b = 1.619, OR = 5.051)
344(1)
Relatively Unbiased Sample #3
344(1)
Can Data Cleaning Help?
345(4)
Summary and Conclusions
349(1)
Enrichment
350(1)
Answer Key
351(2)
References
353(4)
11 Modern And Effective Methods Of Dealing With Missing Data
357(32)
Dealing With Missing or Incomplete Data in Logistic Regression
358(7)
Not All Missing Data Are the Same
360(3)
Categories of Missingness: Why Do We Care If Data Are Missing Completely at Random or Not?
363(2)
How Do You Know If Your Data Are MCAR, MAR, or MNAR?
365(4)
What Do We Do With Missing Data?
369(1)
Data Missing Completely at Random (MCAR)
369(5)
Mean Substitution
371(1)
Strong and Weak Imputation
372(2)
Summary
374(1)
Data Missing Not at Random
374(8)
The Effects of Listwise Deletion (Complex Case Analysis)
375(3)
The Detrimental Effects of Mean Substitution
378(1)
The Effects of Weak Imputation of Values
379(1)
Strong Imputation
380(1)
So Where Does That Leave Us?
381(1)
Multiple Imputation as a Modern Method of Missing Data Estimation
382(1)
How Missingness Can Be an Interesting Variable in and of Itself
383(1)
Summing Up: Benefits of Appropriately Handling Missing Data
384(1)
Enrichment
385(1)
References
386(3)
12 Multinomial And Ordinal Logistic Regression
389(46)
Multinomial Logistic Regression With a Continuous Variable
393(2)
Moving Beyond Simple Multinomial Logistic Regression
395(1)
More Complex Terms in Multinomial Logistic Regression
396(3)
Multinomial Logistic Regression as a Series of Binary Logistic Regression Equations
399(2)
Examples of Data Cleaning Using Binary Logistic Regression
401(4)
Testing Whether Groups Can Be Combined
405(2)
Ordered Logit (Proportional Odds) Model
407(2)
Assumptions of the Ordinal or Proportional Odds Model
409(2)
Interpreting the Results of the Analysis
411(2)
Interpreting the Intercepts
412(1)
Interpreting the Parameter Estimates
412(1)
Data Cleaning and More Advanced Models in Ordinal Logistic Regression
413(2)
Why Not Just Use OLS Regression for This Type of Analysis?
414(1)
Summary and Conclusions
415(1)
Enrichment
415(1)
Answer Key
416(16)
Syntax Examples
432(1)
Reference
433(2)
13 Multilevel Modeling With Logistic Regression
435(16)
What Is HLM?
436(12)
What Is a Hierarchical Data Structure?
436(1)
The Issue With Nested Data
437(2)
How Do Hierarchical Models Work? A Brief Primer
439(2)
A Brief Note About HLM and Statistical Software
441(1)
Residuals in HLM
441(1)
Logits, Odds Ratios, Conditional Odds, Conditional Probabilities, and Relative Risk in HLM
442(1)
Results of DROPOUT Analysis in HLM
443(1)
Cross-Level Interactions in HLM Logistic Regression
444(1)
So What Would Have Happened If These Data Had Been Analyzed via Simple Logistic Regression Without Accounting for the Nested Data Structure?
445(3)
Summary
448(1)
Enrichment
449(1)
References
449(2)
Author Index 451(4)
Subject Index 455
Jason W. Osborne is a thought leader and professor in higher education. His background in educational psychology, statistics and quantitative methods, along with that gleaned from high-level positions within Academia gives a unique perspective on the real-world data factors. In 2015, he was appointed Associate Provost and Dean of the Graduate School at Clemson University in Clemson, South Carolina. As well as Associate Provost, at Clemson University, Jason was a Professor of applied statistics at the School of Mathematical Sciences, with a secondary appointment in Public Health Science. In 2019, he took on the role of Provost and Executive VP for Academic Affairs at Miami University. As Provost, Jason implemented a transformative strategic plan to reposition the institution as one prepared for new challenges with a modern, compelling curriculum, a welcoming environment, and enhanced support for student faculty positions and staff. In 2021, he was named by Stanford University as one of the top 2% researchers in the world, underlining his commitment to world-class research methods across particular domains, ultimately influencing a generation of learners. Currently, Jason teaches and publishes on data analysis "best practices" in quantitative and applied research methods. He has served as evaluator or consultant on research projects and in public education (K-12), instructional technology, health care, medicine and business. He served as founding editor of Frontiers in Quantitative Psychology and Measurement and has been on the editorial boards of several other journals (such as Practical Assessment, Research, and Evaluation). Jason W Osborne also publishes on identification with academics and on issues related to social justice and diversity. He has written seven books covering topics to communicate logistic regression and linear modeling, exploratory factor analysis, best practices and modern research methods, data cleaning, and numerous other topics.