Regression Analysis A Practical Introduction 1st Edition Jeremy Arkes Download PDF

Download the full version of the textbook now at textbookfull.
com
Regression Analysis A Practical Introduction

1st Edition Jeremy Arkes
https://textbookfull.com/product/regression-
analysis-a-practical-introduction-1st-edition-
jeremy-arkes/
Explore and download more textbook at https://textbookfull.com

Recommended digital products (PDF, EPUB, MOBI) that
you can download immediately if you are interested.
Discourse Analysis: A Practical Introduction 1st Edition

Canning
https://textbookfull.com/product/discourse-analysis-a-practical-
introduction-1st-edition-canning/
textbookfull.com
Discourse Analysis: A Practical Introduction 1st Edition

Canning
https://textbookfull.com/product/discourse-analysis-a-practical-
introduction-1st-edition-canning-2/
textbookfull.com
Introduction to Mediation Moderation and Conditional

Process Analysis A Regression based Approach Andrew F.
Hayes
https://textbookfull.com/product/introduction-to-mediation-moderation-
and-conditional-process-analysis-a-regression-based-approach-andrew-f-
hayes/
textbookfull.com
Managing Risk in Nanotechnology Topics in Governance

Assurance and Transfer 1st Edition Finbarr Murphy
https://textbookfull.com/product/managing-risk-in-nanotechnology-
topics-in-governance-assurance-and-transfer-1st-edition-finbarr-
murphy/
textbookfull.com
Mathematical programming with data perturbations II 2nd
Edition Fiacco.
https://textbookfull.com/product/mathematical-programming-with-data-
perturbations-ii-2nd-edition-fiacco/
textbookfull.com
Heal Your Ancestors to Heal Your Life 1st Edition Shelley

A Kaehr
https://textbookfull.com/product/heal-your-ancestors-to-heal-your-
life-1st-edition-shelley-a-kaehr/
textbookfull.com
Communication: Innovation & Quality Miguel Túñez-López
https://textbookfull.com/product/communication-innovation-quality-
miguel-tunez-lopez/
textbookfull.com
Fitness for Work: The Medical Aspects 6th Edition John

Hobson (Editor)
https://textbookfull.com/product/fitness-for-work-the-medical-
aspects-6th-edition-john-hobson-editor/
textbookfull.com
Holographic Entanglement Entropy 1st Edition Mukund

Rangamani
https://textbookfull.com/product/holographic-entanglement-entropy-1st-
edition-mukund-rangamani/
textbookfull.com
Mechanics of Rotor Spinning Machines First Edition Prof.
Dr. Eng. Ibrahim Abdou Elhawary
https://textbookfull.com/product/mechanics-of-rotor-spinning-machines-
first-edition-prof-dr-eng-ibrahim-abdou-elhawary/
textbookfull.com
Regression Analysis
With the rise of “big data,” there is an increasing demand to learn the skills needed to undertake sound
quantitative analysis without requiring students to spend too much time on high-level math and
proofs. This book provides an efficient alternative approach, with more time devoted to the practical
aspects of regression analysis and how to recognize the most common pitfalls.
By doing so, the book will better prepare readers for conducting, interpreting, and assessing
regression analyses, while simultaneously making the material simpler and more enjoyable to learn.
Logical and practical in approach, Regression Analysis teaches: (1) the tools for conducting regressions;
(2) the concepts needed to design optimal regression models (based on avoiding the pitfalls); and
(3) the proper interpretations of regressions. Furthermore, this book emphasizes honesty in research,
with a prevalent lesson being that statistical significance is not the goal of research.
This book is an ideal introduction to regression analysis for anyone learning quantitative methods
in the social sciences, business, medicine, and data analytics. It will also appeal to researchers and
academics looking to better understand what regressions do, what their limitations are, and what they
can tell us. This will be the most engaging book on regression analysis (or Econometrics) you will
ever read!
Jeremy Arkes is Associate Professor at the Graduate School of Business and Public Policy,
Naval Postgraduate School, U.S.A. He conducts research in a variety of fields, with a focus on
military-manpower policy, substance-use policy, determinants of youth outcomes, sports economics,
and using sports outcomes to make inferences on human behavior.
Regression Analysis
A Practical Introduction
Jeremy Arkes
First published 2019
by Routledge
2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN
and by Routledge
52 Vanderbilt Avenue, New York, NY 10017
Routledge is an imprint of the Taylor & Francis Group, an informa business
© 2019 Jeremy Arkes
The right of Jeremy Arkes to be identified as author of this work has been asserted
by him in accordance with sections 77 and 78 of the Copyright, Designs and
Patents Act 1988.
All rights reserved. No part of this book may be reprinted or reproduced or
utilised in any form or by any electronic, mechanical, or other means, now
known or hereafter invented, including photocopying and recording, or in any
information storage or retrieval system, without permission in writing from the
publishers.
Trademark notice: Product or corporate names may be trademarks or registered
trademarks, and are used only for identification and explanation without intent
to infringe.
British Library Cataloguing-in-Publication Data
A catalogue record for this book is available from the British Library
Library of Congress Cataloging-in-Publication Data
A catalog record for this book has been requested
ISBN: 978-1-138-54140-5 (hbk)
ISBN: 978-1-138-54143-6 (pbk)
ISBN: 978-1-351-01109-9 (ebk)
Typeset in Bembo
by Apex CoVantage, LLC
Visit the eResources: www.routledge.com/9781138541405

To my mother, Judy Arkes (1941–2014). She was always a voice of love,
support, and reason.
Contents
List of figures xi
List of tables xiii
About the author xv
Preface xvi
Acknowledgments xviii
List of abbreviations xix
1 Introduction 1
1.1 The problem 2
1.2 The purpose of research 3
1.3 What causes problems in the research process? 4
1.4 About this book 7
1.5 The most important sections in this book 9
1.6 Quantitative vs. qualitative research 10
1.7 Stata and R code 11
1.8 Chapter summary 11
2 Regression analysis basics 13

2.1 What is a regression? 14
2.2 The four main objectives for regression analysis 15
2.3 The Simple Regression Model 17
2.4 How are regression lines determined? 21
2.5 The explanatory power of the regression 26
2.6 What contributes to slopes of regression lines? 29
2.7 Using residuals to gauge relative performance 30
2.8 Correlation vs. causation 32
2.9 The Multiple Regression Model 33
2.10 Assumptions of regression models 35
2.11 Calculating standardized effects to compare estimates 38
2.12 Causal effects are “average effects” 39
viii | Contents
2.13 Causal effects can change over time 40

2.14 A quick word on terminology for regression equations 41
2.15 Definitions and key concepts 42
3 Essential tools for regression analysis 47

3.1 Using binary variables (how to make use of dummies) 47
3.2 Non-linear functional forms using OLS 50
3.3 Weighted regression models 57
4 What does “holding other factors constant” mean? 61

4.1 Case studies to understand “holding other factors constant” 62
4.2 Using behind-the-curtains scenes to understand “holding other
factors constant” 68
4.3 Using dummy variables to understand “holding other factors constant” 72
4.4 Using Venn diagrams to understand “holding other factors constant” 76
4.5 Could controlling for other factors take you further from the true causal effect? 76
4.6 Application of “holding other factors constant” to the story of oat bran
and cholesterol 78
5 Standard errors, hypothesis tests, p-values, and aliens 82

5.1 Setting up the problem for hypothesis tests 83
5.2 Hypothesis testing in regression analysis 86
5.3 The drawbacks of p-values and statistical significance 101
5.4 What the research on the hot hand in basketball tells us about the existence
of other life in the universe 104
5.5 What does an insignificant estimate tell you? 105
5.6 Statistical significance is not the goal 106
6 What could go wrong when estimating causal effects? 111

6.1 How to judge a research study 113
6.2 Exogenous (good) variation vs. endogenous (bad) variation 114
6.3 Setting up the problem for estimating a causal effect 115
6.4 The big questions for what could bias the coefficient estimate 117
6.5 How to choose the best set of control variables (model selection) 146
6.6 What could bias the standard errors and how do you fix it? 149
6.7 What could affect the validity of the sample? 156
6.8 What model diagnostics should you do? 158
6.9 Make sure your regression analyses/interpretations do no harm 159
6.10 Applying the big questions to studies on estimating divorce effects on children 160
6.11 Applying the big questions to nutritional studies 161
6.12 Chapter summary: a review of the big questions163
Contents | ix
7 Strategies for other regression objectives 169

7.1 Strategies for forecasting/predicting an outcome 169
7.2 Strategies for determining predictors of an outcome 172
7.3 Strategies for adjusting outcomes for various factors 176
7.4 Summary of the strategies for each regression objective 180
8 Methods to address biases 183

8.1 Fixed-effects 184
8.2 A thorough example of fixed effects 196
8.3 An alternative to the fixed-effects estimator 200
8.4 Random effects 201
8.5 First-differences 203
8.6 Difference-in-differences 207
8.7 Two-stage Least Squares (instrumental-variables) 211
8.8 Regression discontinuities 217
8.9 Case study: research on how divorce affects children 220
8.10 Knowing when to punt 225
9 Other methods besides Ordinary Least Squares 231

9.1 Types of outcome variables 232
9.2 Dichotomous outcomes 233
9.3 Ordinal outcomes – ordered models 239
9.4 Categorical outcomes – Multinomial Logit Model 241
9.5 Censored outcomes – Tobit models 243
9.6 Count variables – Negative Binomial and Poisson models 245
9.7 Duration models 247
10 Time-series models 252

10.1 The components of a time-series variable 253
10.2 Autocorrelation 253
10.3 Autoregressive models 256
10.4 Distributed-lag models 262
10.5 Consequences of and tests for autocorrelation 266
10.6 Stationarity 269
10.7 Vector Autoregression 274
10.8 Forecasting with time series 275
11 Some really interesting research 281

11.1 Can discrimination be a self-fulfilling prophecy? 281
11.2 Does Medicaid participation improve health outcomes? 286
11.3 Estimating peer effects for academic outcomes 287
11.4 How much does a GED improve labor-market outcomes? 290
Visit https://textbookfull.com
now to explore a rich
collection of eBooks, textbook
and enjoy exciting offers!
x | Contents
12 How to conduct a research project 293

12.1 Choosing a topic 294
12.2 Conducting the empirical part of the study 295
12.3 Writing the report 298
13 Summarizing thoughts 305

13.1 Be aware of your cognitive biases 305
13.2 What betrays trust in published studies 307
13.3 How to do a referee report responsibly 310
13.4 Summary of the most important points and interpretations 311
13.5 Final words of wisdom 313
Appendix of background statistical tools 315

A.1 Random variables and probability distributions 315
A.2 The normal distribution and other important distributions 322
A.3 Sampling distributions 324
A.4 Desired properties of estimators 328
Glossary 330
Index 339
Figures
2.1 Scatterplot of years-of-schooling and income 15

2.2 Data points for Table 2.2 22
2.3 The data points relative to mean income for a four-observation regression 24
2.4 The regression line and residuals 25
2.5 Various combinations of R2 and regression slope 27
2.6 Average income by years-of-schooling 29
2.7 Wins and payroll, Major League Baseball, 2013 31
2.8 Summary of Chapter 2 45
3.1 Income by highest degree 50
3.2 Notional example of study hours and test score 54
3.3 Predicted values for the linear, quadratic, and spline models 57
4.1a Flow chart showing mechanisms for lemon-tree example when excluding “height” 64
4.1b Flow chart showing mechanisms for lemon-tree example when including “height” 64
4.2 Mechanisms for how tipping the first domino affects whether the last domino falls 67
4.3a The behind-the-curtains operations of a regression 69
4.3b What happens behind the curtains when variables are controlled for 70
4.4 Using Venn diagrams to demonstrate “holding other factors constant” 77
4.5 Notional scatterplot of oat-bran consumption and cholesterol 78
5.1 Hypothesis test example for a sample mean 85
5.2 Venn diagrams showing determinants of standard errors 90
5.3 Rejection region for Student’s t-distribution (d.f. = 2762) 93
5.4 Rejection region for a one-sided hypothesis test, Student’s t-distribution (d.f. = 2762) 96
5.5 Summary of Chapter 5 108
6.1 A demonstration of reverse causality with marijuana use and depression 118
6.2 Demonstrating reverse causality in Exogenous and Endogenous Worlds 120
6.3 Reverse causality in football 122
6.4 Omitted-variables bias for the relationship between oat bran and cholesterol 125
6.5 What people eat for breakfast 129
6.6 A demonstration of self-selection bias 132
6.7 Notional example to demonstrate measurement error 135
xii | Figures
6.8 Demonstration of mediating factors 141

6.9 What happens if one controls for a mediating factor 143
6.10 The effect of campaign donations on voting behavior 144
6.11 Using Venn diagrams to demonstrate multicollinearity 150
6.12 A notional example of homoskedasticity 153
6.13 A sample of real data on AFQT and income, demonstrating heteroskedasticity 153
8.1a The cross-sectional representation of the class-size and instructor-evaluation issue 198
8.1b The fixed-effects representation of the class-size and instructor-evaluation issue 198
8.2 The first-difference model for the unemployment rate and teen marijuana use 205
8.3 Difference-in-difference notional example 208
8.4 The problem with omitted variables for IV method setup 212
8.5 The two-stage-least-squares method 213
8.6 Graphical representation of a notional regression-discontinuity model 218
9.1 Frequency of number of children born in first 5 years of NLSY survey 246
10.1 Decomposition of a time-series variable (quarterly gas prices) 254
10.2 San Antonio Spurs winning percentage and change in winning percentage, by year 255
10.3 President Trump’s weekly tweets while president (through April 2018) 256
10.4a Scatterplot of current versus lagged winning percentage for the Spurs 259
10.4b Scatterplot of current versus lagged change in winning percentage for the Spurs 259
10.5a Autocorrelation function (ACF) of the number of weekly Trump tweets 261
10.5b Partial autocorrelation function (PACF) of the number of weekly Trump tweets 261
10.6a The autocorrelation function for U.S. wine consumption 262
10.6b The partial autocorrelation function for U.S. wine consumption 263
10.7a Daily prices for the S&P 500 and oil, 1990–94 272
10.7b Daily prices for the S&P 500 and oil, 2010–14 273
A.1 Probability distribution for the value of a roll of a die 316
A.2 Probability density function of a (continuous) uniform distribution 317
A.3 A cumulative distribution function for a uniform distribution 317
A.4 A non-standard continuous distribution 318
A.5 Sampling distributions of the sample mean for various sample sizes
(for the Central Limit Theorem) 326
Tables
2.1 The four main objectives of regression analysis and the questions addressed 16
2.2 Four-observation example of education and income 21
2.3 Predicted values and residuals for the four-observation sample 25
2.4 A summary of regression assumptions, how reasonable they are, and what action to take 38
2.5 Standardized effect sizes of determinants of the Romanian Peasant Rebellion 39
3.1 Interpretations based on logarithmic forms of the X and Y variables 53
3.2 A comparison of models using various weighting schemes 59
4.1 A summary of the two cases for the lemon-tree example 65
4.2 Notional data on dominos chains to understand “holding other factors constant” 66
4.3 Demonstrating coefficient estimates in the presence of dummy variables 74
4.4 Average income for each group 75
4.5 Average real GDP growth (and defined years) for each group 76
5.1 Stata output for regression of 2003 income on various factors 87
5.2 R output for regression of 2003 income on various factors 88
5.3 The probability that an estimated relationship is real for various p-values and prior
probabilities 102
6.1 Exogenous and endogenous variation in an explanatory variable 115
6.2 Demonstration of the biases from reverse causality 121
6.3 Likely biased (from omitted-variables bias) estimates of top marginal tax rates on
the unemployment rate and real GDP growth rate, 1991–2017 126
6.4 An example for self-selection bias 131
6.5 Regression models demonstrating the effects of measurement error 136
6.6 The effects of state tax rates on Gross State Product growth (n = 1440), 1980–2009 146
6.7 Criteria to use and not to use for determining what control variables to include 148
6.8 Demonstrating the effects of multicollinearity (dependent variable = 2003 income) 150
6.9 Results on how the economy affects alcohol use – from Ruhm and Black (2002) 151
6.10 The correction for heteroskedasticity (n = 2772) 154
6.11 A summary of the 6 big questions 164
7.1 Predicting health status at age 50 for females 174
7.2 Strategies and which big questions need to be asked for the regression objectives 180
xiv | Tables
8.1 The effects of class size on average instructor evaluation from Bedard and Kuhn (2008) 191
8.2 Course characteristics in the simulation 197
8.3 Coefficient estimates on class-size for various models, plus variance within groups 199
8.4 A comparison of a standard and regression-weighted fixed-effects model 201
8.5 The effects of the unemployment rate on youth marijuana use 206
8.6 Summarizing results from Card and Krueger’s (1994) Table 3 (Numbers represent
average FTE workers at fast-food establishments) 210
8.7 The effectiveness and applicability of the methods to address non-random
explanatory variables 226
9.1 Determinants of “whether the youth used marijuana in past 30 days,” using three
different probability models 236
9.2 Demonstration of odds ratios for the probability of marijuana use 238
9.3 Ordered Probit results for “health at age 40” (n = 7705) 240
9.4 Marginal effects from Ordered Probit Model 240
9.5 Coefficient estimates of a marital disruption on problem behavior 244
9.6 Results from Negative Binomial, Poisson, and OLS 247
9.7 Cox proportional-hazards model for the effects of state unemployment rates on
the probability of a couple divorcing (n = 95,472) 249
10.1 Autoregressive models for President Trump weekly tweets 258
10.2 AR(1) model for the Spurs winning percentage and change in winning percentage 260
10.3 The relationship between approval ratings and tweets for President Trump 264
10.4 Granger Causality model for Trump’s weekly tweets and approval rating 265
10.5 Dickey-Fuller tests for nonstationarity 273
10.6 The relationship between oil prices and the S&P 500 index 273
10.7 Vector Autoregression (VAR) for Trump tweets and approval 275
10.8 ARIMA models for daily S&P 500, 2010–14 277
10.9 Forecasts and statistics for first 5 trading days of 2015 for the ARIMA(1,1,1) model 278
11.1 Key results from Glover et al. (2017), Tables III and IV 285
11.2 Summary of the main results from Carrell et al. (2009), Table 3 290
A.1 Example of a joint probability distribution 320
About the author
Jeremy Arkes grew up in Amherst, MA, in the midst of the fields of Amherst College. He left this
bucolic setting to attend Georgetown University for his undergraduate studies and to later earn a Ph.D.
in Economics from the University of Wisconsin. He spent his first ten years after graduate school
working for think tanks: The Center for Naval Analyses (Alexandria, VA) and RAND C orporation
(Santa Monica, CA). His main focus was on military-manpower research, but he sprouted out to other
fields, such as the areas of substance use and divorce effects on children.
Since 2007, he has been teaching Economics and Applied (Regression) Analysis of Military
Manpower in the Graduate School of Business and Public Policy at the Naval Postgraduate School in
Monterey, California. At NPS, besides continuing to conduct research in military manpower, he has
added a line of research that uses sports outcomes to make inferences on human behavior.
When not estimating regression models or writing about them, Dr. Arkes plays basketball and
racquetball, and he hikes around the Big Sur and California Central Coast area with his wife, Jennifer.
And, whenever they get a chance, he and Jennifer like to venture off to New England, the Rockies,
the Cascades, British Columbia, Hawaii, or Alaska for hiking adventures.
Preface
I’ve played a lot of basketball in my life. I was mediocre in high school, being the 6th man (first guy
off the bench) on my high-school team. At 5'9", I played all positions other than point guard (even
center briefly, as I knew how to box out) . . . I could defend, jump, and rebound, but I couldn’t dribble
or shoot well.
Although I never developed a consistent outside shot until my 30s (and 3-point range in my
mid-40s), it was at college (Georgetown University) where I learned to dribble, drive to the basket,
and shoot mid-range jumpers – don’t mind saying this helped me win a few 1-on-1 and 3-on-3
tournaments. And, it was in the Georgetown gym where I started to have experiences with the “hot
hand,” especially on those 95-degree-and-humid spring and summer days. The hot hand is a period
of playing at a systematically higher level than what you normally play at, often called “being in the
zone” or “en fuego.” Those of you who play sports may have experienced it in whatever sport you play.
My wife says she has experienced it in her dancing career. For me, some days, I just had a better feel
for the ball, and I was more likely to make any type of shot I tried.
And, it is in one of those hot-hand events for myself that I briefly got the better of 7'2", future
NBA-Hall-of-Famer, Dikembe Mutombo. (If you ever meet me and ask, I’ll gladly tell you the story.)
So, you can imagine my surprise when I read that a bunch of famous economists, statisticians, and
psychologists (including some Nobel Prize winners), were arguing, based on several studies that failed
to find any evidence for the hot hand, that the hot hand is a figment of our imaginations. Many of
them stated things along the lines of “The hot hand in basketball is a myth.”
How could they conclude this? I had to investigate it on my own, applying some methods to try
to increase the “power” of the test. Sure enough, in 2010, I found some evidence for the existence
of the hot hand and published it. But, the real breakthroughs came in the next several years, with
research by my friends Dan Stone (from Bowdoin College) and Josh Miller and Adam Sanjurjo (both
from Universidad de Alicante). They found biases in the original studies (and mine) that would work
against being able to detect the hot hand. Josh and Adam’s most notable bias is from the Gambler’s
Fallacy (which I’ll discuss briefly in Chapter 6). The bias Dan found (from measurement error) was
more intuitive, as measurement error should be on any checklist for potential biases in a study. But, no
one (including myself) had recognized it.
How could those famous researchers and Nobel Prize winners have made such huge errors?
It wasn’t just that they missed these sources of bias. But, they also used fallacious logic in their
Preface | xvii
interpretations: the lack of evidence for the hot hand was not proof that it did not exist. (I’ll discuss
this more in Chapter 5).
This was a watershed moment for me that so many established researchers bought into this argu-
ment without critically assessing it. And, while I give myself a pass for missing the large biases that Josh
and Adam found, I was guilty for not recognizing the measurement-error bias that Dan had found.
And so, this story arc on the hot hand confirmed a pattern I was seeing on other research topics,
and it hit me.Yeah, most economists know the lingo and the formulas for why things could go wrong
in regression analysis. But, many aren’t trained well in recognizing these when they occur. And, many
do not understand the proper way to interpret the results of regressions. I thought that there must be
a better way to teach regression analysis.
It was at this point that I realized that I had a story to tell. I have come to understand many con-
cepts on regression analysis that elude many researchers. These concepts came to me, not from my
classes, but from my practical experience of estimating thousands of regressions . . . and from the many
mistakes I have made in my research. And, the lessons I learned from poring over results helped me
connect the dots between several of the essential concepts.
And so, from these experiences with various pitfalls of regression analysis, from mistakes I have
made, from what helped me connect the dots between concepts, I have created a story on how to
better understand, conduct, and scrutinize regression analysis.
Acknowledgments
I would first like to thank my wife, Jennifer. For much of my writing of this book, she was a doctoral
student in Education, and we had numerous conversations about the methods I used and her per-
spectives on research from a curriculum focusing on qualitative research. I would also like to thank
several of my colleagues/friends. Robert Eger was instrumental to me for his guidance on the process,
and his encouragement to take a shot. For technical details, I benefited greatly from discussions with
Thomas Ahn (Naval Postgraduate School), William A. Muir (U.S. Air Force and Naval Postgradu-
ate School), John Pepper (University of Virginia), and my hot-hand-in-basketball colleagues, Joshua
Miller (Universidad Alicante), Adam Sanjurjo (Universidad Alicante, and also my surf instructor), and
Daniel Stone (Bowdoin College). And, for telling me when some of my attempts at humor worked
and when they didn’t, I am grateful to my friends at work: Ronda Spelbring, Houda Tarabishi, and
Holly Shrewbridge.
The views expressed in this book are mine and do not reflect those of the Naval Postgraduate
School, the U.S. Navy, the U.S. Department of Defense, or the U.S. government.
Abbreviations
2SLS Two-stage Least Squares

ACF Autocorrelation function
AFA Air Force Academy
AFQT Armed Forces Qualification Test
AIC Akaike Information Criterion
AICc Akaike Information Criterion (corrected)
AR Autoregressive (model)
ARIMA Autoregressive integrated moving average
ARMA Autoregressive moving average
ATE Average Treatment Effect
ATT Average Treatment effect for the Treated
BIC Bayesian Information Criterion
BPI Behavioral Problems Index
CDF Cumulative distribution function
CEO Chief Executive Officer
CPS Current Population Survey
CS Class size
DD Difference-in-difference (model)
d.f. degrees of freedom
EITC Earned Income Tax Credit
ExSS Explained Sum of Squares
FD First-difference (model)
FTE Full-time-equivalent (worker)
GDP Gross Domestic Product
GED General Equivalency Diploma
GPA Grade point average
HS High School
IAT Implicit Association Test
i.i.d. Identically distributed
IQ Intelligence Quotient
Visit https://textbookfull.com
now to explore a rich
collection of eBooks, textbook
and enjoy exciting offers!
xx | Abbreviations
IV Instrumental variable
LATE Local Average Treatment Effect
LPM Linear probability model
MAPE Mean absolute percent error
MJ Marijuana
MNL Multinomial Logit (Model)
MSE Mean Square Error
NBA National Basketball Association
NCAA National Collegiate Athletic Association
NFL National Football League
NIH National Institute of Health
NLSY National Longitudinal Survey of Youth
NRA National Rifle Association
NSDUH National Survey on Drug Use and Health
OLS Ordinary Least Squares
PACF Partial autocorrelation function
PDF Probability density function
PI/C Personal income per capita
PTSD Post-traumatic stress disorder
RCT Randomized-control trial
RD Regression discontinuity
RMSE Root Mean Square Error
RMSFE Root mean square forecast error
RP Relative performance
RR Restricted regression
RSS Residual Sum of Squares
SAT Scholastic Aptitude Test
SE Standard Error
SETI Search for Extraterrestrial Intelligence
SUR Seemingly Unrelated Regression
TS Test score
TSS Total Sum of Squares
UR Unrestricted regression
VAR Vector Autoregression
var ( ) Variance
Introduction
1
1.1 The problem

1.2 The purpose of research
1.3 What causes problems in the research process?
1.4 About this book
1.5 The most important sections in this book
1.6 Quantitative vs. qualitative research
1.7 Stata and R code
1.8 Chapter summary
If you don’t read the newspaper you are uninformed, if you do read the newspaper you are mis-
informed.
– Mark Twain
I’m on a mission, and I need your help.

My mission is to make this a better world.
I want society to make better choices and decisions. Sometimes, these decisions should be based
on what is morally and ethically the right thing to do. I cannot help on that front. But, at other times,
decisions need to be based on how they affect people and institutions. And, in that realm, sometimes
statistical analysis can speak to best practices. Statistical analysis sometimes can tell us what health prac-
tices and interventions are most beneficial to people, what factors lead to better economic outcomes
for people (individually or collectively), what factors contribute to the academic achievement of
children, how to make government functions more cost-efficient (which could reduce the tax burden
on society), and much more.
All that said, sometimes statistical analysis is unable to speak on many of these issues. This may be
because the data are not adequate in terms of having sufficient observations and sufficient accuracy.
Or, it could be because the statistical analysis was flawed, or that there are no solutions to certain
biases. So, we need to be careful in how we interpret and use the results from statistical analyses so
that we draw the correct and prudent conclusions, without overreaching or being affected by our
pre-conceived notions and biases.
2 | Introduction
My goal with this book is not to answer the important questions on how to make the world
better. In fact, I will address some research issues that some of you will care nothing about, such as
whether discrimination is a self-fulfilling prophecy in France, whether a Medicaid expansion in Ore-
gon improved health outcomes for participants, or whether the hot hand in basketball is real or just
a figment of our imaginations. But, these research issues that I use will serve as useful applications to
learn the concepts and tools of regression analysis.
So, my goal is to teach you the tools needed to address important issues. This book is designed to
teach you how to better conduct, interpret, and scrutinize statistical analyses. From this, I hope you will
help others make better decisions that will help towards making this world, eventually, a better place.
1.1 The problem

Jay Leno, in one of his Tonight Show monologues several years ago, mentioned a study that found that
50% of all academic research is wrong. His punchline: there’s a 50% chance this study itself is wrong.
The study Leno referred to may actually understate the true percentage of studies that are inaccu-
rate. The major causes of all these errors in research are likely faulty research designs and improper
interpretations of the results. These accuracy issues bring into doubt the value of academic research.
Most quantitative academic research, particularly in the social sciences, business, and medicine, rely on
regression analyses. The primary objective of regressions is to quantify cause-effect relationships. These
cause-effect relationships are part of the knowledge that should guide society to develop good public
policies and good strategies for conducting business, educating people, promoting health and general
welfare, and more. Regressions are useful for estimating such relationships because they are able to “hold
constant” other factors that may confound the cause-effect relationship in question. That is, regressions,
if done well, can rule out reasons for two variables to be related, other than the causal-effect reason.
Here are some examples of how regressions can be used to estimate the causal effect of one factor
(or a set of factors) on some outcome:
• How does some new cancer drug affect the probability of a patient surviving 10 years after diag-
nosis?
• How does a parental divorce affect children’s test scores?
• What factors make teachers more effective?
• What encourages people to save more for retirement?
• What factors contribute to religious extremism and violence?
• How does parental cell phone use affect children’s safety?
• How does oat-bran consumption affect bad cholesterol levels?
• Do vaccines affect the probability of a child becoming autistic?
• How much does one more year of schooling increase a person’s earnings?
• Does smiling when dropping my shirt off at the cleaners affect the probability that my shirt will
be ready by Thursday?
A regression is a remarkable tool in its ability to measure how certain variables move together, while
holding certain factors constant. A natural human reaction is to be mesmerized by things people do
not understand, such as how regressions can produce these numbers. And so, in the roughly 10 times
that I have used regression results in briefings to somewhat-high-level officials at the Department of
Introduction | 3
Defense (mostly as a junior researcher, with a senior researcher tagging along to make sure I didn’t say
anything dumb), the people I was briefing never asked me whether there were any empirical issues
with the regression analysis I had used or how confident I was with the findings. Most of the time,
based on the leading official’s response to the research, they would act as if I had just given them the
absolute truth on an important problem based on these “magic boxes” called “regressions.” Unfortu-
nately, I was caught up in the excitement of the positive response from these officials, and I wasn’t as
forthright as I should have been about the potential pitfalls (and uncertainty) in my findings. And so,
I usually let them believe the magic.
But, regressions are not magic boxes. The inaccuracy Leno joked about is real, as there are many
pitfalls of regression analysis. And, from what I have seen in research, at conferences, from journal
referees, etc., many researchers (most of whom have Ph.D.s) have a limited understanding of these
issues. And so, published quantitative research is often rife with severely biased estimates and erroneous
interpretations and conclusions.
How bad is it? In the medical-research field, where incorrect research has the potential to result
in lost lives, John Ioannidis has called out the entire field on its poor research methods and records.
The Greek doctor/medical-researcher was featured in a 2010 article in The Atlantic (Freedman 2010).
Ioannidis and his team of researchers have demonstrated that a large portion of the existing medical
research is wrong, misleading, or highly exaggerated. He attributes it to several parts of the research
process: bias in the way that research questions were being posed, how studies and empirical models
were set up (e.g., establishing the proper control group), what patients were recruited for the studies,
how results were presented and portrayed, and how journals chose what to publish.
Along these lines, the magazine The Economist had a much-needed op-ed and accompanying arti-
cle in 2013 on how inaccurate research has become.1 Among the highlights they note are:
• Amgen, a biotech company, could replicate only 6 of 53 “landmark” cancer-research studies;

• Bayer, a pharmaceutical company was able to replicate just one-quarter of 67 important health
studies;
• Studies with “negative results,” meaning insignificant estimated effects of the treatment variables,
constituted 30% of all studies in 1990 and just 14% today, suggesting that important results show-
ing no evidence that a treatment has an effect are being suppressed – and/or extra efforts are
being made to make results statistically significant.
All of this highlights an interesting irony. The potential for valuable research has perhaps never been
greater, with more data available on many important outcomes (such as student test scores, human
DNA, health, logistics, and consumer behavior, and ball and player movements in sports), yet the rep-
utation of academic research has perhaps never been so low.
This is fixable!
This book is meant to effectively train the next generation of quantitative researchers.
1.2 The purpose of research

To understand where research goes wrong, we first have to understand the overall purpose of research.
We conduct research to improve knowledge, which often involves trying to get us closer to under-
standing cause-effect and other empirical relationships. To demonstrate, let’s start with the highly
4 | Introduction
contentious issue of global warming.You may have some belief on the probability that the following
statement is true:
Human activity is contributing to global warming.
And, hopefully that probability of yours lies somewhere between 0.3% and 99.7% – that is, you may
have your beliefs, but you recognize that you probably are not an expert on the topic and so there is a
possibility that you are wrong. I’m guessing that most people would be below 10% or above 90% (or,
even 5% and 95%). But, for the sake of the argument, let’s say that you have a subjective probability of
the statement being true 45% of the time.
Suppose a study comes out that has new evidence that humans are causing global warming. This
may shift your probability upwards. If the new research were reported on cable news channel MSNBC
(which leans toward the liberal side of politics) and you tended to watch MSNBC, then let’s say that
it would shift your probability up by 7 percentage points (to 52%). If you tended to watch Fox News
(a more conservative channel) instead, then the news from MSNBC may shift your probability up by
some negligible amount, say 0.2 percentage points (up to 45.2%). But, ideally, the amount that your
subjective probability of the statement above would shift upwards should depend on:
• How the study contrasts with prior research on the issue;

• The validity and extensiveness of the prior research;
• The extent to which any viable alternative explanations to the current findings can be ruled
out – i.e., how valid the methods of the study are.
With regression analysis, it should be the same thinking of shifting beliefs. People have some prior
beliefs about some issue, say on whether class size is important for student achievement. Using regres-
sion analysis, a new study finds no evidence that class size has an effect on student achievement. This
finding should not necessarily be taken as concrete evidence for that side of the issue. Rather, the
evidence has to be judged based on the strength of the study relative to the strength of other studies, or
the three criteria listed above. And, people would then shift their subjective probability appropriately.
The more convincing the analysis, the more it should swing a person’s belief in the direction of the
study’s conclusions.
This is where it is up to researchers, the media, and the public to properly scrutinize research to
assess how convincing it is. As I will describe below, you cannot always rely on the peer-review process
that determines what research gets published in journals.
1.3 What causes problems in the research process?

The only real fiction is non-fiction.
– Mark Totten
Where do I begin?
Well, let’s discuss some structural issues first, which lead to misguided incentives for researchers.
One major problem in research is publication bias (discussed with more detail in Section 13.2),
which results from the combination of the pressure among academics to publish and journals seeking
Random documents with unrelated
content Scribd suggests to you:
PROTECTIONISM TWENTY
YEARS AFTER
34
I THINK it must be now nearly twenty years since I have made

dinner.
a free-trade speech or been able to take share in a free-trade
When I was invited here this evening I thought I would try to

come for the pleasure of hearing the gentlemen, especially the
members of Congress, who were announced to speak here. I have
been so out of health that it has been impossible for me to sit up
evenings or to attempt public speaking in the evenings, but things
are going a little better and I will make an attempt to say a little—
not very much, as the hour is now late.
Thirty-five or forty years ago I became a free trader for two
great reasons, as far as I can now remember.
One was because, as a student of political economy, my whole
mind revolted against the notion of magic that is involved in the
notion of a protective tariff. That is, there are facts that are
accounted for by protectionism through assertions that are either
plainly untrue or are entirely irrational. The other reason was
because it seemed to me that the protective tariff system nourished
erroneous ideas of success in business and produced immoral results
in the minds and hopes of the people.
I cannot say that I have got any more light on the matter within
the last twenty years; it looks to me still as if the great objections to
protectionism were these two. No man who enjoys the benefit of a
protective tariff, as he believes, can ever tell whether he gets back
anything for the taxes which he pays or not. He never has any
analysis of the operation and never knows whether or not he really
recovers from the action of the tariff what he pays in.
I say now the taxes which he pays, because—let us not make
any mistake about this—the matter we are talking about is one
entirely of Americans and between Americans. If the protective tariff
operates so as to perform what is attributed to it, it prevents things
from being imported into this country. That may be a disadvantage
to the foreigner, it may disappoint him in his hopes, but we may
leave him out of account. Then the increase of the cost of these
commodities for the American consumer at home is the source from
which the American protected manufacturer must obtain his benefit,
if he ever obtains any. Therefore he has to pay also taxes to the
other protected industries on account of the operation of the system.
Therefore he is both paying and receiving, but whether or not he
gets back the part that he hoped to receive is a question which he
never can sift and never can know.
I should myself suppose that possibly the Pennsylvanian on his
coal and iron might stand a good chance of winning something. The
operation is direct and simple in that case, and coal and iron are to-
day the very first conditions of industry. They must be obtained as
raw material, because they enter into everything, and it is possible
that under those circumstances the game might be sufficiently direct
so that its effect could be felt and perceived. But the Connecticut
manufacturer has to pay taxes on coal and iron and copper and the
other metals, and he has to pay also the taxes on wool and the
other raw materials, and then comes the question whether he ever
gets it back again or not. He never knows; he cannot know; he
cannot feel it and he cannot possibly know whether the operation of
the system is to bring him back a return for his outlay or not.
We hear a great deal about a rightly adjusted tariff. It is a
constant ideal that is presented, whenever the tariff subject comes
up again for discussion in Congress, that it ought to be rightly
adjusted, and when it is, it is going to perform its beneficial
operation.
How can a tariff ever be rightly adjusted unless the industry will
stand still? The taxes stand still for years without change. The
industries never stand still. There are new inventions in machinery,
there are new raw materials brought into use, there are new
processes developed, and all that changes the character of the
industry. These inventions and improvements and processes are all
ignored by the protective system. It contains no allowance for them
at all. But our people are full of enterprise, they are fond of
improvements, they like novelties, and they adopt changes. The
consequence is that the industry changes, and then again the
decisions that are made by somebody or other as to the doubtful
questions in the interpretation of the law are also constantly
changing, and then by and by we find a lot of people who want the
tariff changed. They say it needs to be adapted to the time, it is out
of date, it has fallen behind, it does not fit the requirements of the
moment, and they would like to have a tariff revision; but they are
told then that they ought to keep still and not make a disturbance
which will bring up a discussion of the entire tariff system, and that
they ought to allow it to go on for the sake of the “system.”
What is the system then? The system means that the import
duties that we have in this country have raised the prices of all
commodities in our market, I may say thirty or forty per cent on a
very low calculation. Is not that a very extraordinary thing when you
stand off and try to realize it for a minute—that we have raised the
prices in the United States thirty or forty per cent—perhaps more
nearly fifty per cent—above the level of the prices for the same
commodities in the other civilized countries of our grade; and that
we believe that we have done a grand and noble thing by raising
these prices, putting the whole level of life in this country on an
artificial plane that much above the level of the world’s market? In
fact, if you should listen to a protectionist he would make you
believe that this continent would not be habitable if it was not for
the protective tariff that is here working this operation all the time
on the American market.
I am of the opinion—I am not very confident about it—but it
looks to me as if it were true that a protective tariff wears out in a
little while—I mean, so far as its expected beneficial effect is
concerned. Its effects are distributed, they are taken up and they
are allowed for all around the market until the expected benefit to
the protected people is lost and there remains nothing but the dead
weight of the system itself as an interference with the industries.
There is then a call for a new tariff in order to get another impulse
or another fillip, as I have heard it called, to give things a new
impulse, to start them on again.
That has been the history of our tariff now for one hundred
years, that it has been restarted, reinvigorated from time to time in
order to give a new impulse. Then in the very nature of the case,
therefore, it seems to me that a new impulse is constantly required.
As I said at the outset, the tariff system seems to me to teach us
to believe that a man needs a “pull” of some kind or other to make
any industry a success. It is an idea that there must always be a
provision of easy profit in connection with the industry that shall
demand no labor or no expenditure of capital to get it. That is the
pure doctrine of graft. The tariff teaches us to look for a fee or a
gratuity or a rake-off which will be a pure and net profit. People are
told that tariff taxes are a rightful gift to the beneficiary. Those who
do not get that gain seek another one of the same kind somewhere,
and when they do that they have recourse to graft.
It is a shameful fact that this notion of graft, and this word,
should have come to us, as it has within the last four or five years,
and should have extended so far and become so familiar to us in
connection with a great many of the operations of business. It is
customary, as we have known for a long time, in some nations, for
instance in Russia, China, and Turkey; and with us it has seemed to
spread and win acceptance and currency in a most astonishing
manner. I cannot believe but what the tariff system has educated us
in this direction and prepared us to tolerate and accept the
development of this idea. It also seems to me that now, after one
hundred years of this system, the tariff is no longer properly an
economic question. It is a practical political question. The politics
and the business are interwoven in it inextricably. There is no
economic discussion possible of the propositions that are made,
economic in form, in connection with the tariff system. There is only
a war of partial views and of superficial inferences.
Our American protectionism has grown out of the peculiar
circumstances of this country. It is an old idea that has come down
to us from Europe, and, indeed, from the Middle Ages in Europe,
and here it found a chance for a new and very remarkable
development. There were new conditions here, and the chances
were so big and grand that, as a matter of fact, the protective
system has never done more than exact a certain tribute from us on
these chances. It has never really touched us in an acute and
sensible way, and in spite of it we have enjoyed marvelous
prosperity which is due really to the circumstances of advantage and
favor which we have enjoyed here.
In the year 1892 we got an issue on this matter and went to the
electorate with it, with the result that we all know. But the mandate
of the people was neglected and disobeyed by the government and
the purpose that the people showed at that time was defied.
We have also had opportunity to notice the great power of the
protected interests in Congress. The fact is that we are being
governed at the present time by a combination of these protected
interests which have got control of the machinery of government,
and have control of the personnel of the government to such an
extent that it is almost impossible, practically, to make any breach in
this system at all. That is because the political combinations have
been so thoroughly wrought out and so ingeniously developed that
they look at present as if they were impregnable.
I look around to see if I can find some encouragement. I
thought that it was something of an encouragement when Mr. Dalzell
made this speech in Congress that Mr. Williams has referred to, in
which he poured such scorn on the idea of “incidental protection.” I
have never said anything so severe about any protectionist idea as
that which he said about incidental protection. But suppose that the
people of 1850, the middle of the nineteenth century, could come to
life again, the old protectionists of that time. What would they think
to hear a man speak with scorn of incidental protection? It was what
they believed in; it was the whole business to them. When an old
protectionist like Mr. Dalzell can turn around and pour scorn upon
incidental protection I feel as if we never could tell what they might
throw overboard next time, in some paroxysm of some kind or other,
of fear or hope or something else, and we might get a chance that
we have not been able to get in the past.
Then, as has been well said by other gentlemen to-night, there
has been within the last year or two a very great revolt in the public
mind against graft and political and business corruption. How far will
this go? We do not know, but it is, at any rate, an opening in the
public mind that is full of chances. It may go very far; it may have
very great effects; it is certainly something to be noticed and taken
advantage of.
Then, again, there are new conflicts of interests arising. We have
become very great people in the world’s commerce, with a billion
dollars’ worth of exports and imports in a year, and we are so
interwoven with the whole world that it will not be possible for us to
go on with our old policy of discouraging commerce and rejecting it,
and trying to stop it, and paying no attention at all to the
remonstrances of our neighbors. In future we shall be obliged to pay
some attention to these remonstrances. They are just, they are
reasonable, and they will command our attention; and then we shall
have to make concessions to them. In other words, we cannot any
longer afford to reject and neglect these remonstrances.
It may be, therefore, that in the time that is now before us we
shall have better chances for a practical war upon this system than
we have had hitherto. As long, however, as I can remember, and as
long as I have had any share in it, we have got along without any
encouragement in it at all. We have done what we could without
that. We got so we did not expect it. We knew that we should be
neglected and treated as persons whose opinions in these matters
were not of any importance or worthy of any attention, and so we
went on and kept up our arguments, as we considered them, to the
best of our ability and without very much result.
Now, it may be that we are on the eve of a different time, when
the circumstances will be more favorable, more hopeful, more full of
opportunities, and I certainly, for my part, most profoundly hope
that that is so.
I have noticed with some discouragement the efforts that Mr.
Williams has made on the floor of Congress to get some
modifications of the tariff made, or some argument even opened up
there that might give the matter activity and life in the legislative
domain. They did not seem any more encouraging than what we
used to see in the old times. But it is certainly in the nature of things
that the difficulties and absurdities of this system must come out in
practice more and more distinctly as we go on, and the need for
reform will therefore force itself in the shape of a play of interests
that will bring new and counteracting forces into operation to which
we may look for help in the overthrow of the system.
PROSPERITY STRANGLED BY
GOLD
35
S OME of the silver fallacies were stated by Mr. St. John, in his
address before the silver convention, with such precision that
his speech offers a favorable opportunity for dealing with them.
He says that “it is amongst the first principles in finance that the
value of each dollar, expressed in prices, depends upon the total
number of dollars in circulation.” There is no such principle of finance
as the one here formulated. The “quantity doctrine” of currency is
gravely abused by all bimetallists, from the least to the greatest, and
it is at best open to great doubt. When the dollars in question are
dollars of some money of account which can circulate beyond the
territory of the State in which it is issued, the quantity doctrine
cannot be true within that territory. It may be noted, in passing, that
this is the reason why no scheme of the silver people for
manipulating prices in the United States can possibly succeed. Silver
and gold will be exported and imported until their values conform
throughout the world, and prices fixed in one or the other of them
will conform to the world’s prices, after all the trouble and waste and
loss of translating them two or three times over have been endured.
The quantity doctrine, however, means that the value of the
currency is a question of supply and demand, and everybody knows
that to double or halve the supply does not halve or double the
value, or have any other effect which is simple and direct. If it did
have such effect speculation would not be what it is.
Mr. St. John goes on to argue that our population increases two
millions every year, on account of which we need more dollars; that
the production of gold does not furnish enough to meet this need,
and that, therefore, prices fall. This argumentation is very simple
and very glib. Prosperity and adversity are put into a syllogism of
three lines. But, if we can avert the fall in prices and adversity by
coining silver, it must be by adding the silver to the gold which we
now have. “High” and “low” prices are only relative terms. They
mean higher and lower than at another time or place; higher and
lower than we have been used to. If misery depends on ten-cent
corn we are advised to cut the cents in two and we shall get twenty-
cent corn and prosperity. Corn will not be altered in value in gold, or
outside of the United States, and, as all other things will be marked
up at the same time and in the same way, its value in other things
will not be altered by this operation. When we get used to twenty-
cent corn it will seem just as low and just as “hard for the debtor” as
ten-cent corn is now. Then we can divide by ten and get two-dollar
corn, by adding free coinage of copper. When we get used to that
we shall be no better satisfied with it. We can then make paper
dollars and coin them without limit. Million-dollar corn will then
become as bitter a subject for complaint as ten-cent corn is now.
The fact that people are discontented is no argument for anything.
The fact that prices are low is made the subject of social
complaint and of political agitation in the United States. Prices have
undergone a wave since 1850. They arose until about 1872. They
have fallen again. They are lower than they were at the top of the
wave all the world over. This fact, the explanation of which would
furnish a very complicated task for trained statisticians and
economists, is made a topic of easy interpretation and solution in
political conventions and popular harangues, and it is proposed to
adopt violent and portentous measures upon the basis of the
flippant notions which are current about it. But what difference does
it make whether the “plane” of prices is high or low? If corn is at
forty cents a bushel and calico at twenty cents a yard, a bushel buys
two yards. If corn is at ten cents a bushel and calico at five cents a
yard, a bushel will buy two yards. So of everything else. If, then,
there has been a general fall, and that is the alleged grievance,
neither farmers nor any other one class has suffered by it.
It is undoubtedly true that a period of advancing prices
stimulates energy and enterprise. It does so even when, if all the
facts were well known, it might be found that capital was really
being consumed in successive periods of production. Falling prices
discourage enterprise, although, if all facts were known to the
bottom, it might be found that capital was being accumulated in
successive periods of production.
It is also true that a depreciation of the money of account, while
it is going on, stimulates exports and restrains imports.
But who can tell how we are to make prices always go up,
unless by constant and unlimited inflation? Who can tell how we are
to avoid fluctuations in prices or eliminate the element of
contingency, risk, foresight, and speculation?
It is also true that, although high prices and low prices are
immaterial at any one time, the change from one to the other, from
one period of time to another, affects the burden of outstanding time
contracts. Men make contracts for dollars, not for dollar’s-worths.
Selling long or short is one thing; lending is another. Borrowers and
lenders never guarantee each other the purchasing power of dollars
at a future time. If the contracts were thus complicated they would
become impossible. Between 1850 and 1872 the debtors made no
complaint and the creditors never thought of getting up an agitation
to have debts scaled up. The debtors now are demanding that they
be allowed to play heads I win, tails you lose, and Mr. St. John and
others tell us that they have the votes to carry it; as if that made
any difference in the forum of discussion.
Increase in population does not prove an increased need of
money. It may prove the contrary. If the population becomes more
dense over a given area, a higher organization may make less
money necessary. If railroads and other means of communication
are extended, money is economized. If banks and other credit
institutions are multiplied, and if credit operations are facilitated by
public security, good administration of law, etc., less money is
needed. If these changes are going on at the same time that
population is increasing (and such is undoubtedly the case in the
United States), who can tell whether the net result is to make more
or less currency necessary? Nobody; and all assertions about the
matter are wild and irresponsible.
If it was true that an increase of two millions in the population
called for more dollars, how does anybody know whether the current
gold production is adequate to meet the new requirement or not?
The assertion is arithmetical. It says that two quantities are not
equal to each other. The first quantity is the increase in the currency
called for by two million more people. How much more is needed?
Nobody knows, and there is no way to find out. The silver men have
put figures for it from time to time, but the figures rested on nothing
and were mere bald assertions. The second quantity is the amount
of new gold annually available for coinage in the United States. How
much is this? Nobody knows, because if an attempt is made to
define what is meant it is found that there is no idea in the words.
The people of the United States buy and coin just as much gold as
they want at any time. Hence two things are said to be unequal to
each other, when nobody knows how big either one of them is. It
may be added that it makes no difference how big either one of
them is. How much additional tin is needed annually for the increase
of our population? Do the mines produce it? Nobody knows or asks.
The mines produce, and the people buy, what they want. The case is
the same as to gold.
We find, then, that Mr. St. John begins with a doctrine which is
untenable; then he asserts a relation between population and the
need of money which does not exist; then he assumes that this need
is greater than the amount of new gold produced, although neither
he nor anybody else knows how big either one of these quantities is.
This is the argumentation by which he aims to show that prices are
reduced and misery produced by the single gold standard. It is the
argumentation which is current among the silver people. Not a step
of it will bear examination. The inference that we must restore the
free coinage of silver, to escape this strangulation of prosperity, falls
to the ground.
CAUSE AND CURE OF HARD
TIMES
36
I T is an essential part of the case of the silver men that the

country is having “hard times.” The bolters from the
Republican convention say, in their manifesto: “Discontent and
distress prevail to an extent never before known in the history of the
country.” This is an historical assertion. It is distinctly untrue. There
is no such discontent and distress as there was in 1819, or in 1840,
or in 1875, to say nothing of other periods. The writers did not know
the facts of the history, and they made use of what is nowadays a
mere figure of speech. People who want to say that a social
phenomenon is big, and who do not know what has been before,
say that it is unparalleled in history.
There has been an advancing paralysis of enterprise and arrest

of credit ever since the Sherman act of 1890 was passed. The
bolters say that “No reason can be found for such an unhappy
condition of things save in a vicious monetary system.” The reason
for it has been that the cumulative effect of the silver legislation was
steadily advancing to a crisis. The efforts by which the effects of that
legislation had been put off were no longer effective, and it was
evident that the country was on the verge of a cataclysm in which
the standard of value would be changed. What man can fail to see
the effect of such a fear on credit and enterprise? And with such a
fear in the market, how idle it is to try to represent the trouble as
caused by the fact that the existing standard was of gold, or of
silver, or of anything else! Men will make contracts and go on with
business by the use of any medium, the terms of which can be
defined, understood, and maintained until the contract is solved, but
uncertainty as to the terms, or danger of change in them, makes
credit and enterprise impossible. In the whole history of finance no
crisis can be found which was so utterly unnecessary, and so
distinctly caused by the measures of policy which had gone before it,
as that of 1893.
So much being admitted as to “hard times,” it remains true,
however, that by far the greatest part of the current declamation
about hard times is false. Prosperity and adversity of society are not
capable of exact verification. At all times some people, classes,
industries, are less prosperous than others. The fashion has grown
up among politicians and stump orators of using assertions about
prosperity and distress as arguments for their purpose, and parties
come before the public with prosperity policies. They have programs
for “making the country prosperous.” If this country, with its
population, its resources, and its chances, is not prosperous by the
intelligence, industry, and thrift of its population, does any sane man
suppose that politicians and stump orators have any devices at their
control for making it so? The orators of the present day see
prosperity where they need to see it for the purposes of their
argument. They say that all gold-standard countries in Europe are in
distress. Mr. St. John says that Mexico is prosperous. As to Canada,
we have seen no statement. According to some discussions which
are current, the bicycle rivals the gold standard as a calamity-
producer. As the bicycle has certainly gravely affected the
distribution of expenditure and the accumulation of capital, its
efficiency as a crisis-maker, in its degree, whatever that may be, can
be rationally discerned, but nobody has ever been able to show any
rational grounds of belief that the gold standard is a crisis-maker.
A crisis will also be produced whenever capital has been invested
on a large scale in any unproductive investment, whereby it is not
reproduced, but is lost. The enterprises are always made the basis of
engagements and contracts. When the enterprises fail, the
engagements cannot be met; other engagements based on these
also fail, and so on through the whole industrial organization. Such
crises are inevitable in a new country. Enterprises run in fashions. At
any one time great groups of producers tend to one line of industry.
That industry is sure to be overdone and to come to a crisis. In a
free country, where every man is at liberty to direct his enterprise as
he sees fit, what is the sense, when it turns out that he has made a
mistake, of trying to throw the losses on other people? No one
would propose it as to an individual or a number, but when there is a
great interest it makes itself a political power and produces a
platform for the same purpose, generally with inflated principles of
humanity, justice, democracy, and Americanism as wind-attachments
to make it float.
Mr. St. John says that the farmers are spending ten dollars an
acre to get eight or nine dollars an acre. What farmer in the United
States can tell how many dollars he spends on an acre? What is the
sense of these pretendedly accurate figures? But, if they had sense,
what would be the gain of cutting the dollars in two? If the farmer
spent twenty silver dollars on an acre and got back sixteen or
eighteen, how would he be benefited? The dollars of outlay are of
the same kind as the dollars of return in any case. If it is true that
the return does not equal the outlay, it must be on account of some
facts of production, and it requires but a moment’s reflection to see
that changing the currency in which outlay and income are reckoned
cannot change the relation between the two.
A dispassionate view of facts will go to prove that the world is
reasonably and ordinarily prosperous at the present time, except
where particular classes and industries are affected by special
circumstances, as some classes and industries are being affected at
all times. The land-owners of western Europe are in distress on
account of the competition of new land, with cheapened means of
transportation, but now we are told that the holders of the other
side of the competition, the land-owners of the new soil, are victims
of distress. It must be, then, that too much labor and capital are
being expended on the soil the world over, and that, too, in spite of
all the protective tariffs drawing people to the textile and metal
industries. Our silver men say that this is not the correct inference.
They say that the people on the new land suffer because the prices
are set in coins of gold and the debits and credits are kept in terms
of those coins. The prices are fixed in the world’s market in gold.
They will be so fixed, whatever we may do with our coinage laws. If
the proceeds, in being brought home, are converted into silver
value, a new opportunity for brokerage and exchange gambling will
be given to the hated bankers and brokers of Wall Street. That is the
only difference which will be produced. It would be far more sensible
to say that distress is produced by doing the business on the English
system of weights and measures, in bushels and pecks, and that
prosperity would be produced by doing it on the metric system, in
litres and hectolitres, for that charge would at least be harmless. Our
distress could all be dispelled in a week by an act of Congress
making all contracts, beyond political peradventure, that which they
are in law and fact, gold contracts.
There is, however, another cause of hard times for some people
which is far more important in our present case than any other. That
is the case of the boom which has collapsed. We hear a great deal
about “Wall Street gambling.” The gambling in Wall Street is
insignificant compared with the gambling in land, buildings, town
sites, and crops which goes on all over the country, and which is
participated in chiefly by the men who declaim about Wall Street. For
three hundred years our history has been marked by the alternations
of “prosperity” and “distress” which are produced by the booms and
their collapses. When the collapse comes the people who are left
long of goods and land always make a great outcry and start a
political agitation. Their favorite device always is to try to inflate the
currency and raise prices again until they can unload.
It is a very popular thing to tell men that they have a grievance.
That most of them find it hard to earn as much money as they need
to spend goes without saying. Now comes the wily orator and tells
them that this is somebody’s fault. In old times, if a man was sick, it
was always assumed that somebody had bewitched him. The witch
was to be sought. The medicine-man had to name somebody, and
then woe to the one who was named. Our medicine-men say that it
is the gold-bugs, Wall Street, England, who are to blame for hard
times. Whether there is any rational proof of connection is as
immaterial as it always was in witchcraft. It is a case of pain and
passion. The “gold standard” has done it! There is something to hate
and denounce. All would be well if silver could be coined at four
hundred and twelve and a half grains to the dollar. But the
assumption is that while the farmers would sell their products for
twice as many “dollars” as now, in silver, all the prices of things
which they want to buy would remain at the same number of dollars
and cents as now, in gold; that is, it is believed that wheat would be
at, say, one dollar and fifty cents per bushel in silver, instead of
seventy-five cents in gold, but that cloth would remain at fifty cents
a yard in silver, if it is now fifty cents a yard in gold. When this
assumption is brought out into clear words, every one knows that
such can never be the result. The proposed cure is like a witch cure.
It lacks rational basis, and cannot command the confidence of men
of sense. If the times were ever so bad, such a cure could only make
them worse.
THE FREE-COINAGE SCHEME IS
IMPRACTICABLE AT EVERY
POINT
37
The Program.
I N two former articles I have discussed some points which are

presented by the advocates of the free coinage of silver, on
the assumption that their project was feasible and their conception
of its operation correct. They have laid out a program; free coinage,
silver standard, great demand for silver, rise of prices, rise in the
value of silver, cancellation of debts, prosperity. They now admit that
this program would involve a panic, but it would come out, they say,
at the desired result in two or three years. They denounce the gold
standard as having caused hard times, but they plan a program with
a panic as an incident on the way to a silver standard as if it was a
trifle.
There is not a step in this program which could or would be

carried out as planned.
Free Silver Means Fiat Paper Money.

The amount of circulating cash of all kinds in the hands of the
people at the present time is about nine hundred millions. If the
dollar was reduced to half its present value, and if allowance was
made for reserves, two thousand million silver dollars would be the
specie requirement of the country. We already have nearly five
hundred millions of such dollars. Hence the country could not use at
the utmost, if the new silver dollar was worth not more than half the
present gold dollar, and if the total circulation consisted of silver
without any paper, but three times as many more silver dollars as we
have now. But every one knows that such a state of the currency
never would exist. We should have paper “based on silver”; that is to
say, the silver inflation never will be carried out. It will turn to paper
inflation at the first step. Who can believe that, if the silver standard
was adopted, silver would be bought and piled up dollar for dollar
against the paper, and that the paper would be issued only as fast as
the silver could be coined? In fact, silver would no doubt be dropped
and forgotten, and we should have plain and straightforward fiat
money of paper. Such ought to be faced as the only real sense and
probable outcome of the present agitation for the free coinage of
silver.
Limit of the Amount of Silver which could

be Absorbed.
Let us, however, proceed upon the assumption that the plan
proposed is sincere, and that the attempt would be made to carry it
out in good faith. The circulation in the hands of the people would
be paper, for they would become sick of silver and revolt against it.
There would then be two thousand million dollars in paper afloat,
each “dollar” being of silver and worth half a present gold one. We
have now five hundred million silver dollars. At the utmost not more
than another five hundred millions of silver could be absorbed into
the system. That would give reserves of fifty per cent of the total
currency, and that is the maximum of the demand for silver which
could be created if the United States went over to the silver
standard. The supply would come from all over the earth. Mr. St.
John is sure that none would come from Europe, because legal
tender silver there is at a higher ratio than sixteen to one. Not a
nation in Europe which is now under the yoke of silver would
hesitate a moment to demonetize it and send it here if we opened
our mints to it at sixteen to one. He also assures us that none would
come here from the East because the course of silver has always
been from West to East. The course of silver has turned from East to
West more than once when there was a profit on bringing it back,
and that is the only condition necessary to bring it back again. Japan
would adopt a gold currency the moment that the United States
adopted a silver one.
It is Impossible Indefinitely to Increase the

Circulation.
The power of our currency to absorb silver is not unlimited.
People seem to believe that they can go on and increase the
monetary circulation indefinitely. This is possible with paper, which
has no commodity value and cannot be exported, always
understanding that the paper will depreciate as issued, but it is not
possible with any money which has commodity value. When silver
has been put into circulation here to such an amount that all the
fictitious value given to it by the coinage law has been eliminated—
that is to say, when so many silver dollars, or paper bearing the
obligation of silver dollars, have been issued as will equal in value
the present circulation—then there will be no profit in sending silver
here from elsewhere, and no more profit in minting silver here than
in sending it elsewhere. As we have seen, there is no reason to
estimate the amount of silver which would be absorbed in this
operation at more than five hundred millions. The miners are making
all this agitation for the sake of that share which they could get in
furnishing this sum. That share would really not exceed the silver
they had on hand when the law was put in force.
Antagonistic Interests of Miners and

Populists.
What share, then, would the silver-miners get in the results of
the enterprise? They could get none unless the new silver was
bought only of them, and only bought gradually as they produced it,
and bought at a rising price as the demand of debtors acted upon it.
Not one of these conditions would be fulfilled. The debtors and the
silver-miners really have antagonistic interests at every point. It has
been proposed that only American silver should be accepted at the
mint. That plan is impracticable in any case, but, when the Populists
had their victory in hand, does anybody suppose that they would
wait eight or ten years for the realization of their hopes while the
mines were producing new silver, being certain that that delay would
cause all they hoped for to slip through their fingers? I repeat: The
interests of the two factions are all antagonistic to each other, and
one of them is destined inevitably to be the dupe of the other. That
destiny is reserved for the miners who, besides, are paying all the
expenses.
Already, so far as the campaign has proceeded, this antagonism
has begun to manifest itself. Mr. Bryan says that his plan will make
silver worth one dollar and twenty-nine cents per ounce fine. He
thus takes his position with the miners’ faction. Thereupon the
organs of the repudiators’ faction have begun to remonstrate. That
is not at all what they are fighting for. They do not want their
scheme to raise silver at all. But if it does not, the miners gain
nothing. If it does, then again the repudiators take to paper money
and the miners win nothing.
The mechanical difficulty of recoining the silver with the
necessary rapidity could probably be overcome. There are machine-
shops enough to do it if there was a party in power which had that
reckless determination to execute its will which these people show.
We may, therefore, go on to consider the rise of prices.
The Rise of Prices.

The rise in prices would regularly occur only as the new silver or
paper was put out, but as the consequences would all be discounted
it would be sudden and rapid. It would not, however, affect all things
at the same time or to an equal degree. It is here that one of the
first disappointments would occur. It is not possible to put up prices
when and as one would like to do it, even when the rise is due to
inflation. The effect cannot all be distributed at once. An advance in
price reacts on business relations, that is, on the industrial
organization. Many people and many interests find that they cannot
push against others until long after they have been pushed against
themselves. The wages class and the farmers are the ones who are
most clearly in this position, at least as far as the latter do not
produce articles for export. It must be plain that in such a convulsion
of the market everybody will try to save himself at the expense of
others. Who will succeed? Those certainly who spend their lives in
the market and already possess the control of its machinery; not
those whose time is occupied in the details of production.
Where the Expected Gains would Go.

It is said that the farmer would sell his grain and cotton, as now,
for gold; that he would exchange the gold for silver; would get the
silver coined and would pay his debts with it. Would any individual
farmer do this? Would any one man go through the steps of this
operation?—see the buyer of his products, handle the gold and
silver, go to the mint? Certainly not. All these operations would go on
through the commercial and financial machinery. They would be
executed by different individuals, in the way of business, through the
organization, and every one of them would be lost to view. Every
operation would have to be paid for. Every operation would give a
new chance for more middlemen and more charges. Would, then,
the gains of this grand scheme go to the farmer? Not at all. They
would go to the “brokers and speculators of Wall Street.” They would
be lost in commissions and charges. The type of operator whom the
Populist seems to think of when he talks about “Wall Street sharks,”
exists, although his importance in Wall Street is not as great as that
of the political farmer in agriculture; but this type of man does not
care what the currency legislation is, except that he would like to
have a great deal of it, and to have it very mixed. Whatever it is,
when it is made and he sees what it is, he will proceed to operate
upon it.
Playing into the Hands of the Money

Sharks.
We hear fierce denunciations of what is called the “money
power.” It is spoken of as mighty, demoniacal, dangerous, and
schemes are proposed for mastering it which are futile and
ridiculous, if it is what it is said to be. Every one of these schemes
only opens chances for money-jobbers and financial wreckers to
operate upon brokerages and differences while making legitimate
finance hazardous and expensive, thereby adding to the cost of
commercial operations. The parasites on the industrial system
flourish whenever the system is complicated. Confusion, disorder,
irregularity, uncertainty are the conditions of their growth. The
surest means to kill them is to make the currency absolutely simple
and absolutely sound. Is it not childish for simple, honest people to
set up a currency system which is full of subtleties and mysteries,
and then to suppose that they, and not the men of craft and guile,
will get the profits of it?
THE DELUSION OF THE
DEBTORS
38
F IFTY years ago a political agitation was started for the

annexation of Texas. As the enterprise appeared like a
barefaced piece of land-grabbing, it was necessary to invent some
historical, political, and moral theories which would give it another
color. One such theory was that Texas had properly belonged to us,
but that it was given away by Monroe and Adams in 1819. Therefore
the project was presented as one for the re-annexation of Texas.
The Re-monetization of Silver.

An attempt is now made to impugn the coinage act of 1873
under various points of view, in order to lay a foundation for the
claim that it is only sought now to re-monetize silver. Not a single
imputation on the act of 1873 has ever been presented which will
stand examination, but, if that were not so, that act was like any
other act of Congress which has become the law of the land, and
under which we have all been obliged to live for twenty-five years.
We cannot go back and undo the law and live the twenty-five years
over again. All the mistakes and follies of the past are gone into the
past for all classes and all persons amongst us. The men of the past
must be assumed to have acted according to their light, and we who
inherit the consequences of what they did must make the best of
both the good and ill of it, as the case may be, or as we think it is. If
now we make a new coinage law it must stand on its own merits,

Regression Analysis A Practical Introduction 1st Edition Jeremy Arkes Download PDF

Uploaded by

Copyright:

Available Formats

Regression Analysis A Practical Introduction 1st Edition Jeremy Arkes Download PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Regression Analysis A Practical Introduction 1st Edition Jeremy Arkes Download PDF

Uploaded by

Copyright:

Available Formats

Download the full version of the textbook now at textbookfull.

Regression Analysis A Practical Introduction

Explore and download more textbook at https://textbookfull.com

Discourse Analysis: A Practical Introduction 1st Edition

Discourse Analysis: A Practical Introduction 1st Edition

Introduction to Mediation Moderation and Conditional

Managing Risk in Nanotechnology Topics in Governance

Heal Your Ancestors to Heal Your Life 1st Edition Shelley

Communication: Innovation & Quality Miguel Túñez-López

Fitness for Work: The Medical Aspects 6th Edition John

Holographic Entanglement Entropy 1st Edition Mukund

Visit the eResources: www.routledge.com/9781138541405

2 Regression analysis basics 13

2.13 Causal effects can change over time 40

3 Essential tools for regression analysis 47

4 What does “holding other factors constant” mean? 61

5 Standard errors, hypothesis tests, p-values, and aliens 82

6 What could go wrong when estimating causal effects? 111

7 Strategies for other regression objectives 169

8 Methods to address biases 183

9 Other methods besides Ordinary Least Squares 231

10 Time-series models 252

11 Some really interesting research 281

12 How to conduct a research project 293

13 Summarizing thoughts 305

Appendix of background statistical tools 315

2.1 Scatterplot of years-of-schooling and income 15

6.8 Demonstration of mediating factors 141

2SLS Two-stage Least Squares

1.1 The problem

I’m on a mission, and I need your help.

1.1 The problem

• Amgen, a biotech company, could replicate only 6 of 53 “landmark” cancer-research studies;

1.2 The purpose of research

Human activity is contributing to global warming.

• How the study contrasts with prior research on the issue;

1.3 What causes problems in the research process?

I THINK it must be now nearly twenty years since I have made

When I was invited here this evening I thought I would try to

I T is an essential part of the case of the silver men that the

There has been an advancing paralysis of enterprise and arrest

I N two former articles I have discussed some points which are

There is not a step in this program which could or would be

Free Silver Means Fiat Paper Money.

Limit of the Amount of Silver which could

It is Impossible Indefinitely to Increase the

Antagonistic Interests of Miners and

The Rise of Prices.

Where the Expected Gains would Go.

Playing into the Hands of the Money

F IFTY years ago a political agitation was started for the

The Re-monetization of Silver.

You might also like

2 Regression analysis basics 13

2.13 Causal effects can change over time 40

3 Essential tools for regression analysis 47

4 What does “holding other factors constant” mean? 61

5 Standard errors, hypothesis tests, p-values, and aliens 82

6 What could go wrong when estimating causal effects? 111

7 Strategies for other regression objectives 169

8 Methods to address biases 183

9 Other methods besides Ordinary Least Squares 231

10 Time-series models 252

11 Some really interesting research 281

12 How to conduct a research project 293

13 Summarizing thoughts 305

Appendix of background statistical tools 315