Reproducibility in Biomedical Research

Epistemological and Statistical Problems and the Future

2nd Edition - April 26, 2024
Imprint: Academic Press
Author: Erwin B. Montgomery Jr.
Language: English
Paperback ISBN:
9 7 8 - 0 - 4 4 3 - 1 3 8 2 9 - 4
eBook ISBN:
9 7 8 - 0 - 4 4 3 - 1 3 8 3 0 - 0

Reproducibility in Biomedical Research: Epistemological and Statistical Problems, Second Edition explores the ideas and conundrums inherent in scientific research. This second ed… Read more

Purchase options

BLOOM INTO SPRING SAVINGS

Save up to 25% on books and eBooks!

Shop now

Institutional subscription on ScienceDirect

Request a sales quote

Resources

Textbook support for instructors(opens in new tab/window)

Reproducibility in Biomedical Research: Epistemological and Statistical Problems, Second Edition explores the ideas and conundrums inherent in scientific research. This second edition addresses new challenges to reproducibility in biosciences, namely reproducibility of machine learning Artificial Intelligence (AI), reproducibility of translation from research to medical care, and the fundamental challenges to reproducibility. All current chapters are expanded to cover advances in the topics previously addressed. This book provides biomedical researchers with a framework to better understand the reproducibility challenges in the area. Newly introduced interactive exercises and updated case studies help students understand the fundamental concepts involved in the area.

Cover image
Title page
Table of Contents
Copyright
Dedication
Quotes
Preface to the second edition
Looking just over the horizon
Machine learning artificial intelligence
Translational research
Biological realism and Chaos and Complexity
Probability and statistical epistemology
Randomness as fundamental and foundational
Finally…
Preface to the first edition
Chapter 1. Introduction
Abstract
Productive irreproducibility
The multifaceted notion of reproducibility and irreproducibility
Turning from the past with an eye to the future
The fundamental causes of unproductive irreproducibility
Proceeding from what is certain but not useful to what is uncertain but useful
Precision versus accuracy
Dynamics
Machine learning artificial intelligence and the emperor’s new clothes
Knowledge is prediction, prediction is reproducibility or productive irreproducibility
Challenges to prediction and thus biomedical research
When traditional experimental design and statistics breed unproductive irreproducibility
Data do not and cannot speak for themselves
Reductionism and the fundamental problem
Summary
Chapter 2. The problem of irreproducibility
Abstract
Getting a handle on the scope of unproductive irreproducibility
The inescapable risk of irreproducibility
Institutional responses
Who speaks for reproducibility and irreproducibility?
Fundamental limits to reproducibility as traditionally defined
Variability, central tendency, Chaos, and Complexity
Summary
Chapter 3. Validity of biomedical science, reproducibility, and irreproducibility
Abstract
Science must be doing something right and therein lies reproducibility and productive irreproducibility
Legacy of injudicious use of scientific logical fallacies
Science versus human knowledge of it
The necessity of enabling assumptions
Special cases of irreproducible reproducibility
Science as inference to the best explanation
Summary
Chapter 4. The logic of certainty versus the logic of discovery
Abstract
Certainty, reproducibility, and logic
Deductive logic—certainty and limitations
Propositional logic
Syllogistic deduction
Centrality of syllogistic deduction and the Fallacy of Four Terms in biomedical research
Judicious use of the Fallacy of Four Terms
Partial, probability, practical, and causal syllogisms
Induction
The Duhem–Quine thesis
Summary
Chapter 5. The logic of probability and statistics
Abstract
Probability has always been central, statistics only relatively recently
Precision versus accuracy, epistemology versus ontology
The purpose of the chapter
Continuing legacy of notions of probability
The value of the logical perspective in probability and statistics
Metaphysics: ontology versus epistemology and biomedical reproducibility
Probability
Statistics
Key general assumptions whose violation risks unproductive irreproducibility
Summary
Chapter 6. Causation, process metaphor, and reductionism
Abstract
Renewed need for causation
Practical syllogism and beyond
Centrality of hypothesis to experimentation and centrality of causation to hypothesis generation
Ontological sense of cause
Reductionism and the Fallacies of Composition and Division
Other fallacies as applied to cause
Discipline in the Principles of Causational and Informational Synonymy
Process metaphor
Summary
Chapter 7. Case studies in clinical biomedical research
Abstract
Forbearance of repetition
Purpose of clinical research as the standard
Clinical importance
Establishing clinical importance
Specific features to look for in case studies
Case study—two conflicting studies of hormone use in postmenopausal women, which is irreproducible?
Why the dominance of the Women’s Health Initiative Study over the Nurses’ Health Study?
Aftermath
Summary
Chapter 8. Case studies in basic biomedical research
Abstract
Forbearance of repetition
Purpose
Setting the stage
The value of a tool from its intended use
What is basic biomedical research?
Scientific importance versus statistical significance
Reproducibility and the willingness to ignore irreproducibility
Specific features to look for in case studies
Case study—pathophysiology of parkinsonism and physiology of the basal ganglia
Summary
Chapter 9. Case studies in computational biomedical research
Abstract
Theorizing versus computational modeling with simulation
Summary
Chapter 10. Case studies in translational research
Abstract
Translational research as the ultimate goal of basic and clinical research
Contemporary perspective on translational research
Summary
Chapter 11. Case studies in machine learning artificial intelligence
Abstract
The current environment of machine learning artificial intelligence
Machine learning AI in the context of biomedical research
Example of a neural network machine learning AI
The game of warmer/colder
Difference between machine learning AI and other analytic methods
Similarities between machine learning AI and other methods, such as regression analyses
Quality assurance in multivariate regression and implications for machine learning AI
Fallacy of Four Terms
What is the purpose or goal?
To whom or what is the machine learning AI algorithm to apply?
How should one select the training set?
What is the learning methodology?
The notion of error
Only as good as the “gold standard”
Error analyses as quality control
Case study
The “psychology” of machine learning AI
Summary
Chapter 12. Chaotic and Complex systems, statistics, and far-from-equilibrium thermodynamics
Abstract
Chaos and Complexity and the game of pool
Linearization of complex nonlinear systems
The Large Number and Central Limit theorems
Incompleteness
Self-organization
Discovering Chaos and Complexity
Equilibrium and steady-state conditions
Chaos, Complexity, and the basis for statistics
Self-organization
Summary
Chapter 13. The fundamental problem
Abstract
A word at the beginning with an eye to the future
Not for the faint of heart
Implications of the epistemic choice of variety as variation
The necessary transcendental nature of fundamentals
The argument for nonidentity
Simplification as means to willfully ignore or hide
The importance of differences
Physics of the fundamental ontological problem
How to know when no two experiences are exactly alike?
Summary
Chapter 14. Epilogue
Abstract
Implementation science
De-implementation
Metacognition and metaphysics
Reclaiming philosophy
Scientism
Some suggestions—Good Manufacturing Practices
Ethical obligations as applied to biomedical research
Summary
Appendix A. An introduction to the logic of logic
Misperception of what is logic
Logic is a discipline used to help understand reality
Proceeding from what is most certain
Proceeding to what is not certain but useful and dangerous
Extension to syllogistic deduction
Where it gets more uncertain, from state-of-being linking verbs to causation
Appendix B. Introduction to the logic of probability and statistics
Purpose
The notion of “Introduction”
Probability by enumeration of past experiences
Combinatorics to avoid uniform probability distributions and assure utility
The arithmetic mean as probability calculus
Defined statistical distributions and their models
Accuracy, precision, population, mean, and variance
The shaky ground upon which traditional statistics rests
Measures of randomness as alternatives
Experimentation is a technical matter; science is altogether different and much more difficult
Appendix C. Moving away from sample-based analyses for translational research
Importance of normal distributions in the data
Mill’s Method of Differences
Possible only when enough is ignored by what appears to be reasonable presuppositions
Preserving information of the particular individual subject
Multidimensional Shannon’s entropy
Glossary
Index

Life Sciences

Physical Sciences & Engineering

Social Sciences & Humanities

Health