Modeling Human–System Interaction: Philosophical and Methodological Considerations, with Examples

By Thomas B. Sheridan

This book presents theories and models to examine how humans interact with complex automated systems, including both empirical and theoretical methods.

• Provides examples of models appropriate to the four stages of human-system interaction
• Examines in detail the philosophical underpinnings and assumptions of modeling
• Discusses how a model fits into "doing science" and the considerations in garnering evidence and arriving at beliefs for the modeled phenomena

Modeling Human-System Interaction is a reference for professionals in industry, academia and government who are researching, designing and implementing human-technology systems in transportation, communication, manufacturing, energy, and health care sectors.

• Language: English
• Publisher: Wiley
• Release date: Dec 19, 2016
• ISBN: 9781119275282

Book preview

Table of Contents

COVER

TITLE PAGE

PREFACE

INTRODUCTION

1 KNOWLEDGE

GAINING NEW KNOWLEDGE

SCIENTIFIC METHOD: WHAT IS IT?

FURTHER OBSERVATIONS ON THE SCIENTIFIC METHOD

REASONING LOGICALLY

PUBLIC (OBJECTIVE) AND PRIVATE (SUBJECTIVE) KNOWLEDGE

THE ROLE OF DOUBT IN DOING SCIENCE

EVIDENCE: ITS USE AND AVOIDANCE

METAPHYSICS AND ITS RELATION TO SCIENCE

OBJECTIVITY, ADVOCACY, AND BIAS

ANALOGY AND METAPHOR

2 WHAT IS A MODEL?

DEFINING MODEL

MODEL ATTRIBUTES: A NEW TAXONOMY

EXAMPLES OF MODELS IN TERMS OF THE ATTRIBUTES

WHY MAKE THE EFFORT TO MODEL?

ATTRIBUTE CONSIDERATIONS IN MAKING MODELS USEFUL

SOCIAL CHOICE

WHAT MODELS ARE NOT

3 IMPORTANT DISTINCTIONS IN MODELING

OBJECTIVE AND SUBJECTIVE MODELS

SIMPLE AND COMPLEX MODELS

DESCRIPTIVE AND PRESCRIPTIVE (NORMATIVE) MODELS

STATIC AND DYNAMIC MODELS

DETERMINISTIC AND PROBABILISTIC MODELS

HIERARCHY OF ABSTRACTION

SOME PHILOSOPHICAL PERSPECTIVES

4 FORMS OF REPRESENTATION

VERBAL MODELS

GRAPHS

MAPS

SCHEMATIC DIAGRAMS

LOGIC DIAGRAMS

CRISP VERSUS FUZZY LOGIC (SEE ALSO APPENDIX, SECTION MATHEMATICS OF FUZZY LOGIC)

SYMBOLIC STATEMENTS AND STATISTICAL INFERENCE (SEE ALSO APPENDIX, SECTION MATHEMATICS OF STATISTICAL INFERENCE FROM EVIDENCE)

5 ACQUIRING INFORMATION

INFORMATION COMMUNICATION (SEE ALSO APPENDIX, SECTION MATHEMATICS OF INFORMATION COMMUNICATION)

INFORMATION VALUE (SEE ALSO APPENDIX, SECTION MATHEMATICS OF INFORMATION VALUE)

LOGARITHMIC‐LIKE PSYCHOPHYSICAL SCALES

PERCEPTION PROCESS (SEE ALSO APPENDIX, SECTION MATHEMATICS OF THE BRUNSWIK/KIRLIK PERCEPTION MODEL)

ATTENTION

VISUAL SAMPLING (SEE ALSO APPENDIX, SECTION MATHEMATICS OF HOW OFTEN TO SAMPLE)

SIGNAL DETECTION (SEE ALSO APPENDIX, SECTION MATHEMATICS OF SIGNAL DETECTION)

SITUATION AWARENESS

MENTAL WORKLOAD (SEE ALSO APPENDIX, SECTION RESEARCH QUESTIONS CONCERNING MENTAL WORKLOAD)

EXPERIENCING WHAT IS VIRTUAL: NEW DEMANDS FOR HUMAN–SYSTEM MODELING (SEE ALSO APPENDIX, SECTION BEHAVIOR RESEARCH ISSUES IN VIRTUAL REALITY)

6 ANALYZING THE INFORMATION

TASK ANALYSIS

JUDGMENT CALIBRATION

VALUATION/UTILITY (SEE ALSO APPENDIX, SECTION MATHEMATICS OF HUMAN JUDGMENT OF UTILITY)

RISK AND RESILIENCE

TRUST

7 DECIDING ON ACTION

WHAT IS ACHIEVABLE

DECISION UNDER CONDITION OF CERTAINTY (SEE ALSO APPENDIX, SECTION MATHEMATICS OF DECISIONS UNDER CERTAINTY)

DECISION UNDER CONDITION OF UNCERTAINTY (SEE ALSO APPENDIX, SECTION MATHEMATICS OF DECISIONS UNDER UNCERTAINTY)

COMPETITIVE DECISIONS: GAME MODELS (SEE ALSO APPENDIX MATHEMATICS OF GAME MODELS)

ORDER OF SUBTASK EXECUTION

8 IMPLEMENTING AND EVALUATING THE ACTION

TIME TO MAKE A SELECTION

TIME TO MAKE AN ACCURATE MOVEMENT

CONTINUOUS FEEDBACK CONTROL (SEE ALSO APPENDIX, SECTION MATHEMATICS OF CONTINUOUS FEEDBACK CONTROL)

LOOKING AHEAD (PREVIEW CONTROL) (SEE ALSO APPENDIX, SECTION MATHEMATICS OF PREVIEW CONTROL)

DELAYED FEEDBACK

CONTROL BY CONTINUOUSLY UPDATING AN INTERNAL MODEL (SEE ALSO APPENDIX, SECTION STEPPING THROUGH THE KALMAN FILTER SYSTEM)

EXPECTATION OF TEAM RESPONSE TIME

HUMAN ERROR

9 HUMAN–AUTOMATION INTERACTION

HUMAN–AUTOMATION ALLOCATION

SUPERVISORY CONTROL

TRADING AND SHARING

ADAPTIVE/ADAPTABLE CONTROL

MODEL‐BASED FAILURE DETECTION

10 MENTAL MODELS

WHAT IS A MENTAL MODEL?

BACKGROUND OF RESEARCH ON MENTAL MODELS

ACT‐R

LATTICE CHARACTERIZATION OF A MENTAL MODEL

NEURONAL PACKET NETWORK AS A MODEL OF UNDERSTANDING

MODELING OF AIRCRAFT PILOT DECISION‐MAKING UNDER TIME STRESS

MUTUAL COMPATIBILITY OF MENTAL, DISPLAY, CONTROL, AND COMPUTER MODELS

11 CAN COGNITIVE ENGINEERING MODELING CONTRIBUTE TO MODELING LARGE‐SCALE SOCIO‐TECHNICAL SYSTEMS?

BASIC QUESTIONS

WHAT LARGE‐SCALE SOCIAL SYSTEMS ARE WE TALKING ABOUT?

WHAT MODELS?

POTENTIAL OF FEEDBACK CONTROL MODELING OF LARGE‐SCALE SOCIETAL SYSTEMS

THE STAMP MODEL FOR ASSESSING ERRORS IN LARGE‐SCALE SYSTEMS

PAST WORLD MODELING EFFORTS

TOWARD BROADER PARTICIPATION

APPENDIX

MATHEMATICS OF FUZZY LOGIC (CHAPTER 4, SECTION CRISP VERSUS FUZZY LOGIC)

MATHEMATICS OF STATISTICAL INFERENCE FROM EVIDENCE (CHAPTER 4, SECTION SYMBOLIC STATEMENTS AND STATISTICAL INFERENCE)

MATHEMATICS OF INFORMATION COMMUNICATION (CHAPTER 5, SECTION INFORMATION COMMUNICATION)

MATHEMATICS OF INFORMATION VALUE (CHAPTER 5, SECTION INFORMATION VALUE)

MATHEMATICS OF THE BRUNSWIK/KIRLIK PERCEPTION MODEL (CHAPTER 5, SECTION PERCEPTION PROCESS)

MATHEMATICS OF HOW OFTEN TO SAMPLE (CHAPTER 5, SECTION VISUAL SAMPLING)

MATHEMATICS OF SIGNAL DETECTION (CHAPTER 5, SECTION SIGNAL DETECTION)

RESEARCH QUESTIONS CONCERNING MENTAL WORKLOAD (CHAPTER 5, SECTION MENTAL WORKLOAD)

BEHAVIOR RESEARCH ISSUES IN VIRTUAL REALITY (CHAPTER 5, SECTION EXPERIENCING WHAT IS VIRTUAL; NEW DEMANDS FOR MODELING)

MATHEMATICS OF HUMAN JUDGMENT OF UTILITY (CHAPTER 6, SECTION VALUATION/UTILITY)

MATHEMATICS OF DECISIONS UNDER CERTAINTY (CHAPTER 7, SECTION DECISION UNDER CONDITION OF CERTAINTY)

MATHEMATICS OF DECISIONS UNDER UNCERTAINTY (CHAPTER 7, SECTION DECISION UNDER CONDITION OF UNCERTAINTY)

MATHEMATICS OF GAME MODELS (CHAPTER 7, SECTION COMPETITIVE DECISIONS: GAME MODELS)

MATHEMATICS OF CONTINUOUS FEEDBACK CONTROL (CHAPTER 8, SECTION CONTINUOUS FEEDBACK CONTROL)

MATHEMATICS OF PREVIEW CONTROL (CHAPTER 8, SECTION LOOKING AHEAD (PREVIEW CONTROL))

STEPPING THROUGH THE KALMAN FILTER SYSTEM (CHAPTER 8, SECTION CONTROL BY CONTINUOUSLY UPDATING AN INTERNAL MODEL)

REFERENCES

INDEX

END USER LICENSE AGREEMENT

List of Tables

Chapter 02

TABLE 2.1 A taxonomy of model attributes

Chapter 09

TABLE 9.1 Fitts’ list

TABLE 9.2 The original levels of automation scale

List of Illustrations

Chapter 04

FIGURE 4.1 Trends in telephone company data (hypothetical).

FIGURE 4.2 Gaussian probability density function. .

FIGURE 4.3 Hypothetical supply–demand curves.

FIGURE 4.4 Map of the United States. .

FIGURE 4.5 Rasmussen’s schematic diagram depicting levels of behavior.

FIGURE 4.6 Wickens’ (1984) model of human multiple resources (modified by author).

FIGURE 4.7 Forward chaining tree.

FIGURE 4.8 Backward chaining tree, where AND indicates necessity and OR indicates sufficiency.

FIGURE 4.9 Kanizsa square illusion.

Chapter 05

FIGURE 5.1 The complexity of communication with a person or a machine.

FIGURE 5.2 Interpretation of Brunswik lens model (after a diagram by Kirlik, 2006).

FIGURE 5.3 Wickens’ SEEV model of attention. .

FIGURE 5.4 Senders’ model: sampling matches the Nyquist criterion.

FIGURE 5.5 Properties of mental workload (effects of very low workload not shown).

FIGURE 5.6 Regions of workload accommodation. .

FIGURE 5.7 Two images of a video showing superposition of computerized truck images on actual driver view in a test drive on a country road. White objects on trees along the roadway are fiduciary markers to enable continuous geometric correspondence of the AR image to the real world.

FIGURE 5.8 Variables contributing to presence in VR.

FIGURE 5.9 Relationship of VR created by computer and telepresence resulting from high‐quality sensing and display of events at an actual remote location. The dashed line around the remote manipulator arm suggests that the remote arm can be either real or virtual, and that if the visual and/or tactile feedback are good enough, there will be no difference in the human operator’s perception (mental model, shown in the cloud) of the (real or virtual) reality.

Chapter 06

FIGURE 6.1 A hypothetical form for performing a task analysis.

FIGURE 6.2 An example of calibration for a three‐dimensional problem space.

FIGURE 6.3 Stress–strain analogy to resilience.

FIGURE 6.4 Variables affecting trust (after Lee and See, 2004).

Chapter 07

FIGURE 7.1 Example of determining the space of what is achievable within the space defined by what is aspired to and what is acceptable (in a simple two‐dimensional problem space).

FIGURE 7.2 Tulga’s task for deciding where to attend and act.

Chapter 08

FIGURE 8.1 Fitts’ index of difficulty test.

FIGURE 8.2 Classical feedback control system.

FIGURE 8.3 Ferrell (1965) results for time to make accurate positioning movements with delayed feedback.

FIGURE 8.4 Response times of nuclear plant operator teams to properly respond to a major accident alarm. For the particular mathematical function used (log normal), using specialized graph paper (logarithm of response time on y‐axis, Gaussian percentiles on x‐axis) reduces that function to a straight line. The 95th percentile mark is seen to be roughly 100 s. .

FIGURE 8.5 Reason’s taxonomy of human error.

FIGURE 8.6 Capture error.

FIGURE 8.7 The Swiss Cheese model of accident occurrence as a result of penetrating multiple barriers. After Reason (1991).

Chapter 09

FIGURE 9.1 Four stages of human operator activity.

FIGURE 9.2 Supervisory control, as originally proposed for lunar rover operations (Ferrell and Sheridan, 1967).

FIGURE 9.3 Functions of the supervisor in relation to elements of the local human‐interactive computer (Figure 9.2) and multiple remote task‐interactive computers.

FIGURE 9.4 Supervisory control in relation to degree of automation and task entropy.

FIGURE 9.5 Distinctions with and between trading and sharing control. .

FIGURE 9.6 Adaptable control (from Sheridan, 2011).

FIGURE 9.7 Model‐based failure detection.

Chapter 10

FIGURE 10.1 The ACT‐R cognitive architecture (after Byrne et al., 2008).

FIGURE 10.2 An example of Moray’s 1990 lattice model of the operation of a pump: (a) causality relations and (b) purpose relations.

FIGURE 10.3 Formation of neuronal packets in Yufik’s model of understanding.

FIGURE 10.4 Multiple model representations in teleoperation. .

Chapter 11

FIGURE 11.1 The Leveson STAMP model. .

FIGURE 11.2 An example of system dynamics. .

FIGURE 11.3 Relationships in a policy flight simulator. .

bapp

FIGURE A.1 Hypothetical fuzzy membership functions for basketball players.

FIGURE A.2 Information relationships.

FIGURE A.3 How often to sample.

FIGURE A.4 Payoff matrix for signal detection.

FIGURE A.5 Probability densities for evidence in signal detection.

FIGURE A.6 Receiver operating characteristic (ROC).

FIGURE A.7 The definition and experimental elicitation of a person’s utility function.

FIGURE A.8 Pareto frontier and utility curve intersection determine optimal choice.

FIGURE A.9 Sample payoff matrix for decisions under probabilistic contingencies.

FIGURE A.10 Dominating and nondominating strategies (at left) and prisoner’s dilemma (right).

FIGURE A.11 Dynamic programming model of preview control.

FIGURE A.12 Kalman model of control.

STEVENS INSTITUTE SERIES ON COMPLEX SYSTEMS AND ENTERPRISES

William B. Rouse, Series Editor

WILLIAM B. ROUSE

Modeling and Visualization of Complex Systems and Enterprises

ELISABETH PATE‐CORNELL, WILLIAM B. ROUSE, AND CHARLES M. VEST

Perspectives on Complex Global Challenges: Education, Energy, Healthcare, Security, and Resilience

WILLIAM B. ROUSE

Universities as Complex Enterprises: How Academia Works, Why It Works These Ways, and Where the University Enterprise Is Headed

THOMAS B. SHERIDAN

Modeling Human–System Interaction: Philosophical and Methodological Considerations, with Examples

MODELING HUMAN–SYSTEM INTERACTION

Philosophical and Methodological Considerations, with Examples

THOMAS B. SHERIDAN

logo.gif

Copyright © 2017 by John Wiley & Sons, Inc. All rights reserved

Published by John Wiley & Sons, Inc., Hoboken, New Jersey

Published simultaneously in Canada

No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per‐copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750‐8400, fax (978) 750‐4470, or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, (201) 748‐6011, fax (201) 748‐6008, or online at http://www.wiley.com/go/permissions.

Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Neither the publisher nor author shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages.

For general information on our other products and services or for technical support, please contact our Customer Care Department within the United States at (800) 762‐2974, outside the United States at (317) 572‐3993 or fax (317) 572‐4002.

Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic formats. For more information about Wiley products, visit our web site at www.wiley.com.

Library of Congress Cataloging‐in‐Publication Data:

Names: Sheridan, Thomas B., author.

Title: Modeling human‐system interaction : philosophical and methodological considerations, with examples / Thomas B. Sheridan.

Description: Hoboken, New Jersey : John Wiley & Sons, [2017] | Series: Stevens Iinstitute series on complex systems and enterprises | Includes bibliographical references and index.

Identifiers: LCCN 2016038455 (print) | LCCN 2016051718 (ebook) | ISBN 9781119275268 (cloth) | ISBN 9781119275299 (pdf) | ISBN 9781119275282 (epub)

Subjects: LCSH: Human‐computer interaction. | User‐centered system design.

Classification: LCC QA76.9.H85 S515 2017 (print) | LCC QA76.9.H85 (ebook) | DDC 004.01/9–dc23

LC record available at https://lccn.loc.gov/2016038455

Cover Image: Andrey Prokhorov/Gettyimages

PREFACE

This book has evolved from a professional lifetime of thinking about models and, more generally, thinking about thinking. I have previously written seven books over a span of 42 years, and they all have all talked about models, except for one privately published as a memoir for my family. One even dealt with the concept of God and whether God is amenable to modeling (mostly no). So what is new or different in the present book?

The book includes quite a bit of the philosophy of science and the scientific method as a precursor to discussing human–system models. Many aspects of modeling are discussed: the purpose and uses of models for doing science and thinking about the world and examples of different kinds of models in what has come to be called human–system interaction or cognitive engineering. Along with new material, the book also includes many modeling ideas previously discussed by the author. When not otherwise cited, illustrations were drawn by the author for the book or were original works under the author’s copyright or previously declared by the author to be in public domain prior to publication.

I gratefully acknowledge contributions to these ideas from many colleagues I have worked with, especially Neville Moray, who has been my friend and invaluable critic over the years, and Bill Rouse, who shepherded the book as Wiley series editor. Modeling contributions of past coauthors Russ Ferrell, Bill Verplank, Gunnar Johannsen, Toshi Inagaki, Raja Parasuraman, Chris Wickens, Peter Hancock, Joachim Meyer, and many other colleagues and former graduate students are gratefully acknowledged.

Finally, I dedicate this effort to Rachel Sheridan, my inspiration and life partner for 63 years.

INTRODUCTION

This is a book about models, scientific models, of the interaction of individual people with technical environments, which has come to be called human–system interaction or cognitive engineering. The latter term emphasizes the role of the human intelligence in perceiving, analyzing, deciding, and acting rather than the biomechanical or energetic interactions with the physical environment.

Alphonse Chapanis (1917–2002) is widely considered to be one of the founders of the field of human factors, cognitive engineering, or whatever term one wishes to use. He coauthored one of the (if not THE) first textbooks in the field (Chapanis et al., 1949). I had the pleasure of working with him on the original National Research Council Committee in our field (nowadays called Board on Human Systems Integration, originally chaired by Richard Pew). I recall that Chapanis, while a psychologist by training, repeatedly emphasized the point that our field is ultimately applied to designing technology to serve human needs; in other words it is about engineering. Models are inherent to doing engineering.

More generally, models are the summaries of ideas we hang on to in order to think, communicate to others, and refine in order to make progress in the world. They are cognitive handles. Models come in two varieties: (1) those couched in language we call connotative (metaphor, myth other linguistic forms intended to motivate a person to make his or her own interpretation of meaning based on life experience) and (2) language we call denotative (where forms of language are explicitly selected to minimize the variability of meaning across peoples and cultures). Concise and explicit verbal statements, graphs, and mathematics are examples of denotative language. There is no doubt that connotative language plays a huge role in life, but science depends on denotative expression and models couched in denotative language, so that we can agree on what we’re talking about.

The book focuses on the interaction between humans and systems in the human environment of physical things and other people. The models that are discussed are representations of events that are observable and measurable. In experiments, these necessarily include the causative factors (inputs, independent variables), the properties of the human operator (experimental subject), the assigned task, and the task environment. They also include the effects (outputs, dependent variables), the measures of human response correlated to the inputs.

Chapters 1–3 of the book are philosophical, and apply to science and scientific models quite generally, models in human–system interaction being no exception. Chapter 1 begins with a discussion of what knowledge is and what the scientific method is including the philosophical distinction between private (subjective) knowledge and public (objective) knowledge, the importance of doubt, using and avoiding evidence, objectivity and advocacy, bias, analogy, and metaphor.

Chapter 2 defines the meaning of model and offers a six‐factor taxonomy of model attributes. It poses the question of what is to be gained by modeling and the issue of social