Most books on human-computer interaction (HCI) and usability give recommendations based on empirical research, guidelines fit to observed user behavior, and cognitive models after the fact. Peter Pirolli, the father of information foraging theory, has written a new book that models and predicts what users will do before they navigate a website. Using mathematical models of human behavior, Pirolli lays out the foundation of information foraging theory, a relatively new field based in part on optimal foraging theory in animals (Stephens & Krebs 1986). The result is a seminal work in Oxford University Press’ series on Human-Computer Interaction. We were fortunate to review a proof of Pirolli’s new book Information Foraging Theory: Adaptive Interaction with Information, due out April 2007.
The book builds up Pirolli and Card’s current thinking on Information Foraging Theory, chapter by chapter (Pirolli & Card 1999). The first eight chapters proceed from simple foraging models to more complex and realistic models of informating foraging on the Web. Each chapter uses mathematical forumlas to model human behavior with a brief discussion along with illustrative examples and graphs to represent key concepts. By necessity Pirolli has created a new field, borrowing from economic decision-making, Bayesian mathematics, and rational choice under uncertain and limited conditions. Overall the book is technical and more of a college textbook than usability handbook in the tradition of Nielsen and Spool. You may need to brush up on your advanced math skills before tackling this book. However, the last three chapters on cooperative foraging, practical guidelines, and future research and applications are well worth the price of admission.Like animals foraging for food with time and energy constraints, humans forage for information looking for answers.
The problem is that we are awash in a glut of information, which is growing at an alarming rate. Humans have naturally adopted adaptive strategies to optimize their intake of useful information per unit cost. Information foraging theory aims to model and predict our evolutionary behavior in adapting to an increasingly complex world. The book both models our behavior in mathematical formulae and validates its predictions empirically to refine the model. Armed with this knowledge, designers can make informed decisions on web designs before they make a heavy investment in implementation.
Answers the Big Questions
The book answers following questions that have vexed web developers for over a decade.
- Why we abandon one website for another
- Why a small decrease in “information scent” can exponentially increase search costs
- How breakthrough discoveries occur
The Father of Information Foraging Theory
After he joined the Palo Alto Research Center (PARC) in 1992, Pirolli became involved in problems of intelligent information access. A group of researchers including Stuard Card, Dan Russell, Mark Stefik, and John van Gigch were doing studies of analysts and MBA students doing information-intensive searches. Pirolli realized that a new approach would be needed to analyze and model these free-wheeling tasks. Building on the work of John Anderson (1990), who first studied the mind-environment link and the work of E. A. Smith (1987) whose behavioral ecology holds that people’s strategies can be predicted in part by scrutinizing their environment, Pirolli wondered if a similar approach could be applied to the analysis of the information environment and to how it shapes our seeking behavior. Researching the topic he found optimal foraging theory (Stephens and Krebs 1986) and literally had an “ah-ha” experience one night while brainstorming the idea with his fiancee, Jacqui LeBlanc. Pirolli realized that optimal foraging theory and the core mathematics behind it could be applied to how humans search for information.
In July 1992, Pirolli wrote a working paper titled “Notes on Adaptive Sense Making in Information Ecologies,” to introduce the idea that conventional foraging models could be used on idealized information foraging tasks. The paper was met with scepticism (“information is not food”) and awe (“audacious”). After encouragement from his manager and HCI pioneer Stuard Card, Pirolli presented a talk in 1993 titled “Information foraging: A new view on problems in human-computer interaction,” for Human-Computer Interaction Consortium Winter Workshop in Atlanta, Georgia. His first published article on information foraging was in 1995 titled “Information foraging in information access environments” (Pirolli & Card 1995). The work most cited is the paper “Information Foraging,” Pirolli and Card published in the prestigious Psychology Review in 1999 (Pirolli & Card 1999). His new book “Information Foraging Theory: Adaptive Interaction with Information” is the culmination of more than a decade of work on the topic of information foraging theory.
Pirolli recounts his second “ah-ha” moment in the Preface where information foraging was used to bring “messy data into crystal clear focus.” When analyzing log file data from a Scatter/Gather study detailed in Chapter 6, simple analysis of the user interaction logs showed a seemingly nonsensical allocation of their time and navigational choices. Once he applied optimal foraging methods, the data fell neatly into place (Figure 7).
What Is Information Foraging?
Unlike cognitive engineering models of the early 1990s where tasks were analyzed in well-defined domains and user interfaces, people seeking information are largely shaped by the structure of the content, that is, their information environment. People routinely sift massive volumes of information under deadline pressure to make complex search decisions under uncertain conditions. We are constantly evaluating the expected costs and benefits of our actions in this largely probabilistic textured information environment (Brunswik 1952) to maximize our gain of valuable information. Pirolli calls this behavior information foraging.
Information foraging strategies optimize the use of our time and effort to gain valuable information. We satisfice, getting a good enough outcome under uncertain and limited conditions (bounded rationality). As in the real world where animals forage for food, the online world of the Web is a patchy environment with information arranged into clumps. Patches of useful information reside in different websites, it is our job as “informavores” to seek out the richest patches and extract useful information. As we forage for more information, it becomes harder to find additional useful information as we pick off the “low hanging fruit.” Diminishing returns cause us to “feed” at a patch until our rate of gain of useful information falls below the perceived average. Once we think the grass is greener in another “patch,” we switch pages, websites, or search criteria, seeking out more fruitful patches of information. Pirolli’s mathematics describe the environment and predicts with some accuracy our behavior in the Web’s patchy environment.
The optimal information forager “maximizes the value of knowledge gained per unit cost of interaction.” We select designs that improve our returns on information foraging. This natural selection implies that sites that have high information scent and usability would fare better than sites that have lower scent and usability. This gives an entirely new meaning to the survival of the fittest when applied to the Web. Next page » information foraging theory ».
Here are the chapter titles of the book:
- Information Foraging Theory: Framework and Methods
- Elementary Foraging Models
- The Ecology of Information Foraging on the World Wide Web
- Rational Analysis of Information Scent and Web Foraging
- A Cognitive Model of Information Foraging on the Web
- A Rational Analysis and Computational Cognitive Model of the Scatter/Gather Document Cluster Browser
- Stochastic Models of Information Foraging by Information Scent
- Social Information Foraging
- Design Heuristics, Engineering Models, and Applications
- Future Directions, Upward, Downward, Inward, and Outward