Search and the Paradox of Choice

Back when most of us were still learning HTML, Peter co-authored the IA Bible–Information Architecture for the World Web. In no time at all, this best-selling book ushered in an entirely new profession—information architects or IAs—designers of complex information systems, focused on improving structure, organization, navigation, labeling, and search.

Peter then went on to write Ambient Findability: What We Find Changes Who We Become, which examines attention, authority, and wayfinding at the crossroads of ubiquitous computing and the Internet. And, Peter’s latest book is Search Patterns: Design for Discovery (co-created with Jeffery Callender). This new book delves into the arcane topics of search and search design. In the opening pages, Peter writes “search is a wicked problem of terrific consequence.”

We caught up with Peter to talk about Project Information Literacy and our latest findings about college students, especially their difficulties with finding information and search.

PIL: Recently, PIL finished conducting a large-scale survey of 8,400 college students on 25 different campuses in the U.S. We asked students how they found, evaluated, and used information for course work and solving information problems in everyday life. Here’s a finding that intrigued us: The three most frequently reported difficulties students had with course-related research were: getting started, defining a topic, and narrowing down a topic. If you notice, reading research articles, distilling what’s been found, or writing didn’t make the list—search predominates. Why is search so difficult for college students, especially the first few steps of search?

Peter: This finding is emblematic of the intimate relationship between search, learning, and decision making, and it brings to mind the paradox of choice. After all, the search box offers unrivaled selection. You can ask it any question. Or at least it often feels that way. For a student, this freedom can be simultaneously exhilarating and totally paralyzing. Also, most students lack a useful mental model of search. They don’t know how search works or what’s being searched, which may be fine for casual Googling but not for navigating dozens of research databases. Finally, selecting a topic is inherently difficult. It’s like buying a house or finding a spouse. The process is fuzzy and uncomfortable because we’re not sure what we want. So, all too often, we procrastinate. We wait until the last minute to begin, which is a shame because getting started is half the battle. The key is to recognize that search can be an iterative, interactive journey of discovery that not only helps us find what we need but also lets us learn what we want to find. When we embrace this more playful model of exploratory search, it’s not so hard to get started.

PIL: We also asked students about everyday life research—finding news, health information, employment opportunities, and such. With everyday life research, the three most frequent difficulties were—sorting through irrelevant results, trying to find an “answer” online that they knew existed, and determining the credibility of a source they have found (usually on the Web). Why is processing search results particularly problematic for students?

Peter: Search is inherently difficult. Given the ambiguity of language and the subjectivity of relevance, it’s easy to use “the wrong keywords” and hard to deliver “the right results.” How do we bridge the gap between the simple search terms of students and the sophisticated vocabularies of subject matter experts? How can we offer useful results without understanding what the searcher already knows? These intrinsic problems are magnified by the sad state of search design. Let’s face it. Many scholarly database products are slow, ugly, and hard to use. It’s no wonder students prefer Google.

Of course, students are also part of the problem. Despite being more comfortable with technology than their parents (and teachers), they often lack the skills and literacy needed for optimal learning and decision making. At a time when we can increasingly select our sources and choose our news, we’re not effectively building information literacy into education. Students rarely consider the bias or motives of authors and publishers, and all of us are guilty of seeking sources that reinforce our existing beliefs while ignoring ideas that challenge or contradict what we think we know. And, this problem isn’t just academic. Information literacy shapes our lives. I told a personal story about search, sources, authority, and trust in this excerpt from Ambient Findability. (See excerpt.)

In short, I view search (within a broad framework that includes information literacy) as one of the most important and difficult skills we should teach our children and students. It’s absolutely critical to the success of their education, their careers, and their lives.

PIL: In our work at PIL, we’ve found students see research as a competency learned by rote. A large majority of college students we’ve studied turn to the same small set of commonplace sources—course readings, Google, library databases, Wikipedia—over and over again, whether they are conducting course-related or everyday life research. With all the sources that are available to them, why do today’s college students use a search strategy that is risk-averse, non-exploratory, narrow, and that relies on repeat behaviors?

Peter: Librarians often make the same mistake as economists. We assume that people are rational. This leads to all sorts of seductively clear models that are quite simply wrong. I call this the “people problem” and it’s been addressed by lots of smart folks over the years. George Kingsley Zipf wrote about the Principle of Least Effort:

Each individual will adopt a course of action that will involve the expenditure of the probably least average of his work (by definition, least effort).

Herbert Simon added that since we often “satisfice” under conditions of “bounded rationality” this behavior is neither perfect nor perfectly predictable. And, Calvin Mooers revealed the irrationality at the very heart of our information use (or lack thereof):

An information retreival system will tend not to be used whenever it is more painful and troublesome for a customer to have information than for him not to have it. [*]

It’s worth noting that Calvin Mooers was not talking about usability. Consider his explanation of what subsequently became known as Mooers’ Law:

It is now my suggestion that many people may not want information, and that they will avoid using a system precisely because it gives them information….Having information is painful and troublesome. We have all experienced this. If you have information, you must first read it, which is not always easy. You must then try to understand it….Understanding the information may show that your work was wrong, or may show that your work was needless….Thus not having and not using information can often lead to less trouble and pain than having and using it.[‡]

Again, this is why I believe that explicit education is required to build information literacy. Left to our own devices, most people will adopt a fast food approach to information consumption. Changing this natural behavior isn’t easy. But it’s not impossible either.

PIL: Let’s talk about design. PIL has found that 9 out of 10 students turn to their campus library’s OPAC when looking for scholarly information. Many library sites use federated search (i.e., search functionality that allows for the simultaneous search across multiple databases or Web resources). Is federated search an effective search design for making the information students need findable?

Peter: Federated search is an intriguing design pattern because it can be a solution and a symptom at the same time. Here’s how we explained it in Search Patterns. (See excerpt.)

In short, I believe federated search is necessary, but we must insist on well-designed interfaces and sub-second response times. Speed is essential to iterative, interactive search. We simply can’t take slow for an answer. And I recognize that achieving these goals is not easy and will take time and passionate advocacy on behalf of our users.

PIL: What’s your answer for making students have fewer difficulties with search, Peter? What should librarians and college professors keep in mind when working with students?

Peter: I’d love to see search and information literacy embedded more deeply in education. For instance, a research paper assignment could include an initial (graded) step in which the student must identify a research topic, cite a few potential sources, and explain why this topic is important and inspiring. That could help soften the procrastination problem. Then, after submission of the paper, the professor could assign a second paper, in which the student must argue the opposite point with source citations. Or, I can simply imagine classroom discussions in which students are asked to defend or critique their sources. Why do you trust this person? Why do you believe what you read in this book?
​
But education isn’t enough. We must make search better. Librarians, in particular, must improve our own search systems and demand better interfaces and faster performance from vendors. This won’t be easy. In fact, I love search precisely because it’s so hard. It truly is a wicked problem that requires radical innovation, continuous improvement, and multi-disciplinary collaboration. It’s both a project and a process. It’s a thorny problem that’s never solved. And yet it’s a powerful force at the center of change. It’s transforming the way we learn and make decisions. So, search is one thing we better get right.

---

[*] Remarks by Calvin N. Mooers during a panel discussion at the Annual Meeting of the American Documentation Institute, October 24, 1959.[‡] Remarks by Calvin N. Mooers on October 24, 1959. Reprinted in the Bulletin of the American Society for Information Science, October/November 1996.

---

Peter Morville is president and founder of Semantic Studios, a consultancy in user experience, information architecture, and findability. Clients have included AT&T, Harvard, IBM, the Library of Congress, Microsoft, the National Cancer Institute, Vodafone, and the Weather Channel. Peter also serves on the Board of Project Information Literacy.

Peter’s books include Search Patterns: Design for Discovery (co-created with Jeffery Callender, O’Reilly, 2010), Ambient Findability: What We Find Changes Who We Become (O’Reilly, 2005), and Information Architecture for the World Wide Web (co-authored with Lou Rosenfeld, first edition: O’Reilly, 1998).

Smart Talks is an occasional series produced by Project Information Literacy (PIL). PIL hosts interviews with leading experts about PIL’s findings and their thoughts about the challenges of finding information and conducting research in the digital age. 

​Smart Talks are open-access; no permission is required from PIL for re-use, though we ask that this source is cited as: “Peter Morville: Search and the Paradox of Choice,” Project Information Literacy Smart Talks, no. 1, Peter Morville, June 1, 2010.