xv
provide. We can think of the variables in these models as hypotheses for measurable
properties of persons. The statistical relationships between them and cognitively
motivated features of items appear to move us in the direction of intersubjectivity.
But an understanding that is emerging across fields as varied as linguistics, sociology, reading research, subject domain learning, and cognitive, situative, and social
psychology opens the box further still. This perspective, which I will call “sociocognitive” for short, provides insights into how people learn, act, and think. It moves
us to think more deeply about objectivity, intersubjectivity, and latent variable models in educational measurement.
This sociocognitive perspective
3
is a kind of grand elaboration of Bartlett’s
schema theory. Every situation we experience, though unique, builds around recurring patterns in language, culture, and subject matter—from elements of grammar
and vocabulary to schemas like Bartlett’s, to cultural models, and to expectations
and behaviors of recurring activities, including taking a standardized test. None of
these patterns is a fixed thing; rather they are regularities, with variation, across
many individual unique events, across myriad individuals. Similarly, through participating in such events, every individual develops resources to recognize the patterns, learn what actions are possible, and create situations around them. Every
individual’s capabilities are unique, shaped by their history of experience.
Comprehension is activating such resources to construct, largely below the level of
consciousness, what Walter Kintsch referred to (and this book references) a situation model. Everybody’s comprehension of a given text, for example, is personal
and unique. A reading passage about a dog walking down a road draws on each
reader’s history of experiences with dogs, roads, and typical situations in which
dogs might walk along roads. Persons from environments in which dogs are friendly
pets and those in which dogs are dangerous wild animals will understand the passage differently. This is an example of a reason that questions for a given text can be
easy for some people, hard for others, and incomprehensible to still others. An
example concerning the choice of questions to ask is that “who, what, where” questions can be answered by intuitive transformations of explicit propositions in a text,
while a question about “what is missing?” demands a richer situation model (this
question can be the most important aspect of comprehending a real estate contract).
It is a commonplace observation that tasks that are hard for some people can be
easy for others, depending on our life experiences. So what does a task variable, like
an IRT difficulty variable, actually mean? How can it be a property of an item? Joe
Redish, my colleague from the Physics Education Research Group at Maryland,
explains IRT as a mean field approximation. Comprehensive explanations of all the
unique resources and processes that produced all the responses from all the people
in a group to all the items in a collection are overwhelmingly complicated. Yet while
3 More technically, sociocognitive complex adaptive systems (CASs). The special issue of
Language Learning (Volume 59, Supplement 1) provides a readable overview of CASs, summarizes ways the CAS perspective has revolutionized the field of linguistics, and offers illustrative
articles on topics such as the evolution of grammatical structures, children’s language learning, and
language testing.
Foreword
provide. We can think of the variables in these models as hypotheses for measurable
properties of persons. The statistical relationships between them and cognitively
motivated features of items appear to move us in the direction of intersubjectivity.
But an understanding that is emerging across fields as varied as linguistics, sociology, reading research, subject domain learning, and cognitive, situative, and social
psychology opens the box further still. This perspective, which I will call “sociocognitive” for short, provides insights into how people learn, act, and think. It moves
us to think more deeply about objectivity, intersubjectivity, and latent variable models in educational measurement.
This sociocognitive perspective
3
is a kind of grand elaboration of Bartlett’s
schema theory. Every situation we experience, though unique, builds around recurring patterns in language, culture, and subject matter—from elements of grammar
and vocabulary to schemas like Bartlett’s, to cultural models, and to expectations
and behaviors of recurring activities, including taking a standardized test. None of
these patterns is a fixed thing; rather they are regularities, with variation, across
many individual unique events, across myriad individuals. Similarly, through participating in such events, every individual develops resources to recognize the patterns, learn what actions are possible, and create situations around them. Every
individual’s capabilities are unique, shaped by their history of experience.
Comprehension is activating such resources to construct, largely below the level of
consciousness, what Walter Kintsch referred to (and this book references) a situation model. Everybody’s comprehension of a given text, for example, is personal
and unique. A reading passage about a dog walking down a road draws on each
reader’s history of experiences with dogs, roads, and typical situations in which
dogs might walk along roads. Persons from environments in which dogs are friendly
pets and those in which dogs are dangerous wild animals will understand the passage differently. This is an example of a reason that questions for a given text can be
easy for some people, hard for others, and incomprehensible to still others. An
example concerning the choice of questions to ask is that “who, what, where” questions can be answered by intuitive transformations of explicit propositions in a text,
while a question about “what is missing?” demands a richer situation model (this
question can be the most important aspect of comprehending a real estate contract).
It is a commonplace observation that tasks that are hard for some people can be
easy for others, depending on our life experiences. So what does a task variable, like
an IRT difficulty variable, actually mean? How can it be a property of an item? Joe
Redish, my colleague from the Physics Education Research Group at Maryland,
explains IRT as a mean field approximation. Comprehensive explanations of all the
unique resources and processes that produced all the responses from all the people
in a group to all the items in a collection are overwhelmingly complicated. Yet while
3 More technically, sociocognitive complex adaptive systems (CASs). The special issue of
Language Learning (Volume 59, Supplement 1) provides a readable overview of CASs, summarizes ways the CAS perspective has revolutionized the field of linguistics, and offers illustrative
articles on topics such as the evolution of grammatical structures, children’s language learning, and
language testing.
Foreword