THE PSYCHOLOGY OF ACTIONS: MENTAL MODEL, ERRORS

<< COGNITIVE PROCESSES: LEARNING, READING, SPEAKING, LISTENING, PROBLEM SOLVING, PLANNING, REASONING, DECISION-MAKING

DESIGN PRINCIPLES: >>

Human Computer Interaction (CS408)

Lecture

Lecture 11. The Psychology of Actions

Learning Goals

As the aim of this lecture is to introduce you the study of Human Computer

Interaction, so that after studying this you will be able to:

Understand mental models

Understand psychology of actions

Discuss errors.

Mental model

11.1

The concept of mental model has manifested itself in psychology theorizing and HCI

research in a multitude of ways. It is difficult to provide a definitive description,

because different assumption and constraints are brought to bear on the different

phenomena it has been used to explain. A well-known definition, in the context of

HCI, is provided by Donald Norman: `the model people have of themselves, others,

the environment, and the things with which they interact. People form mental models

through experience, training and instruction'.

It should be noted that in fact the term mental model was first developed in the early

1640s by Kenneth Craik. He proposed that thinking `...models, or parallels reality':

`If the organism carries a "small-scale model" of external reality and of its own

possible actions within its head, it is able to try out various alternatives, conclude

which is the best of them, react to future situations before they arise, utilize the

knowledge of past events in dealing with the present and future, and in every way to

react in a much fuller, safer, and more competent manner to emergencies witch face

it.'

Just as an engineer will build scale models of a bridge, in order to test out certain

stresses prior to building the real thing, so, too, do we build mental models of the

world in order to make predictions about an external event before carrying out an

action? Although our construction and use of mental models may not be as extensive

or as complete as Craik's hypothesis suggests, it is likely that most of us can probably

recall using a form of mental simulation at some time or other. An important

observation of these types of mental models is that they are invariably incomplete,

unstable, and easily confusable and are often based on superstition rather than

scientific fact.

Within cognitive psychology the term mental model has since been explicated by

Johnson-Laird (1983, 1988) with respect to its structure and function in human

reasoning and language understanding. In terms of structure of mental models, he

argues that mental models are either analogical representations or a combination of

analogical and prepositional representations. They are distinct from, but related to

images. A mental model represents the relative position of a set of objects in an

Human Computer Interaction (CS408)

analogical manner that parallels the structure of the state of objects in the world. An

image also does this, but more specifically in terms of view of a particular model.

An important difference between images and mental models is in terms of their

function. Mental models are usually constructed when we are required to make an

inference or a prediction about a particular state of affairs. In constructing the mental

model a conscious mental simulation may be `run' from which conclusions about the

predicted state of affairs can be deduced. An image, on the other hand, is considered

to be a one-off representation. A simplified analogy is to consider an image to be like

a frame in a movie while a mental model is more like a short snippet of a movie.

So, after this discussion we can say that while learning and using a system, people

develop knowledge of how to use the system and, to a lesser extent, how the system

works. These two kinds of knowledge are often referred to as a user's mental model.

Having developed a mental model of an interactive product, it is assumed that people

will use it to make inferences about how to carry out tasks when using the interactive

product. Mental models are also used to fathom what to do when something

unexpected happens with a system and when encountering unfamiliar systems. The

more someone learns about a system and how it functions, the more their mental

model develops. For example, TV engineers have a deep mental model of how TVs

work that allows them to work out how to fix them. In contrast, an average citizen is

likely to have a reasonably good mental model of how to operate a TV but a shallow

mental model of how it worked.

To illustrate how we use mental models in our everyday reasoning, imagine the

following scenario:

· You arrive home from a holiday on a cold winter's night to a cold house. You

have small baby and you need to get the house warm as quickly as possible.

Your house is centrally heated. Do you set the thermostat as high as possible

or turn it to the desired temperature (e.g., 70F)

Most people when asked the questions imagine the scenario in terms of what they

would do in their own house they choose the first option. When asked why, a typical

explanation that is given is that setting the temperature to be as high as possible

increases the rate at which the room warms up. While many people may believe this,

it is incorrect.

There are two commonly held folk theories about thermostats: the timer theory and

the valve theory. The timer theory proposes that the thermostat simply controls the

relative proportion of time that the device stays on. Set the thermostat midway, and

the device is on about half the time; set it all the way up and the device is on all the

time; hence, to heat or cool something most quickly, set the thermostat so that the

device is on all the time. The valve theory proposes that the thermostat controls how

much heat comes out of the device. Turn the thermostat all the way up, and you get

maximum heating or cooling.

Thermostats work by switching on the heat and keeping it going at a constant speed

until the desired temperature set is reached, at which point they cut out. They cannot

control the rate at which heat is given out from a heating system. Left a given setting,

thermostats will turn the heat on an off as necessary to maintain the desired

temperature. It treats the heater, oven, and air conditioner as all-or-nothing devices

that can be either fully on or fully off, with no in-between states. The thermostat turns

the heater, oven, or air conditioner completely on--at full power--until the

temperature setting on the thermostat is reached. Then it turns the unit completely off.

Human Computer Interaction (CS408)

Setting the thermostat at one extreme cannot affect how long it takes to reach the

desired temperature.

The real point of the example is not that some people have erroneous theories; it is

that everyone forms theories (mental models) to explain what they have observed. In

the case of the thermostat the design gives absolutely no hint as to the correct answer.

In the absence of external information, people are free to let their imaginations run

free as long as the mental models they develop account for the facts as they perceive

them.

Why do people use erroneous mental models?

It seems that in the above scenario, they are running a mental model based on general

valve theory of the way something works. This assumes the underlying principle of

"more is more": the more you turn or push something, the more it causes the desired

effect. This principle holds for a range of physical devices, such as taps and radio

controls, where the more you turn them, the more water or volume is given. However,

it does not hold for thermostats, which instead function based on the principle of an

on-off switch. What seems to happen is that in everyday life people develop a core set

of abstractions about how things work, and apply these to a range of devices,

irrespective of whether they are appropriate.

Using incorrect mental models to guide behavior is surprisingly common. Just watch

people at a pedestrian crossing or waiting for an elevator (lift). How many times do

they press the button? A lot of people will press it at least twice. When asked why, a

common reason given is that they think it will make it lights change faster or ensure

the elevator arrives. This seems to do another example of following the "more is

more" philosophy: it is believed that the more times you press the button, the more

likely it is to result in he desire effect.

Another common example of an erroneous mental model is what people do when the

cursor freeze on their computer screen. Most people will bash away at all manner of

keys in the vain hope that this will make it work again. However, ask them how this

will help and their explanations are rather vague. The same is true when the TV starts

acting up: a typical response is to hit the top of the box repeatedly with a bare hand or

a rolled-up newspaper. Again, as people why and their reasoning about how this

behavior will help solve the problem is rather lacking.

Indeed, research has shown that people's mental models of the way interactive

devices work is poor, often being incomplete, easily confusable, based on

inappropriate analogies, and superstition. Not having appropriate mental models

available to guide their behavior is what caused people to become very frustrate--

often resulting is stereotypical "venting' behavior like those described above.

On the other hand, if people could develop better mental models of interactive

systems, they would be in a better position to know how to carry out their tasks

efficiently and what to do if the system started acting up. Ideally, they should be able

to develop a mental model that matches the conceptual; modal developed by the

designer. But how can you help users to accomplish this? One suggestion is to

educate them better, however, many people are resistant to spending much time

learning about how things work, especially if it involves reading manuals and other

documentation. An alternative proposal is to design systems to be more transparent,

so that they are easier to understand.

People do tend to find causes for events, and just what they assign as the cause varies.

In part people tend to assign a causal relation whenever two things occur in

Human Computer Interaction (CS408)

succession. If I do some action A just prior to some result R, then I conclude that A

must have caused R, even if, there really was no relationship between the two.

Self-blaming

Suppose I try to use an everyday thing, but I can't: where is the fault, in my action or

in the thing? We are apt to blame ourselves. If we believe that others are able to use

the device and if we believe that it is not very complex, then we conclude that any

difficulties must be our own fault. Suppose the fault really lies in the device, so that

lots of people have the same problems. Because everyone perceives the fault to be his

or own, nobody wants to admit to having trouble. This creates a conspiracy of silence,

maintaining the feeling of guilt and helplessness among users.

Interestingly enough, the common tendency to blame ourselves for failures with

everyday objects goes against the normal attributions people make. In general, it has

been found that normal attribute their own problems to the environment, those of

other people to their personalities.

It seems natural for people to blame their own misfortunes on the environment. It

seems equally natural to blame other people's misfortunes on their personalities. Just

the opposite attribution, by the way, is made when things go well. When things go

right, people credit their own forceful personalities and intelligence. The onlookers do

the reverse. When they see things go well for someone else, they credit the

environment.

In all cases, whether a person is inappropriately accepting blame for the inability to

work simple objects or attributing behavior to environment or personality, a faulty

mental model is at work.

Reason for self-blaming

Learned helplessness

The phenomenon called learned helplessness might help explain the self-blame. It

refers to the situation in which people experience failure at a task, often numerous

times. As a result, they decide that the task cannot be done, at least not by them: they

are helpless. They stop trying. If this feeling covers a group of tasks, the result can be

severe difficulties coping with life. In the extreme case, such learned helplessness

leads to depression and to a belief that the person cannot cope with everyday life at

all. Some times all that it takes to get such a feeling of helplessness is a few

experiences that accidentally turn out bad. The phenomenon has been most frequently

studied as a precursor to the clinical problem of depression, but it might easily arise

with a few bad experiences with everyday life.

Taught helplessness

Do the common technology and mathematics phobias results from a kind of learned

helplessness? Could a few instances of failure in what appear to be straightforward

situations generalize to every technological object, every mathematics problem?

Perhaps. In fact, the design of everyday things seems almost guaranteed to cause this.

We could call this phenomenon taught helplessness.

With badly designed objects--constructed so as to lead to misunderstanding--faulty

mental models, and poor feedback, no wonder people feel guilty when they have

trouble using objects, especially when they perceive that nobody else is having the

Human Computer Interaction (CS408)

same problems. The problem is that once failure starts, it soon generalizes by self-

blame to all technology. The vicious cycle starts: if you fail at something, you think it

is your fault. Therefore you think you can't do that task. As a result, next time you

have to do the task, you believe you can't so you don't even try. The result is that you

can't, just as you thought. You are trapped in a self-fulfilling prophecy.

The nature of human thought and explanation

It isn't always easy to tell just where the blame for problem should be placed. A

number of dramatic accidents have come about, in part, from the false assessment of

blame in a situation. Highly skilled, well-trained people are using complex equipment

when suddenly something goes wrong. They have to figure out what the problem is.

Most industrial equipment is pretty reliable. When the instruments indicate that

something is wrong, one has to consider the possibility that the instruments

themselves are wrong. Often this is the correct assessment. When operators

mistakenly blame the instruments for an actual equipment failure, the situation is ripe

for a major accident.

It is spectacularly easy to find examples of false assessment in industrial accidents.

Analysts come in well after the fact, knowing what actually did happen; with

hindsight, it is almost impossible to understand how the people involved could have

made the mistake. But from the point of view of the person making decisions at time,

the sequence of events is quite natural.

Three Mile Island Nuclear Power Plant

At the Three Mile Island nuclear power plant, operators pushed a button to close a

valve; the valve had been opened (properly) to allow excess water to escape from the

nuclear core. In fact, the valve was deficient, so it didn't close. But a light on the

control panel indicated that the valve position was closed. The light actually didn't

monitor the valve, only the electrical signal to the valve, a fact known by the

operators. Still, why suspect a problem? The operators did look at the temperature in

the pipe leading from the valve: it was high, indicating that fluid was still flowing

through the closed valve. Ah, but the operators knew that the valve had been leaky, so

the leak would explain the high temperature; but the leak was known to be small, and

operators assumed that it wouldn't affect the main operation. They were wrong, and

the water that was able to escape from the core added significantly to the problems of

that nuclear disaster. Norman says that the operators' assessment was perfectly

reasonable: the fault wan is the design of the lights and in the equipment that gave

false evidence of a closed valve.

Lockheed L-1011

Similarly many airline accidents happened just due to misinterpretations. Consider

flight crew of the Lockheed L-1011 flying from Miami, Florida, to Nassau, Bahamas.

The plane was over the Atlantic Ocean, about 110 miles from Miami, when the low

oil pressure light for one of the three engines went on. The crew turned off the engine

and turned around to go back to Miami. Eight minutes later, the low-pressure lights

for the remaining two engines also went on, and the instruments showed zero oil

pressure and quantity in all three engines. What did the crew do now? They didn't

believe it! After all, the pilot correctly said later, the likelihood of simultaneous oil

exhaustion in all three engines was "one in millions I would think." At the time,

sitting in the airplane, simultaneous failure did seem most unlikely. Even the National

Human Computer Interaction (CS408)

Transportation Safety Board declared, "The analysis of the situation by the flight crew

was logical, and was what most pilots probably would have done if confronted by the

same situation."

What happened? The second and third engines were indeed out of oil, and they failed.

So there were no operating engines: one had been turned off when its gauge registered

low, the other two had failed. The pilots prepared the plane for an emergency landing

on the water. The pilots were too busy to instruct the flight crew properly, so the

passengers were not prepared. There was semi-hysteria in the passenger cabin. At the

last minute, just as the plane was about to ditch in the ocean, the pilots managed to

restart the first engine and land safely to Miami. Then that engine failed at the end of

the runway.

Why did all three engine fail? Three missing O-rings, one missing from each of three

oil plugs, allowed all the oil to seep out. The O-rings were put in by two different

people who worked on the three engines (one for the two plugs on the wings, the

other of the plug on the tail). How did both workers make the same mistake? Because

the normal method by which they got the oil plugs had been changed that day. The

whole tale is very instructive, for there were four major failures of different sorts,

from the omission of the O-rings, to the inadequacy of the maintenance procedures, to

the false assessment of the problem, to the poor handling of the passengers.

Fortunately nobody was injured. The analysts of the National Transportation Safety

Board got to write a fascinating report.

Find an explanation, and we are happy. But our explanations are based on analogy

with past experience, experience that may not apply in the current situation. In the

Three Mile Island incident, past experience with the leaky valve explained away the

discrepant temperature reading; on the flight from Miami to Nassau, the pilots' lack of

experience with simultaneous oil pressure failure triggered their belief that the

instruments must be faulty. Once we have an explanation--correct or incorrect--for

otherwise discrepant or puzzling events, there is no more puzzle, no more

discrepancy. As a result, we are complacent, at least for a while.

How people do things

To get something done, you have to start with some notion of what is wanted--the

goal that is to be achieved. Then, you have to do some thing to the world , that is, take

action to move yourself or manipulate someone or something. Finally, you check to

see that your goal was made. So there are four different things to consider: the goal,

what is done to the world, the world itself, and the check of the world. The action

itself has two major aspects: doing something and checking. Call these execution and

evaluation Goals do not state precisely what to do--where and how to move, what to

pick up. To lead to actions goals must be transformed into specific statements of what

is to be done, statements that are called intentions. A goal is some thing to be

achieved, often vaguely stated. An intention is specific action taken to get to the goal.

Yet even intentions are not specific enough to control actions.

Suppose I am sitting in my armchair, reading a book. It is dust, and the light has

gotten dimmer and dimmer. I decide to need more light (that is the goal: get more

light). My goal has to be translated into the intention that states the appropriate action

in the world: push the switch button on the lamp. There's more: I need to specify how

to move my body, how to stretch to reach the light switch, how to extend my finger to

push the button (without knocking over the lamp). The goal has to be translated into

an intention, which in turn has to make into a specific action sequence, one that can

Human Computer Interaction (CS408)

control my muscles. Note that I could satisfy my goal with other action sequences,

other intentions. If some one walked into the room and passed by the lamp, I might

alter my intention form pushing the switch button to asking the other person to do it

for me. The goal hasn't changed, but the intention and resulting action sequence have.

Action Cycle

Human action has two aspects, execution

Goals

and evaluation. Execution involves doing

What we want to

something. Evaluation is the comparison of

happen

what happened in the world with what we

wanted to happen

Evaluation

Execution to

Stages of Execution

What we do

Comparing what

the world

Happened with what we

Start at the top with the goal, the state that is

wanted to happen

to be achieved. The goal is translated into an

intention to do some action. The intention

must be translated into a set of internal

commands, an action sequence that can be

THE

WORLD

performed to satisfy the intention. The

action sequence is still a mental event:

noting happens until it is executed, performed upon the world.

Stages of Evaluation

Evaluation starts with our perception of the world. This perception must then be

interpreted according to our expectations and then compared with respect to both our

intentions and our goals

Seven stages of action

Goals

Intention to act

Evaluation of the

Interpretations

The stages of execution (intentions, action

sequence, and execution) are coupled with

the stages of evaluation (perception,

sequence of

Interpreting the

interpretation, and evaluation), with goals

actions

perception

common to both stages.

Errors

11.2

execution of

Perceiving the state

The action sequence

of the world

Human capability for interpreting and

manipulating

information

quite

impressive. However, we do make mistake.

Whenever we try to learn a new skill, be it

skiing, typing, cooking or playing chess, we

are bound to make mistakes. Some are

THE WORLD

trivial, resulting in no more than temporary

inconvenience or annoyance. Other may be more serious, requiring substantial effort

to correct. In most situations it is not such a bad thing because the feedback from

making errors can help us to learn and understand an activity. When learning to use a

computer system, however, learners are often frightened of making errors because, as

100

Human Computer Interaction (CS408)

well as making them feel stupid, they think it can result in catastrophe. Hence, the

anticipation of making an error and its consequences can hinder a user's interaction

with a system.

Why do we make mistakes and can we avoid them? In order to answer the latter part

of the question we must first look at what is going on when we make an error. There

are several different types of errors. Some errors result from changes in the context of

skilled behavior. If a pattern of behavior has become automatic and we change some

aspect of it, the more familiar pattern may break through and cause an error. A

familiar example of this is where we intend to stop at the shop on the way home from

work but in fact drive past. Here, the activity of driving home is the more familiar and

overrides the less familiar intention.

Other errors result from an incorrect understanding, or model, of a situation or system.

People build their own theories to understand the casual behavior of systems. These

have been termed mental models. They have a number of characteristics. Mental

models are often partial: the person does not have a full understanding of the working

of the whole system. They are unstable and are subject to change. They can be

internally inconsistent, since the person may not have worked through the logical

consequences of their beliefs. They are often unscientific and may be based on

superstition rather than evidence. Often they are based on an incorrect interpretation

of the evidence.

A classification of errors

There are various types of errors. Norman has categorized them into two main types,

slips and mistakes:

Mistakes

Mistakes occur through conscious deliberation. An incorrect action is taken based on

an incorrect decision. For example, trying to throw the icon of the hard disk into the

wastebasket, in the desktop metaphor, as a way of removing all existing files from the

disk is a mistake. A menu option to erase the disk is appropriate action.

Slips

Slips are unintentional. They happen by accident, such as making typos by pressing

the wrong key or selecting wrong menu item by overshooting. The most frequent

errors are slips, especially in well-learned behavior.

101

Table of Contents: