Question 2

CONTEXTUAL OBSERVATION/ INTERVIEW DATA COLLECTION—

TEMPLATE WITH SAMPLE DATA
This template can be used for data collection in any contextual observation/ interview. The sample data
is adapted from a property processing task in a metropolitan police department. In this sample, a police
officer brings in a prisoner with property to process.

Stationhouse Contextual Observation/interview

Data Collection
Actor(user) Trigger Use Case(Task) Task Scenario Sequence Error, Problem, Comments
Police officer Arrest Fill out property 1. Respond to 911 call. Police officer may do
form for all this type of task only
property associated 2. Pick up perpetrators and take property (knife, four times per year.
with an arrest (in bag, of marijuana) into custody
this case a domestic Use case may
3. Report to station commander at police station currently take two to
assault).
4. Write up property summary so station six hours to complete.
commander can make entry in station log. High frequency of
5. Secure prison. missing data, incorrect
data, missing forms.
6. Store property in locker.
Very high frequency of
7. Fill out arrest paperwork obtain arrest number. interrupts and
distractions during
8. Contact district attorney task.
9. Dispose of prisoner (must be done within two
hours of arrest

Property clerk Make corrections to Look up previous incident in station log by date High frequency of
interrupts with a previous property and police officer name missing info in station
questions regarding form log
a previous property Trace info from station log to property log
form High frequency of
Get info from property log, and supply it to missing info in
property form can e corrected property log

(return to arrest (Continue filling out 10. Retrieve property from locker High frequency of
task) property forms missing numbers.
from arrest) 11. Back to station commander to get property
forms and property packaging Property forms
drafted before typed
12. Fill out draft property form. due to high level of
13. Back to station commander to approve draft errors and difficulty
property form. correcting typewritten
forms.
 SAMPLE WORK ENVIRONMENT ANALYSIS
This sample is adapted from excerpts from a document describing the work environment in the local
police stations of a metropolitan police department.

Stationhouse Work Environment Analysis

The police station environment is described in terms of both its physical characteristics and the
sociocultural environment, which is a distinct subculture within the overall police department
"corporate culture." Both the physical and sociocultural work environments have specific implications
for the design of user interfaces to computer systems that will be used within that environment, and
these are highlighted in indented text and set off by rules above and below.

Physical Work Environment

While the exact physical layout of each local police station is different, there are important
commonalities.

Open layout: With the exception of the arrest room and the property closet, which are distinct rooms
with ceiling-height walls and in some cases doors, all areas are part of a large open room, defined only
by desks and low walls. Even in the arrest room, there can be a lot of traffic coming and going. Thus
there is almost no privacy, potentially a high volume of noise, and constant distractions and
interruptions.

A software user interface in such an. interrupt-driven, noisy, and distracting environment must add
minimal cognitive load to the user's job. Thus, the interface needs to be very intuitive and easy to
understand and use. There need to be

 Effective attention-getting cues to highlight key information,

 A means to suspend an activity and pick it up again, perhaps after completing another task
altogether,
 Plenty of context information to help users keep track of their tasks between interruptions, and
 Lots of support for the limited capacity of human short-term memory for transient information.

Oppressive atmosphere: In general, the police station buildings are fairly old, and they look it. The
interiors, especially when compared to a typical business office, can at best be described as gloomy,
dismal, and run-down. Wall color tends to be dark and neutral, adding to the gloomy atmosphere. The
walls are plastered with miscellaneous notices and reminders, mostly reflecting time pressures (e.g.,
"arrest info must be entered on-line within one hour") and ominous demands to follow rules and
procedures in order to avoid penalties of various kinds (e.g., "do not allow prisoners to use the water
fountain," "be sure to check the prisoner for potential weapons").

There are few if any windows, and the police station is lit with bright, harsh overhead lighting. Some
police stations are air-conditioned and some are not. On one of our site visits a station commander
commented that "everyone is especially tense as it is so hot." In another, a radio was playing during our
whole visit, adding to the already noisy environment.

The station commander's desk is raised up from floor height and tends to be an area of central focus in
the police station. Its size, height from floor level, and a metal bar surrounding it all create a distance
between the station commander and anyone approaching the station commander’s desk, which seems
to emphasi2e the authority of the station commander’s rank within the police station. Although there
seems to be an adequate amount of work space in this area, there is usually a good deal of activity
around the area, with people shuffling back and forth behind the desk, and prisoners and police officers
repeatedly approaching the front of the desk to interact with the station commander. The tension
surrounding the station commander's desk is palpable, even in the quieter moments in the police
station. The station commander has a high-accountability, highly stressful job, and this combined with
the constant yet erratic and unpredictable arrival of police officers with prisons and numerous other
complex problems, seems to permeate the general atmosphere of the police station, which can be
described as stressed, pressured, tense, and at times chaotic. Station commanders appear to be sullen
and not terribly positive, due to the stressful nature of their jobs.

Human performance tends to degrade under stressful, high-pressure conditions. Poor lighting and air
can also affect performance. This again means that

 The user interface to any system used in this environment must provide lots of cognitive,
perceptual, and motor support.

In addition, the interface can be made more attractive and appealing by using

 Light colors,
 Pleasant graphics, and
 A positive, helpful (rather than demanding and penalty-oriented) tone.
 Adding guide/tip facilities to a software program may reduce the number of times users need to
ask someone what to do, especially since someone is probably the station commander who has
the most to do and most responsibility.

In sum, in the typical police station, the appearance of the interior is dark, run-down, and cluttered, the
lighting is harsh and artificial, and the air is close and sometimes very hot. The noise level can be high,
the work areas are cramped and cluttered, and the overall atmosphere is tense and high-pressure at
best, chaotic and sometimes riotous at worst.
These conditions most likely have a general impact on morale and certainly will have an impact on
cognitive functioning, which in turn impacts productivity and effectiveness. The user interface must be
carefully designed to

 Mitigate the and possibly extreme degradations of human performance under these conditions.

Sociocultural Work Environment

Almost all the upper-level staff in this organization have worked up through the ranks. Organizations
where this is true tend to have fairly static cultures, because workers absorb the culture when they
come in at low levels and then perpetuate the culture as they rise up into managerial positions. Change
in such corporate cultures is usually slow and resisted, and can often only come from strong leadership
at the top of the organization in the person of someone who has come from outside the organization.

The observations made here are thus most likely based on historical factors, rather than on the
philosophy and style of the current management. It is not our intention to be critical' of or make
recommendations for change regarding either the physical environment or the corporate culture—this
is beyond both our expertise and the scope of this project. We make these observations simply because
any computer system will be introduced in the context of both the physical and socio-cultural
environment, and these factors will heavily influence its reception and usage. We cannot change the
environment, but by being aware of it, we can design the user interface to any on-line system in a way
that best addresses the unique requirements of its users in their environment.

Low teamwork: One main factor in the sociocultural environment is the natural hostility and distrust
between police officers and prisoners. This seems to be exacerbated by an apparent lack of trust and
support between—and even within---ranks within the police department. Every-one is under a great
deal of stress and pressure, and trying to avoid responsibility for errors and problems. The relationship
between ranks seems to be authoritarian and status conscious, rather than supportive and teamwork
oriented. There is high turnover, and to add to this, there is shift rotation among police officers and
station commanders. These in turn mean that a culture of teamwork is not readily fostered. In addition,
everyone is usually extremely busy with his or her own overload of responsibilities, leaving little time or
energy to support one another.

In many organizations, users look to each other for help in learning and mastering new computer
systems on the job. We probably cannot assume this will happen here due to the factors above. This in
turn means that the user interface:

 should be as self-explanatory as possible, and

 should have built into it as many of the business rules as possible, so that users need minimal
help to use it successfully even if they use it infrequently.
To summarize, this seems to- be a sociocultural environment in which (especially compared with a
professional business environment) office morale is low, negative pressure and stress is high, teamwork
is not fostered, the only reward is escape, and perfect performance is demanded of workers under
conditions that seem almost specifically designed to impede performance. In addition, staff have had
mostly negative experiences—if any—with automation to date.

 Ease of use,
 ease of learning, and
 task and user orientation in any system user interface in this environment will be critical to its
success.

SAMPLE TASK ANALYSIS DOCUMENT

This sample is adapted from excerpts from two sections of a Task Analysis document describing the
current task flow and its weaknesses for processing property in the local police stations of a
metropolitan police department. In the next section I separately offer sample Task Scenarios, which
would also be included in the Task Analysis document.

Stationhouse Task Analysis

Purpose and Structure of This Document
This document presents a user-oriented perspective and analysis of the tasks that will be automated by
the application. The functional spec describes processes and activities in traditional systems analysis
terms. Here we are describing user tasks from the perspective of the people performing those tasks in
their actual work environment.

Whereas the formal description of abstract and general processes and activities in the functional
spec documents the way things should be done and how they are done when everything goes smoothly
and correctly, here we focus on real-life instances of tasks and highlight where things can go wrong or
don't work the way the worker needs them to in order to perform his or her job effectively. This focus is
key

to meeting user interface design requirements for this system and ensuring that the functionality
provided best supports the real-life tasks users will perform with the new application in their real work
environment.
 This document thus presents an analysis of the tasks to be automated from a user interface
design perspective. Direct implications for user interface design are highlighted in indented text and set
off by rules above and below.

Analysis of Current Work and Tasks

In this section, we make general observations from our contextual observations/interviews regarding
the user tasks being automated, highlighting problem areas, bottlenecks; and opportunities for
improvement, where the design of the user interface can have an impact.

Broad range of task complexity: The complexity of the property processing task can vary significantly.
One police officer told us she once had to complete twelve separate property forms for a single incident;
another time she saw an officer come into the police station with fifteen large bags of narcotics and
related materials associated with a single Incident, all of which had to be logged on property forms. By
contrast, we watched another police officer come in with no prisoners and a single piece of property to
process: a spent shell from a handgun. Certain property types, such as automobiles, guns, and cash,
always entail quite a few complex forms, while others involve only simple paperwork.

The user interface must keep simple transactions simple but make complex transactions possible.

High degree of interrupts during tasks: The property processing tasks are subject to a high level of
interruptions. In the case of the police officer, the source of interruptions may be prisoners, other police
officers in the arrest room, other aspects of processing an arrest besides property, or having to leave the
arrest room to gather data for the property forms, for example, to check an auto for information that
needs to be included on the property form.

A software user interface in such an interrupt-driven, noisy, and distracting environment must add
minimal cognitive load to the user's job. Thus, the interface needs to be very intuitive and easy to
understand and use. There need to be

 effective attention-getting cues to highlight key information;

 plenty of context information on the screen at all times, so that if users are distracted, when
they turn their attention back to the screen, they can quickly reconstruct where they were and
what they were doing and carry on; -
 lots of support for the limited capacity of human short-term memory for transient information;
 Multitasking, so that users can suspend one on-line task, start another, and then return to the
first task without loss of work or context.

SAMPLE TASK SCENARIOS

Task Scenarios represent concrete, detailed, real, and typical work tasks a user might encounter in his or
her job. They are drawn from and reconstructed from the contextual observations/interviews. They
might represent composite scenarios made up of parts of scenarios actually observed. They are written
strictly in terms of abstract user tasks and goals, and make little or no reference to software or other
tools currently used to get the task done. In the parlance of DOSE, they represent instances of Use
Cases.

The first Task Scenario is written in the format and style that would be used in the Task Analysis
document. Note the use of the OOSE concept of an extension (in this case, an interruption). Following
that is a Task Scenario that has been rewritten from the Task Analysis document to serve as a prop in a
usability test.

Stationhouse Task Scenario

Task: Complete property form User: Police officer
Description: Domestic dispute—couple filing charges against each other. He beat her with a. broomstick,
she attacked and cut him with a knife. Someone called 911 to report the incident, and the police officer
picked them up and brought them in, with the knife as evidence, a beeper for safekeeping, and a bag of
marijuana, which was taken from the premises where the arrests were made.

Task Flow:

1. Police officer (PO) goes to station commander's (SC) desk with perpetrators and shows all
property, filling out a summary form for the SC. SC makes entry in command log.
2. PO secures the prisoners, takes fingerprints, and checks for warrants.
3. PO stores property in her own personal locker.
4. PO writes up an arrest worksheet and Prisoner Release forms, questioning the prisoners to
obtain required information.
5. PO records data from the arrest worksheet and obtains an arrest number, which she writes on
her hand.
6. PO contacts the assistant district attorney (ADA), faxes arrest worksheet to ADA office.
7. PO disposes of the prisoners (she must within two hours of the arrest, or the SC will have to file
a special report explaining why she did not). Property forms are not yet done, so prisoners do
not get copies, as they should.
8. PO fills out a complaint worksheet.
9. PO records data from the complaint worksheet and obtains complaint 'writhe! s.
10. PO retrieves property from her locker.
11. PO returns to the desk, and she gets property forms and bags, for which she signs out on the
scratch property log (in other police stations, the SC distributes the bags and property forms and
keeps the scratch property log).
12. PO bags the property and enters the relevant property form number on each bag.
13. PO fills out property form worksheets (optional).
14. PO takes bags and property form worksheets to SC for checking (optional).
 15. PO types up the property forms herself, recording the arrest number, complaint number, and
bag numbers, and also types up a letter of transmittal for the narcotics, a Controlled. Substance
Analysis form, a Domestic Incident form, and a Medical Treatment form (with a great deal of
repeated header info on all forms).
16. PO takes all forms to the SC for approval. He checks especially all the cross-referencing of
numbers (arrest number, complaint numbers, property form numbers, bag numbers), signs off,
takes copies of forms, and makes entries on scratch property log and in station log.
17. PO seals all bags and attaches property forms.
18. PO disposes of all property and property forms: narcotics in a locked narcotics locker, other
property in property closet.

Task Closure: This scenario took from 1 p.m., when the arrests were made, until almost 4 p.m. to
process.

To support this task, the user interface should

 Eliminate all redundant entry of data, such as information about the arresting police officer,
prisoners, incident; and
 Facilitate the cross-referencing between such things as the complaint, the arrest, the property
forms, and the property bags.

Insurance Task Scenario

Enter application data on-line for a new insurance policy for an existing customer and handle two
customer phone calls.

The phones are quiet, there are no sales agents around, so you decide to enter data in an on-
line application form for a new policy that has been sitting in your in-basket. Humphrey Bogart, an
existing client, has applied for a new insurance policy. -

All Humphrey Bogart's information is available on-line, and you don't need to reenter any of it,
you just need to complete an on-line application form.

Procedure
Get a new policy application started for Humphrey Bogart, and start filling in all the necessary
information to complete application.

The phone rings, and you answer it. Lauren Bacall is calling to report a change in address. She gives you
a policy number. You need to suspend your data entry task and locate Lauren Bacall's information to
make changes.

When you find Lauren Bacall's record, check to see if she has any business that includes other people,
such as family members. Since she does, ask her if their addresses will be changing as well. She says yes.
Make the changes as she specifies.
 Wrap up with Lauren Bacall-and hang up. Return to the application you were in the middle Of for
Humphrey Bogart. Complete this application.

The phone rings, and it is John Wayne asking to cancel a policy. He does not have his policy number.
Look him up.

When you find his record, you note that he has several policies. Ask him if he wants to cancel just one or
all. He says all He is in financial difficulties and says he cannot afford to keep any of his policies. Tell John
Wayne that you have all the necessary information and that his sales agent will be contacting him soon.

SAMPLE FIRST PASS AT CURRENT USER TASK ORGANIZATION MODEL

This sample is adapted from part of a design team's original task hierarchy for the customer support
function in an insurance company. In this illustration, the text beneath the boxes represents the basic
user tasks identified from the contextual observations/interviews. The text in boxes represents the
hierarchy of groups developed by 'the design team, with labels they created—again, all based on the
contextual observations/ interviews. (For simplicity of the example, not all groups or low-level tasks are
included.)

Task Organization Model—Customer Support

Customer Support
Representative

Sales Support Customer Support Follow-Up Office

Administration

 Supplies
 Facility
Change Information  Directory
New Sale
 Telephone Bill
Requests Requests  Backup
Policy Type A  Postage Meter
 Beneficiary Change
 Bank Deposits
Policy Type B  Renewal Date Quote
 Address
 Phone
 Claims
 Change
Application
 Quote

Information
Change Requests Requests

 Address Policy New Sale

 Phone
Management

 Incomplete
Task Sorting Exercise—Housekeeping Analogy Application
Imagine you are a housekeeper. Say we are designing your house and deciding where to storeall your
Missing Evidence
housekeeping tools and supplies in a way that best supports your job as housekeeper. Suppose we listed
all your basic housekeeping tasks and asked you to organize them into groups, so we could better
understand how to design your house and store your tools where you would find them and access them
quickly.

Housekeeping Tasks--
A Random Ordering of Lowest-Level Tasks

Vacuum furniture (e.g., couches) Polish furniture

Serve beverages clear the table

Wash floors Vacuum floors and rugs

Rinse dishes Put meal on table

Combine ingredients for cooking Wash pots and pans

Wash rugs/towels in washing machine Dust

Sweep floors Mix up ingredients for cooking

Set the table Wash fixtures (sinks, toilets, tubs)

Cook prepared food Serve food from table

Put dishes in dishwasher

With your (the end user's) input, we hope to end up with something like the task hierarchy shown on
page 119. If we had never done house-keeping ourselves, and we didn't consult you, the task expert, we
might have come up with an organization that didn't support your work at all and made it unnecessarily
tedious and difficult. Imagine how unusable the household would be if instead of organizing your house
along the lines of the hierarchy below, we put your pantry in your attic, stored your dishes in the cellar,
or placed your dishwashing liquid in a bathroom closet along with other types of cleaning liquids!
Something analogous to this usually happens, unfortunately, in the way we organize functionality in
software applications, because as developers; we don't really understand the work users are doing. We
organize according to some logic, but it usually isn't according to user logic or actual work organization.

This hierarchy of housekeeping tasks shows both task organization groupings and fools supporting tasks
(in parentheses) that should be kept together.

Housekeeping Tasks Hierarchy

 SAMPLE CURRENT USER TASK ORGANIZATION MODEL
This sample of part of one user's task hierarchy is adapted from the customer support function in an
insurance company. In the model Lillis_ traced here, the text beneath the boxes represents the low-level
tasks Presented to the user, each task on a separate index card. The bold text in boxes represents the
groups that one user formed from the low-level tasks, with labels they suggested. (This hierarchy
represents a partial set of the functionality required by a customer support rep. For simplicity of the
example, not all groups or low-level tasks are included.)

Note how this hierarchy is different from the one generated by the design team (see the earlier
section Sample First Pass at Current User Task Organization Model). For example, the design team
organizes all types of changes (e.g., Address, Beneficiary) under the category 'Change Requests" within
the category Sales Support. Meanwhile, this user distinguishes between those changes that have to do
with customer information and those that .relate to policy details, but locates both types of changes
under the category "Customer Support" rather than "Sales Support!' As it turns out, customer support
staff can make simple changes to customer information on their own, but more complex paperwork and
approvals are required for policy changes. Thus, these are two distinct categories of tasks to users. And
even though requests for these types of changes often come from customers via sales agents,-this user
still regards them as reflecting customer support.

Similarly, the design team lumps all Office Admin tasks in one group, while this user distinguishes
between Daily and Occasional office tasks. It turns out that different people are assigned to these .types
of tasks, so again, this is a meaningful distinction to users.

The designers imagine a category unto itself called "Follow-Up," where incomplete tasks of all types
(e.g., Incomplete App) are all located; while this user considers follow-up activities as belonging with the
original task type with which they are associated.

Finally, note that the user regards New Sales (entering data in new policy applications) as its own
category, whereas the designers see this as a subcategory of Sales Support, a very different perception.

One important thing to note in comparing this user's task organization model with the designers' first
pass at a task organization model is that for any set of low-level tasks, there may be a large number of
very different logical task organizations possible, but there will only be a small number of rather similar
ones that will make sense to users, given their current work models. A traditional systems analysis will
uncover all the tasks, but it typically does not ensure that the task organization most consistent with the
Current User Task Organization Model will be presented to the user.

Also note that this user's model is most likely not exactly the same as other users' models, although in
most cases, there will be a lot of similarity across individual users' models. As part of step 11 (Obtain
Current User Task Organization Model) in the Contextual Task Analysis technique described in this
chapter, recall that you consolidate all the sampled users' task organization models and generate one
model in the same format that captures as many commonalities as possible. Then, in a later task in the
lifecycle—Work Reengineering—this consolidated user model will be reengineered
 Question 1
USER PROFILE QUESTIONNAIRE TEMPLATE
The following template can be used as a starting point for any User Pro- file questionnaire. Some notes
on how the template is laid out will help when creating your own User Profile questionnaire from this
template: First, in the questionnaire itself, the titles in parentheses should be removed (they are, there
to cue you regarding how these questions relate to the Data Summary Template shown later).

Some questions should be kept together and in the order they Wet in the template, such as questions 1-
4, which establish membership 01 predefined user categories, and questions 17-23, which refer to one
another and therefore should be kept together and in order. Other questions should be randomly
distributed across the questionnaire to red response bias; for example, questions 5-9 poll attitudes and
motivation, and questions 10-16 address knowledge and experience.

Many of these sample questions must be tailored to fit what is already known about the intended
respondents. For example, question I should, list the known job types within the target user population,
and question 2 should be much more specific, listing actual locations of offices if possible.

A short cover letter should he provided with the questionnaire, pointing out the purpose and benefits of
User Profiles. It is helpful to have the cover letter signed by a well-known and highly placed person, to
help motivate respondents to participate (obviously, permission must be obtained from such a person
before using his or her name). An example cover letter is offered here, but you should tailor it to the
particular circumstances.

The template uses two formatting techniques that I highly recommend. First, bold is used to highlight
key aspects of questions and answers. This enhances clarity and readability. Second, considerable white
space is used, especially where respondents are expected to write in answers. Although this makes the
questionnaire appear longer and thus may put some users .off, the increased readability and ease of
response outweigh any drawbacks. If length is an issue, you can produce a double-sided questionnaire
and give users an estimate of the time it will take to complete it, pointing out that it is mostly multiple
choice.

When adding questions that poll attitudes or opinions, order choices in an ascending or descending
order (e.g., "like," "neutral," "dis-like"), but vary the order from question to question, so that some have
the positive choice first and others have the negative choice first. This helps to avoid introducing a
response bias due to implicitly leading the user in one direction or the other. However, in self-
assessment questions, such as typing skill, educational level, or computer literacy, order from lowest to
highest skill level to avoid encouraging users to overinflate their assessments.

In general, always design questions to be neutral, so as not to inadvertently lead users in their answers
and introduce a response bias into the data. Finally, make sure your multiple-choice answers are
mutually exclusive. For example, don't have "1-3" as one choice and "3-6" as another: which would you
choose if your answer was 3? And make them exhaustive (don't make your last category "10-15" if
somebody's answer might be 16 or greater).

XYZ Insurance Application User Profile Questionnaire

Dear XYZ Insurance Application User:

This questionnaire has been prepared by the XYZ Insurance Application project team to help us learn
more about you, the potential end users of our system. The information you and other future users of
the XYZ Insurance System provide through this questionnaire will help us to design and develop a
higher-quality application that will be better tailored to your needs and thus easier to learn and easier
to use.

The questionnaire is anonymous, and we will be summarizing all responses to describe whole categories
of users, rather than referring back to any single questionnaire. The more candid and accurate you are in
your responses, the more useful the information gathered through this questionnaire will be in helping
us to meet your needs.

It should only take you about 15-20 minutes to fill out this questionnaire. It may look long, but most
questions are simple multiple choice. Your input will have the most impact if you return your completed
questionnaire in the enclosed stamped, addressed envelope by Friday, October 22, 1999. However, we
will find your input useful no matter when it gets to us, so even if you miss this important deadline,
please return it whenever you can. To ensure your anonymity, do not put a return address on the
envelope. Your participation is greatly appreciated.
 Best regards,

Mr. Important

XYZ Insurance Company

XYZ Insurance Application User Profile Questionnaire

(User Category Identifiers)
1. Check the job title that best describes your current job:

Clerical

Internal Claim Rep

External Claim Rep

Supervisor

Manager

Other (please. describe)

2. In which geographic area is your main office located?

Northeast

Southeast

Midwest

Southwest

Northwest

California

3. Please estimate how many people in your job title are working in your geographic area:

(If you have no idea,, write "N/A.")

4. Describe the current level of automation of your job title in your office by checking one choice
below:

None (No users in my job title have or use a computer workstation.)

Low (All users in my job title who use the computer share a workstation with
other users.)
 Medium (Some users in my job title who use the computer share a workstation
with other users, but some have their own workstations.)

High (All users in my job title have their own workstations.)

(Attitude and Motivation)

5. In general, how do you feel about working with computers?

I don't like working with computers.

I have no strong like or dislike for working with computers.

I like working with computers.

______________ Other (please explain)

6. How have computers affected your job?

_________________ Computers have made my job easier.

___________________ Computers have not affected my job in any particular way.

____________________ Computers have made my job more difficult.

______________ Other (please explain)

7. Is the amount of time it takes to learn new software applications usually worth it?
Yes, it pays off because computer systems usually help me do my job
better or faster.

____________________ Sometimes it pays off, and sometimes it doesn't.

No, computer systems are : usually not useful enough to justify the
training time.

__________ Other (please explain) ______________________________________

8. Do you enjoy learning how to use new software applications?

_________________ Yes, it's usually challenging and interesting.

____________________ Sometimes, depending on the application.

_______________ No, it's usually tedious and frustrating.

_______________ Other (please explain)

In General, are you interested in computer?

 I am not interested in computers and would avoid using them if I
could.

I am interested in computers but only as a means to help me

do my job better and fa
I am interested in computers in general, and I enjoy using them.
Other (please explain)

(Knowledge and Experience)

10. What is your level of typing skill?

"Hunt and peek"-typist (less than 15 words per minute)

Moderately skilled touch typist (between 15 and 50 words per minute)

Highly skilled touch typist (greater than 50 words per minute)

11. What is your highest academic degree?

No degrees

High school degree

Trade or vocational school degree (beyond the high school level)

College degree (for example, B.A., B.S., Associate College degree)

Graduate degree (for example, M.A., M.S., Ph.D., Ed.D., M.D., R.N.)

Other (please explain)

12. How would you describe your experience level in your current job title?

Novice (less than 1 year)

Experienced (1-3 years)

Expert (more than 3 years)

Other (please describe)

13. What is your native language?

English (go to question 16)

Spanish

Other (please name)

14. If your native language is not English, how well do you speak English (leave blank if English is your
native language)? Poorly (I have trouble communicating with English speakers.)

Adequately (I speak well enough to get around.)

Fluently (I speak almost as well as a native speaker.)

Other (please describe)

15. If your native language is not English, how well do you read English (leave blank if English is your
native language)?

Poorly (I have trouble reading documents in English.)

Adequately (I read well enough to get around.).

Fluently (I read almost as well as a native speaker.)

Other (please describe)

16. How would you describe your general level of computer experience?

None (I have never used any software applications.)

Low (I have used only one or two software applications.)

Moderately low (I have learned and used between

Three and ten different software applications.)

Moderately high (I have learned and used more than ten different software applications
but have no programming skills.)

High (I have used many different software applications and have some programming
skills.)

Other (please describe)

(Job and Task Characteristics)

17. Please name all the software applications you currently use in your job, and indicate how long you
have been using them. Under "Business Applications" (1-5), list any specialized applications built by or
through MIS at XYZ Insurance Co. to service your business. Only list systems you personally use. Under
"Office Applications" (A-E), list any general-purpose commercial packages such0 as word processors,
spreadsheets, and so on that you use.
 Years/Months

Application Experience
Business Application:

1.

2.

3.

4.

5.

Office Application:

A.

B.

C.

D.

E.

18. On the average in your geographic area, how long do people stay in your Job title before
leaving the company, being promoted to other titles, or leaving the job for any other
reason? .

I don't know

Less than six months

Six months to a year

Over a year, up to three years

Over three years

19. Approximately what percentage of your total work time do you spend doing your
work in the following different locations (your answers should add up to 100
percent)?
Your field office
 Your home

A hotel room

A claimant's location

Another field office

Other (please name)

(Physical Characteristics)
20. Are your

Male

Female

21. Are you:

Right-handed
Left-handed
Ambidextrous (equally coordinated with both hands)
22. Are you color blind in any way?
No
Yes (please describe)

23. How old are you?

18-25

26-40

41-55

over 55

24. Do you wear glasses or contact lenses?

No

Yes (Please describe your vision problem and correction method, for
example, nearsighted, farsighted; bifocals, contact lenses.)
 25. Do you have any physical handicaps other than vision deficiencies that computer
technology would need to accommodate or support (for example, hard of hearing;
arthritis in hands, wheelchair)?

No

Yes (Please describe)

USER PROFILE DATA ENTRY AND ANALYSIS TEMPLATES

Separate data entry and analysis should be done for each major user category, if relevant.
For example, the data from all clerical staff should be entered and analyzed separately from
the data for the internal claims reps, external claims reps, supervisors, and managers.
Similarly, data from all physicians should be entered and analyzed separately from the data
for technicians, nurses, and receptionists. An analysis of the whole population can always
be synthesized from the separate analyses, but you will want to draw conclusions and
implications for each user category separately as well.

Here I offer two User Profile data entry and analysis templates, one assuming a
spreadsheet program, and the other assuming just paper, pencil, and calculator. The sample
spreadsheet below shows the data from two questions and three users. Raw data is
entered from each respondent's questionnaire, putting a 1 in any cell that represents an
answer chosen for that question by that user, and leaving blank or entering 0 in cells
representing all other answers to that question. Then, sums and percentages across users
for each question are calculated by entering simple formulas in the cells of the Sums and %
rows.

As you can see in the spreadsheet at the bottom of this page, after the raw data was
entered, the spreadsheet calculated that 67 percent of users answered "Yes" to question 8
about whether they enjoy learning new computer systems, and 33 percent answered
"Sometimes." In response to question 9 about how interested in computers they are, 67-
percent answered "Only as a means to get my work done and 33 percent answered "I am
interested in computers." In most cases percentages across answers within a single
question should add up to 100 percent (unless there is missing data, i.e., some respondents
failed to answer some questions).
 If data entry and analysis cannot be automated, you can still make up a simple data
entry sheet onto which raw data can be collated and analyzed. A sample of such a format,
using the same two questions and three respondents appears on page 57. As you go
through each questionnaire, you would enter a tick mark in the Tally column for the answer
the respondent chose for each question (here you can see that two users chose "Yes" to
question 8, and one user chose "Sometimes:' etc.). Then compute the percentage by
dividing each sum by the total number of respondents. Again, in most cases percentages
across answers within a single question should add up to 100 percent unless there is
missing data.
 USER PROFILE DATA SUMMARY TEMPLATE WITH SAMPLE DATA
Following is a template for summarizing the data from the data entry and analysis
templates in a more readable way. The header section identifies the set of users whose
data is summarized on this form, and' a separate summary should be developed for each
major user category, if relevant.

Each question from the original questionnaire is recorded on the form, using question
numbers and abbreviated phrases for .both questions and answers (only the first several
questions from the questionnaire are represented here to give the idea of the format). The
questions are grouped in related categories, rather than in the exalt order in which they
appear on the questionnaire (where order was intentionally randomized). The percentage
of respondents who answered in each multiple-choice category is recorded for each
question. For example, in question 5, an attitude question, the data might indicate that 13
percent "don't like," 48 percent are "neutral," and 39 percent "like" computers. In most
cases percentages should add up to 100 percent (unless there is missing data, i.e., some
respondents failed to answer some questions).

XYZ insurance Application—

User Profile Data Summary

Overall Business Organization: Individual Insurance

User Category:

Job title: Supervisors All

Geographic area(s): All

Total respondents: 43
% of total at geographic area: 33%

Level of automation: High

Attitude and Motivation:

5. Feel about computers 6. Affected your job, 7. Learning pays

.13 don't like .16 more difficult .24 no

.48 neutral .56 neutral. 39 neutral

.39 Like .28 easier .37 Yes

8.Enjoy learning apps 9. Interested in computers

.74 no .10 not interested

.14 sometimes .78 only as a means

12 yes .12 interested

Knowledge and Experience:

10. Typing skill 11. Highest degree 00 no 12. Job experience

.91 < 15 wpm .00 no .05 < 1 year

.06 15-50 wpm .05 high school .07 13 years

.03 > 50 wpm .00 trade .88 > 3 years

.76 college .00 others

.19 graduate
 SAMPLE USER PROFILE CONCLUSIONS ANDDESIGN IMPLICATIONS
The sample offered here is based on a User Profile of seven categories of plant floor workers in a
manufacturing company, which is in turn based on a questionnaire and data analyses and summaries
similar to those offered in the preceding templates. The categories of users in the complete User
profile included production workers, managers, supervisors, quality assurance specialists, clerical
workers, skilled tradespeople, and engineers. This sample contains the analysis and conclusions in the
User Profile for two user categories: production workers and engineers. Pay particular attention to the
differences in user characteristics between the two categories and the corresponding difference in
user interface requirements. Also included is a summary table liming key, general usability goals to
user categories according to user characteristics revealed in the User Profiles. This table gives an
example of how to link information gained during the User Profiles task with goals set 'during the
Usability Goal Setting task.

ABC Factory Applications—User Profile Conclusions

Production Workers

General Description

Production workers are hourly workers, including machine operators, assemblers, and hourly
personnel other than skilled trades. They include people performing job functions as an integral part
of the production process, such as an assembler on a final assembly line. Other production workers
work as machine operators, responsible for the operation of one or more pieces of production
equipment.

There are a total of 7,834 production workers, representing 62 percent of the total plant floor
workforce, working in four different plants all in the same city.

User Characteristics

Among production workers, general attitude and motivation towards computers are fairly high, but
not as high as most other User categories (e.g., clerical and QA). However, nearly 73 percent of these
users do not perceive computers to be important to their jobs.

Educational level is generally high school degree or less—lower than any other user category.
By inference, reading skill would probably aver-age around the eighth-grade level.

Job experience levels are-quite high, comparable to other user categories. -Turnover is low,
and generally lower than other categories.
 Computer experience, by contrast, is quite low, and considerably lower than other user
categories. Frequency of computer use is quite low, and significantly lower than other user categories,
and use is most often - discretionary.

These users get very little, and brief, training support for new systems—less than any other
user category.

A majority have no typing skills, and only a small fraction are experienced typists, significantly
less than in other user categories.

A majority of these users are male—a mix not terribly discrepant from other user categories.
By inference, a little over 7 percent (599) have some form of color vision deficiency. A substantial
majority (69 percent) wear corrective lenses, comparable -to other user categories. Virtually 100
percent wear protective eyeglasses' on the job.

This user category is currently about 67 percent over forty in age comparable to other categories
except engineers, who are generally younger.

Usability Requirements

Production workers have a high need for ease of learning, especially in the general computer (as
opposed to job and task) aspects of usage, due to their lower educational levels, low computer
experience, low frequency of use, discretionary use, minimal training support, and their perception
that computers are currently not particularly important to their jobs.

However, they are very experienced on the job, their attitude and motivation regarding
computer usage are high, and their turnover rate is low. This suggests that power without complexity
(i.e., simplicity) is also important. That is, the power ought to be immediately perceivable, not hidden
by a complex and difficult-to-learn user interface.

Very low typing skills suggest an interface with absolutely minimal typing requirements.

Low reading skill and the prevalence of corrective lenses suggest that icons and visual displays
(rather than verbal ones) will be useful. Any text that is displayed should be written at about the fifth-
grade reading level. The prevalence of corrective lenses and the general older age of these users also
suggest that text and symbols should be adequately large.

The fact that most production workers are male suggests that the use of color must take into
consideration a significant (8 percent) incidence of color blindness.
 Engineers

General Description

Engineers are trained, salaried employees who design, purchase, install, and support production
processes and equipment. They include process engineers, plant engineers, controls engineers, and
other professions that support plant processes and operations.

There are a total of 972 engineers, representing 9 percent of the total workforce, working in
four different plants all in the same city.

User Characteristics

Among engineers, attitude and motivation towards computers are generally high, higher than among
production workers, comparable to skilled tradespeople and managers, but not as high as among
clerical and.QA users. Of these users, 82 percent perceive computers to be important to their jobs.

Educational level is very high, with 96 percent possessing at least one college degree. This is
significantly higher than any other category. By inference, reading skill would conservatively average
between tenth-and twelfth-grade level.

Job and task experience levels are moderate to high, with 54 percent possessing eleven years
or more experience, although this-percentage is significantly less than any other user category. A
significant portion of the total category (35 percent) possess three years or less experience in their
current job. As compared with other categories, engineers have the highest turnover rate, with 21
percent turnover per year.

Computer experience is moderate to high, generally a bit higher than most of the other
nonhourly categories, and considerably higher than the hourly categories. Frequency of computer use
is moderate, higher than production workers and skilled tradespeople but lower than other
categories. This category is second only to supervisors in the percentage of members using four or
more different computer applications (66 percent). Use is most often discretionary.

Engineers have moderate levels of training support for new systems available to them, less
than some categories but more than others. Training is typically eight hours or less in duration.

A majority are experienced typists, second only to the clerical user Category.

A majority of these users are male—a mix not terribly discrepant from other user categories.
By inference, a little over 7 percent (65) have some form of color vision deficiency. A substantial
majority wear corrective lenses (77 percent), comparable to other user categories.

Only about 38 percent of engineers are currently over forty in age, and engineers are thus
significantly younger than any other category.
 Usability Requirements

Engineers are very experienced on the job, and their frequency of use is relatively high, their
educational levels are high, their computer experience is moderate to high, their training support Is
moderate, and their perception is that computers are important to their jobs. This suggests that ease
of use and power are important.

High levels of typing skills suggest that typing requirements will not be a problem.

High educational and reading skill levels suggest that icons and visual displays are not as
important as to-other user categories. The prevalence of corrective lenses suggests that text and
symbols should be adequately large.

The fact that most engineers are male suggests that the use of color must take into
consideration a significant (7 percent) incidence of color blindness.

Usability Requirements Summary

The following table summarizes the important usability requirements according to user-category.

Key: blank not important, x important, xx = very important

Usability requirements named in the table are defined below.

Ease of learning: How easily and quickly can users learn to use new computer systems, with or
without training? This is important for infrequent users and/or users who may not have access to
formal training. It is also important for users with negative 'attitudes and low motivation regarding
computers-and-their jobs, arid/or with low computer literacy.

Ease of use/power: How quickly, easily, and efficiently can users accomplish tasks once they have
been learned and what range of tasks can be easily accomplished? This is important-for users with
high experience levels, high frequency of use, positive attitudes, high motivational levels, and a need
for efficiency and speed.

Simplicity: Is there a need for a low level of Complexity in order to accomplish tasks? This includes
conceptual complexity (minimizing new concepts required to perform additional tasks) and the
Complexity of actually performing human-computer interactions (minimizing the actions required to
communicate with the system).

Use visuals /icons: Should information be presented as icons and in other visual, graphical formats, as
opposed to text and numbers? This includes the use of visual cues such as color, reverse video, bold,
and so on. Many users will benefit from the use of graphics and visuals, but particularly users with low
ability in reading or other language skills.

Minimize typing: How significant are typing skills and familiarity with keyboards? "Point and select" is
an alternative interaction style to "remember and type."

Color vision deficit: How common is color blindness? Special care must be taken in using colors in the
interface if a significant number of users have some form of color blindness. In particular, pure,
saturated reds and greens should be avoided or used only with some other redundant cue.

Other vision deficit: How common are other vision problems? The interface must be usable by users
with imperfect eyesight if significant numbers of such users are part of the target population. This
means text must not be too small, and graphics and selections must be large and detailed enough so
that users can discriminate among them.