Interviews can be used to gain information about a system and how it is, or will be used.
Often an audio tape recorder is used to record the dialog but pencil and paper can be
used with or without the tape. Generally, three types of interview can be used.
Interviews can be unstructured, semi-structured or structured.
1. Semi-structured interview. When the designer has a better understanding of the
system requirements, a more focused, or semi-structured interview technique can be
used to focus on the points of interest. However, there can still be a degree of flexibility
to allow the user to expand on an answer.
The interviewer has a set of prepared questions which are put to a single subject at a
time and their answers are recorded. The advantage of having the interviewer there,
rather than just asking the subject to fill in the questionnaire, is that responses can be
followed up on the spot with supplementary questions, requests for clarification, etc.
which have not been prepared in advance.
2. Questionnaires. These might be filled in by a group of people while the designer is
present (eg in an office) or, more usually, they will be sent out and returned by post. The
disadvantage, of course, is that up to 75% of the people who receive a questionnaire
never bother to return it, so a large number have to be distributed in the first place in
order to get a reasonable sample. [The problem then might be that the kind of people
who return questionnaires may not be representative of the user population at whom
you are aiming the software!]
A well-designed questionnaire should adhere to the following points:
it should be concise - usually not more than 2 pages
it should be unambiguous - scale ranges shouldn't intersect, questions should be
well posed and explained where necessary.
different kinds of ranking scales used shouldn't be mixed too much - users tend
to answer only the easier ones.
it should have at least one open question for respondent's comments.
the designer should personally perform a pilot study to test the questionnaire.
3. Incident diary. Over a specified period of time users of the system are asked to keep
a diary of any significant incidents that occurred relating to their use of the product. This
can be effective, but it is probably best used when the system is almost completed and
it is known that it no longer contains a large number of flaws.
4. Feature checklist. A list is made of all the features in the product. For example, a list
for a word processing package might contain items like : insert footnote; move
paragraph; change font style; change font size; insert date; change margins; and so on.
In using the software, a tick is placed against a particular feature each time it is used.
This gives very useful information about the most commonly used features of the
product (usually around 20% of all the possible features) and these might be
concentrated on for enhancement in later versions of the package.
Checklists can be used to ensure that all requirements set out by the client have been
met. These might include the type of information available to the user, functions
available on each screen, 'help' available, or time required to perform tasks. Other more
abstract checks may be made, such as the look of the screen, colours used,
background-to-text contrast, or the learnability of the interface. However abstract these
may appear, decisions on measuring or defining aspects of the interface, should have
been made at the outset.
Other checklists can be utilised, such as a 'user profile' checklist. This would be used by
the designer to identify the skill level and experience of prospective users.
5. Focus group : A group of people representing different facets of a product -
designers, programmers, managers, users etc. who get together to discuss particular
strengths and weaknesses of the product (or prototype(s)).
They can be described as informal meetings used as a forum to openly discuss topics.
The group members discuss experiences with interfaces, or features they would like to
see in a future design. The group leader has an agenda, co-ordinates the meeting,
makes sure everyone has their say, while allowing the group members to discuss
6. Think-aloud : This can be a very effective method of determining how a user feels
about a product, any conceptual difficulties they might have in understanding it, or
difficulties in remembering sequences of operations, and so on. The user is asked to
work through a set of specified tasks and is encouraged to talk about what they are
doing, what they are thinking, etc. Comments and thoughts are recorded by an
7. Interactive experiment : A large group of people would be asked to work through a
specified set of tasks and their performance would be observed and measured in some
way. Quite an expensive method : a sample of 30 or more would normally be deemed
minimum for obtaining significant results.