Human Computer Interaction 7th Homework

1. Explain where evaluation sits in a user-centred design process

Unlike waterfall model' summative evaluation, in UCSD process, evaluation is done after the prototype implementation. But after the evaluation, it can be iterative process. Therefore, it called formative evaluation.

And as soon as an idea is generated, it can be tested. It can say evaluation is done all the time.

2. Discuss the key evaluation techniques

There are five key evaluation techniques.

● User observation

It is good to know what users thinking and actions for designer. but it is hard to know about this naturally. So there are kinds of the method to user observation. Video recording, Think-aloud, computer logging, theoretical memos and diary study, etc...

Based on the outcome of observation, decisions need to be made about what to do with the prototype system by considering whether it is good enough to be fully imlemented and Whether modifications need to be made.

It overcomes many disadvantages of experments. But it may be difficult to replicate findings and it enjoys less scientific rigour.

There are something to avoid subjectivities to improve reliability of observational study.

- Hawthorne effect : Users increase performance to please the observer. Users know that observer looks to themselves. So they want to show good actions and mind. It is not good to user observation.

- Observer bias : An observer only sees and records what they want to see. So if there are bad points about design, they would be ignored.

- Halo effect : The observer’s decision is influenced by another positive judgment. If users are handsome and polite man, observer would be system is also good.

● User experiment

Usually define a hypothesis(or a question) and adopt a clear experimental method(a way of ansering the question). The experiment itself has two distinct stages: implementing the experiment and analyzing the results.

A good experiment tends to be systematic, with a repeatable approach to testing based on scientific rigour. However, a reduced consideration of specific variables and unrealistic artificial settings.

It several aspects of experimental design.

1. Hypothesis is some prediction we wish to test.

2. Need to be explicit about what kind of difference we expect to see

- The things we will measure in the experiment called ‘dependent variable’. Therefore designers need to be things that are easily observed and measured.

3. Need to state what we will change between different sets of subjects

- The things we change called ‘independent variable’

4. Avoid to confounding variables

● Cognitive walkthrough

A cognitive walkthrough is an approach to formative evaluation without users. Members of the design team perform it. The user is 'represented' by a set of relevant questions that are applied to each action in the system.

The cognitive walkthrough is based on the assumption that users explore the interface to learn. This learning is example-based where users look to learn how the system in general works by trying some of its features and learning general principles.

● User interviews / questionaries

It tend to be used to find out information about the users' backgrounds, their views and their opinions about the system. These method help us understand the user population.

Two kinds of interviews

- Structured interviews provides a strict sequence of questions. It is useful for comparing the responses of interviewees, but It doesn’t allow the evaluator to ask follow-up questions.

- Unstructured interviews allow the inteviewer to choose the order and nature of the questions. It allows the interviewer to follow up interesting issues and topics, but difficult to collate and analyze data provided by this method.

And I get another kinds of user interviews.

- informal, conversational interview. There are no prepared questions and subject. They talk freely any of subjects.

- general intervies guide approach. It prepared subjects. So it points subjects but in the subject, they talk freely.

- standardized, open-ended interview. It prepared subjects and questions. So all the interviewee got the same questions. It is good to get the data and interview progress is fast.

Unlike interviews, questionnaires usually involve subjects filling in responses to questions themselves rather than responding to an interviewer. Therefore questionnaires can involve a larger set of people as there is no need for the interviewer to ask each subject for their answers. Extensive use of questionnaires is usually referred to as a survey. But it is hard to get the pool as people don’t devote time to answering long lists of questions and questions may be confusing or need clarification.

There are also two types of questionaries

- Open questions is the subject is free to write their answers.

- Closed questions is the subject selects an answer from possibilities, with a response form.

- Simple checklist : simple response such as ‘yes’, ‘no’, or ‘don’t know’

- Multiple checklist : Multiple response such as 'What is team leaders role?' 'responsibility of project', 'make the order of the day', 'solve the complications between members', 'choose the membe', 'etc'

- Ranked order : require uses to specify their preferences for items in a list

- Multi-point rating scale : a number of points on a scale w.r.t. a range of responses

● Heuristic evaluation

It is a checklist for testing usability (e.g., Nielson’s heuristics). For example, in Nielson’s heuristics, there is 'Aesthetic and minimalist design' statement. Then designer check the interface is complicated or not.

Reference

Serengul Smith-Atakan, <Human-Computer Interaction>, THOMSON, 2006, pp.137~154

Professor Chang Wook Ahn, Chapter9(Nov_7).pdf

김희철, <인간과 컴퓨터의 상호작용:인컴학을 향하여>, SciTech, 2006, pp.280~

Human Computer Interaction 7th Homework.pdf

크리에이티브 커먼즈 라이선스

이 저작물은 크리에이티브 커먼즈 코리아 저작자표시-비영리-변경금지 2.0 대한민국 라이선스에 따라 이용하실 수 있습니다.