# Evaluation of UIs

# Recall

The circular/iterative approach. Design to implementation to evaluation and repeat. Let's focus on the evaluation aspect.

# Evaluation

There are three kinds of evaluations: passive, active, and predictive. The purpose of conducting these are to determine if the goals of UI are achieved. Recall the previous lecture:

• Security
• Efficiency
• Consistency
• Usability
• Reliability
• Functionality

# Example - Restaurant ordering

Goal	Satisfied?	Yes/No
Functionality	when ordering, we can only add menu items but can't remove them	No
Consistency	every time I click a menu item, I see additional info about the item	Yes
Efficiency	every time I click a menu item, it goes back to the ordering page	No
Usability	from menu page, you cannot access the order page without adding an item to get there	No
Reliability	the calculated total is wrong	No
Security	credit card system used encrypted approach	Yes

# Passive Evaluation

Gets feedback from users 'after the fact'.

• customer reports (must be designed in a way to compel users to fill it out)
• logs (observing)
• surveys

# Problem Report Monitoring

Users should have an easy way to report their problems. Ideally, it is built into the software. This can take the form of form that the user fills out describing their issues.

# Automatic Software Log

Gathers data about user usage (telemetry). Privacy may be a concern.

# Surveys

Surveys are a useful tool to get statistics on a large number of users. The survey must be well designed and worded to gather the best information.

# Active Evaluation

• observations (video)
• scientific experiments
• usability testing

# Observing Users

Users can be observed by video recording sessions who are actively using the UI. This can be done In-Vivo ('at home' in the person's environment) or In-Vitro (lab setting). The users can be prompted to 'think aloud' and vocalize their thoughts.

# Scientific Experiments

Develop a hypothesis with independent and dependent variables, perform it, and analyze the results.

# Usability Testing

A process to identify whether the UI meets the usability requirements previously established. This is partly engineering and partly science. It isn't as rigorous as scientific testing.

# Predictive Evaluation

• heuristic evaluation

The most popular approach of the three is heuristic evaluation. See future lectures for coverage.

# UI Evaluation Vocabulary

Proper functionality

• The system does not function properly because ...
• The UI does not allow all the desired (or necessary) functions, such as ...

Coherence

• The UI is incoherent because ...
• The UI does not use standard icons (or messages), it makes it incoherent with ...

Efficiency

• The system is inefficient because it requires additional steps to ...
• The system's efficiency is reduced because ...

Usability

• The user will have difficulty using the system because...
• The desired option is not visible... and reduce the usability of the system.○The error message (feedback) is confusing, and will make the user less able to use the system ...

Reliability

• The errors obtained by the operation ... show the lack of reliability of the system
• The system does not seem reliable because ...

Security and data integrity

• The user's personal information seems compromised because ...
• The system is not secure as it shows ...