About the Project
This section provides a brief description of several theories and models for Human-Computer Interaction (HCI) as applied to design of Web-based survey instruments.
The perceptual system first tries to establish what the central object of view is (figure) and everything else is considered the context (ground) e.g. the faces/vase
Proximity: the tendency of individual elements to be associated more strongly with nearby elements than those far away
Similarity: elements that share common visual variables such as color will be more associated than those that do not
Continuity: the preference for unbroken figures with the simplest possible explanation
Closure: the tendency to interpret forms as complete whole figures even though some pieces are missing
Area: smaller of two overlapping figures will be interpreted as figure while larger is interpreted as ground
Symmetry: grouping based on properties of the entire form instead of its parts
Illusions present us with conflicting or confusing information that cause distortions in perceptions.
Depth cues: smaller, loss of detail
Relative sizes: depth
Texture gradient: depth
Aereal perspectives: loss of detail Ð color shift
Relative height: relationship to horizon
Interposition: whole object seen as in front
Linear perspective: depth
Cognitive Model: this model can be applied to characterize the cognitive abilities, skills and knowledge of representative users, so that such a model could serve as a basis to then validate the System Model of the Operator, it means, that the Survey Instrument Design really matches the sample of subjects who will participate answering the questionnaire.
Mental Model: this model could be used to assess how the task of answering a survey instrument online is being perceived and conceptualized by a sample of subjects willing to participate in answering it. It would be interesting to describe their perceptions of paper-based questionnaires, and then compare with their perceptions about answering it online. Hopefully, this model would influence the Conceptual Model.
Conceptual Model: this model would represent the procedures to be carried out by a participant who is willing to answer the questionnaire online. It should be described by using online help, FAQ, and descriptions of the project sponsoring the survey. It should be consistent with the Mental Model.
System Model of the Operator: this model would represent the expected behavior of the sample of subjects participating in the survey. It should also be consistent with the cognitive abilities, skills and knowledge of representative users, as defined with the Cognitive Model.
Interface Object Models: this model would represent with graphics or symbols the interaction objects which allow a subject to access, answer, edit, and submit the questionnaire in a transparent way.
Interface Model: this model would be composed of specifications of the purpose and basic functions of the online survey
A good HCI designer should be aware of the following issues when using or generating HCI models:
System power and user's proficiency will generally combine to determine the overall user performance, so that interface designers need to carefully consider these two aspects when designing and evaluating preliminary designs.
Recursiveness: "models model models", there is a theoretical concern when modeling users, tasks and processes, since the whole perspective can get complex and confusing.
Residence: "Is it the model of the user's knowledge of the task, or the system model of the user, or the user's perception of the system?", there is also a theoretical concern on who is modeling what, and how can we accurately represent abstract processes and complex tasks.
The system should be able to accomplish a specific functionality level when performing those tasks and procedures expected by the user.
Systems' conventions and mappings between actions and consequences should be the same within a questionnaire and, if possible, across questionnaires and other interviewer tools
The system should provide some feedback such as confirmation message or movement to next screen, for every user action
The system should carry out certain functions (s.a. error checking) without drawing user's attention to them
The system should make it obvious what actions are possible and how they are to be performed
The system should avoid jargon, abbreviations, and arbitrary conventions
The system should allow for errors, incorporating facilities to prevent, detect, and correct errors
The system should minimize user effort by, for example, simplifying actions needed to carry out common operations
The system should recognize the cognitive limits of the users and make it unnecessary for them to memorize large numbers of commands, providing ready access to online help instead
The system should avoid both over simplification and extreme complexity.
Users are uncomfortable if screens are too simple or too complex.
Contrary to "Whitespace is a virtue."
Take advantage of knowledge that is available externally, as well as knowledge stored in long-term memory
Simplify the structure of tasks. E.g., require numerical answer
Make both controls and the actions they perform visible to the user
E.g., help bars with reminders about which key to press to back up or to advance
E.g., display answer back to respondent, making the action of keying the answer more visible
Rely on natural mappings between actions and their consequences
Exploit both physical and cultural constraints
Allow for errors. E.g., allow users to back up and change answers
Rely on standardization when other design principles do not apply