Usability evaluation techniques

A field study gives us insight into how people think and act and reveals aspects of their behavior that would be hard or even impossible to watch otherwise. A field study is needed in order to find critical details and understand working relationships, usually in large and complex organizations. A good example of this would be an Intranet of a large organization that may involve people, processes and software that is unknown to the system designers. With the field study, the designers should be able to understand the social, technical and organizational pressures within the company as well as the individual needs of the system’s stakeholders.  The outcome of the evaluation will be mostly qualitative results. A field study may provide the designers with a lot of information about the use of the system but is also time consuming. In addition, the results might not be accurate, since workers will be expected to modify their behavior if they know they are being watched.  
While field study takes place in the beginning, formative evaluation can be used in every stage of the lifecycle. Formative evaluation aims to expose usability problems during the iterative design cycle. To identify whether the interface prototype fits the requirements of the users, heuristic evaluation and other methods can assist the process. Formative evaluation increases user involvement and it is used to identify problematic areas really early in the design process. In comparison with other methods, undertaking a formative evaluation may require a lot of careful planning and preparation.
Contrary to a field study, a controlled experiment involves strictly controlled environmental settings were users are observed and timed while performing specifically given tasks. The outcome of the experiment is quantitative data that is used to calculate performance and identify errors. Experiments have a large number of participants and are performed under strongly controlled conditions. In comparison with field studies, data from controlled experiments are validated statistically. User feedback is collected through questionnaires and interviews while field studies collect data from observation and interviews. Emphasis is given in measurable information like the time needed to complete a task and the number and type of errors).  Due to their nature, controlled experiments are used in cases where we need to assess corresponding quality criteria and test quality assurance.
Predictive evaluation, on the other hand, uses an engineering model to predict usability. The advantages of predictive evaluation are that it doesn’t require the participation of real users or a prototype. Design alternatives can be compared without the implementation of a working prototype. Predictive evaluation is also more efficient in understanding the nature of usability problems because it breaks down user tasks in smaller sub-tasks and thus, makes it easier to identify where the problem lies. Predictive evaluation can be used to measure results in comparison with another method.
An expert evaluation provides for a quick identification of usability problems. Expert users undertake the job of completing specific tasks and while they are more likely to identify major problems with ease, they may miss minor problems because of their expertise. All in all, expert users may not represent the typical user of the system.  Typically, we would use expert evaluation to identify major problems quickly and easily.