CAPÍTULO II: MARCO TEÓRICO-CONCEPTUAL
2.1 ANTECEDENTES INVESTIGAVIVOS
2.2.25. Indicadores de Auditoría de Gestión
2.2.25.1. Mediciones Cualitativas
Before using the application in the hospital with patients, it is important to have as few usability and design errors as possible. The experiment described in section 7.5 can only be executed once; changing the application during the testing on patients could alter the results and is therefore not possible. Time wise, it is also sensible to prevent design errors to creep in to the
implementation, since fixing this after the implementation has been completed would take more time. The designs presented in the previous section are therefore tested on usability issues.
This usability testing is also in line with the iteration and prototyping paradigm suggested by Dix, Finlay, Abowd, & Beale (2004, p. 220). This checking of
(possibly paper-‐based) prototypes “whether they are acceptable and where there is room for improvement” can go on until there are “no more problems
that can economically be fixed”. For this project there is only time for one iteration of usability testing.
A problem with summative usability studies is that the system is judged based on predefined performance goals. Setting these goals beforehand is difficult and achieving the goals may not necessarily mean that the final system has high usability. Usability metrics “rely on measurements of very specific user actions in very specific situations”. In the early stages of design, the designer does not know yet which actions and situations there will be (Dix, Finlay, Abowd, & Beale, 2004, p. 241). This usability testing will therefore be formative, and focus more on qualitative information (e.g. users’ ideas about improvements) than
measuring e.g. efficiency, effectiveness and satisfaction levels.
The usability testing is done in a Wizard-‐of-‐Oz simulation where the designs of section 5.2 are used as a paper-‐prototype to simulate the application.
Participants of the usability study should generally be as close as possible to the target group of the application. However, it’s not possible to test the design on patients of the hospital, and therefore healthy people are used. Since the persuasion elements target hospital patients with their specific behaviour,
persuasion is not part of this usability testing. The target group of the application is very broad, so there are no specific conditions for the testing participants only that it is generalized and non-‐specific. The well-‐known article of Nielsen and Landauer (1993) shows that after five users the number of usability problems found decreases considerable. They also estimate that the optimal number of users for user testing in a small project would be seven. Although it has to be mentioned that this static number of users is controversial (Lewis, 2001), and some more difficult approaches have been proposed (Woolrych & Cockton, 2001).
First, all participants have to sign a consent form, which states that they give informed consent to the usability testing. The consent form also serves as a standardized way of explaining the usability test. Therefore it will give some background information about the research, where and how the application is going to be used and explains the procedure. The consent form is based on the template from the User Experience Group of Indiana University (2005) and can be found in appendix A.1.
Second, the participants in the usability testing do a walkthrough, which means that they go through the (simulated) system by doing some tasks. These tasks should of course not be too specific (e.g. press the green button) because that would contradict the goal of the testing, which is finding unclear or problematic design issues. The users’ actions are video recorded and users are asked to think aloud. Thinking aloud has the advantage of getting insight into the reasons why users do certain actions, and can be the basis for questions after the tasks (Dix, Finlay, Abowd, & Beale, 2004, p. 343-‐347).
There are three tasks the participants have to do. The time it takes the user to complete this task is measured. If the user cannot succeed, a hint will be given, but this is mentioned in the results. The three tasks are:
• 15 minutes before lunch you receive a notification (figure 5.1); open it
• After lunch you have to fill in what you ate and drank (figure 5.3). Select
something, answer the questions (figure 5.4) and go to the nutritional overview (figure 5.6).
• Go to the social section (figure 5.7) and visit Anna’s profile (figure 5.8).
Third, to obtain some general information and impressions of the user, a small questionnaire will be held. Usability questions have a five-‐point scale and two questions are open-‐ended. The questionnaire is based on the Standardized Computer System Usability Questionnaire at IBM (Lewis, 1995). This survey is also recommended by Lazar, Feng, & Hochheiser (2010) as one of the
recommended tools for usability testing and evaluation. However, seven questions (about completion of work, productiveness, error messages and documentation) not related to the application under observation were removed. The questionnaire also includes questions about the age, gender, experience with tablets and hospital visits of the user. The complete questionnaire can be found in appendix A.2.
As has been mentioned, the focus of this usability testing will be on qualitative information. The fourth and last part of the usability testing is therefore a semi-‐ structured interview with questions related to the observation. This interview is done after the questionnaire, to prevent biased responses (Brooke, 1996). The thinking aloud and other remarks made by the participants – in combination with this interview – will be used to help clarifying reasons why users behave or think a certain way (Dix, Finlay, Abowd, & Beale, 2004, p. 348).