Investigating the Impact of Usability on Energy Efficiency of Web-based Personal Health Records (DATASET)

This dataset contains data from usability evaluations performed by 4 experts and energy consumption measurements obtained on a host machine while performing representative tasks on 4 PHR web portals. The dataset consists of 10 variables: 8 usability variables and 2 energy consumption variables. The energy consumption was obtained from each PHR and the usability results from each expert. The data collection methodology was based on the completion of 20 tasks. The description of the tasks can be found in the paper. The experts rated a set of 8 heuristics on a Likert scale. The heuristics were based on those proposed by the author, Jakob Nielsen, and were specifically designed for PHRs. Each heuristic contained two questions as a checklist to guide the usability evaluation at the end of each task. Energy consumption data was obtained using an energy meter connected to the power supply and monitor of a host machine under test. The device, known as Energy Efficient Tester (EET), was developed by the ALARCOS group of the University of Castilla-La Mancha. During the execution of the tasks by one of the experts, the energy consumption of the above-mentioned components was measured. As a result, instantaneous consumption measurements were obtained. The measurements were repeated 5 times to calculate the standard deviation in each iteration. These standard deviation values were averaged over the 5 iterations, obtaining a representative energy consumption value of each power source and monitor when performing each task.

In the correlations tab the correlations between the energy consumption measures and the average scores of the usability evaluations can be seen. Significant correlations are marked in green and those close to being significant are marked in orange. In the t-student tab, the statistical analysis was performed to see if there were significant differences between the energy consumption and usability evaluation sets of data. For the usability values, a threshold was determined, which distinguished when a task was considered to have excellent or fair usability. Those calculations where significant differences were found, are marked in colour.