27.03.2014 Views

SEKE 2012 Proceedings - Knowledge Systems Institute

SEKE 2012 Proceedings - Knowledge Systems Institute

SEKE 2012 Proceedings - Knowledge Systems Institute

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

inspection and in usability. Table III shows the data obtained<br />

from Characterization Form.<br />

B. Study Execution<br />

The inspection phase was carried out with each subject<br />

individually. They were provided with the instruments to<br />

accomplish the inspection and received instructions about the<br />

evaluation by the moderator. Once the inspector understood the<br />

procedures, the inspection process began. One researcher acted<br />

as the observer, being responsible for conducting the detection<br />

phase. It is worth to mention that the subjects did not receive<br />

training on WE-QT technique, usability or inspections.<br />

At the end of the inspection phase, a meeting attended by<br />

the researchers and a control group formed by usability<br />

specialists took place. A list of all usability problems identified<br />

by the subjects was discriminated to classify these problems<br />

into real defects or false positives. The authors of the technique<br />

did not influence the discrimination.<br />

V. QUANTITATIVE DATA ANALYSIS<br />

As a result of the inspection, we identified a total of 85<br />

usability defects. We computed the number of detected defects,<br />

time spent during the inspection phase, efficiency and efficacy<br />

for each inspector. Table III presents these results including<br />

and their experience level.<br />

N°<br />

TABLE III.<br />

Usability<br />

Experience<br />

SUMMARY OF THE INSPECTION RESULTS PER SUBJECT<br />

Inspection<br />

Experience<br />

Defects<br />

Time<br />

(hours)<br />

Defects<br />

/Hours<br />

% Founds<br />

Defects<br />

01 None None 25 1,85 13,51 29,41<br />

02 None None 48 0,77 62,61 56,47<br />

03 Low Low 26 0,67 39,00 30,59<br />

04 Low None 27 0,92 29,45 31,76<br />

05 Medium High 22 0,72 30,70 25,88<br />

06 Low Medium 21 0,87 24,23 24,71<br />

07 Medium Low 35 1,47 23,86 41,18<br />

Table IV shows the averages for the time, and effectiveness<br />

and efficiency indicators. Regarding the efficiency indicator,<br />

inspectors detected an average of 31.91 defects per hour using<br />

the WE-QT technique.<br />

Total Known<br />

Defects<br />

TABLE IV.<br />

DISCREPANCIES AND DETECTION TIME<br />

Effectiveness<br />

Average Indicator<br />

(%)<br />

Average Time<br />

(Hours)<br />

Efficience Average<br />

Indicator<br />

85 34,29% 1,04 31,91<br />

Regarding the effectiveness indicator, each inspector found<br />

an average of 34.29% of known defects. Considering the time<br />

of inspection, each inspector spent an average of 1.04 hours (or<br />

62 minutes) detecting the usability problems.<br />

As mentioned before, despite the goal of this study being<br />

eliciting the process used by the inspectors when applying our<br />

technique, the quantitative analysis results provide an important<br />

opportunity to evaluate the performance of the technique.<br />

Considering the effectiveness indicator in the last study [10],<br />

the WE-QT technique resulted in 29,37%, while the new<br />

version of WE-QT technique resulted in 34,29% (see Table<br />

IV). These results are an indication that the WE-QT v2 is more<br />

effective than its first version. Still considering this indicator,<br />

we compared our results to the results of a study presented in<br />

[6]. According to Gomes et al. [6], the effectiveness indicator<br />

of the WDP technique resulted in 13% and WDP-RT technique<br />

resulted 29%. This indicates that our technique is considerably<br />

more effective than WDP and lightly more effective than<br />

WDP-RT. As the number of usability defects depends on the<br />

application, it is not recommended to compare the efficiency<br />

indicator with previous studies.<br />

VI. QUALITATIVE DATA ANALYSIS<br />

We analyzed the inquisition data using coding procedures<br />

[12] in order to understand the subjects’ perception about the<br />

technique. We a lso merged the inquisition data with the<br />

observational data.<br />

At the end of this analysis, the coding processes produced<br />

altogether 21 codes which were associated to 3 categories:<br />

Positive Aspects of the technique, Negative Aspects of the<br />

technique and Improvements Suggestions. Although the<br />

study’s goal was to identify how the inspectors apply our<br />

technique, the Post Inspection Questionnaire had questions<br />

about the appropriateness and ease of use of the technique,<br />

which was useful to provide us with the subjects’ perception of<br />

the technique.<br />

Regarding the category “Positive aspects of the technique”,<br />

we identified codes as: “It is simple and easy to apply the<br />

technique”; “Using the technique contributes to detect a greater<br />

number of usability problems” and “I would use and<br />

recommend the WE-QT technique”.<br />

Concerning the category “Negative aspects of the<br />

technique” (Fig. 2) we identified the code “The questions have<br />

ambiguous items”, we observed that the subject had doubt with<br />

38% of the questions of the WE-QT technique. Another cited<br />

code is “There is no reference to pop-ups and some elements<br />

present in Web applications”, the observation data show that<br />

the technique also does not cover elements such as logins<br />

authentication, despite this element being evaluated as regular<br />

data forms, there is n o specific evaluation criteria for its<br />

features. The code “It was not trivial knowing what kind of<br />

answer to provide to the questions” was considered as a delay<br />

factor, and it also indicates that the technique requires more<br />

guidance. The code “Using the technique is a little tiring” was<br />

considered a harm factor to the inspector enjoyment when<br />

using our technique.<br />

591

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!