35 lines
3.8 KiB
TeX
35 lines
3.8 KiB
TeX
\subsection{Questionnaire}
|
|
\label{questions}
|
|
|
|
\figref{results_questions} presents the questionnaire results for each visual hand augmentation.
|
|
Friedman tests indicated that all questions had statistically significant differences (\pinf{0.001}).
|
|
Pairwise Wilcoxon signed-rank tests with Holm-Bonferroni adjustment are reported (results are shown as mean $\pm$ standard deviation):
|
|
\begin{itemize}
|
|
\item \response{Difficulty}: \level{Occlusion} (\num{4.7 \pm 1.5}) was considered more difficult than \level{Contour} (\num{3.4 \pm 1.6}, \p{0.02}), \level{Skeleton} (\num{3.3 \pm 1.3}, \p{0.01}), and \level{Mesh} (\num{3.3 \pm 1.3}, \p{0.03}). No difference evidences were found for \level{None} (\num{4.6 \pm 1.9}) or \level{Tips} (\num{4.3 \pm 1.7}).
|
|
\item \response{Fatigue}: \level{None} (\num{4.9 \pm 1.3}) was found more fatiguing than \level{Mesh} (\num{3.6 \pm 1.3}, \p{0.04}); And \level{Occlusion} (\num{5.0 \pm 1.4}) more than \level{Skeleton} (\num{3.7 \pm 1.4}, \p{0.02}) and \level{Mesh} (\p{0.02}). No difference evidences were found for \level{Tips} (\num{4.5 \pm 1.6}) or \level{Contour} (\num{3.9 \pm 1.5}).
|
|
\item \response{Precision}: \level{None} (\num{2.8 \pm 1.8}) was considered less precise than \level{Skeleton} (\num{4.6 \pm 1.1}, \p{0.02}) and \level{Mesh} (\num{4.6 \pm 1.6}, \p{0.02}); And \level{Occlusion} (\num{2.8 \pm 1.7}) more than \level{Contour} (\num{4.6 \pm 1.8}, \p{0.02}), \level{Skeleton} (\p{0.006}), and \level{Mesh} (\p{0.02}). No difference evidences were found for \level{Tips} (\num{3.8 \pm 1.7}).
|
|
\item \response{Performance}: \level{Occlusion} (\num{3.0 \pm 1.5}) was lower than \level{Contour} (\num{4.9 \pm 1.6}, \p{0.02}), \level{Skeleton} (\num{4.9 \pm 1.4}, \p{0.006}), and \level{Mesh} (\num{4.6 \pm 1.4}, \p{0.03}). No difference evidences were found for \level{None} (\num{3.4 \pm 1.8}) or \level{Tips} (\num{4.2 \pm 1.6}).
|
|
\item \response{Efficiency}: \level{Occlusion} (\num{3.2 \pm 1.7}) was found less efficient than \level{Contour} (\num{4.9 \pm 1.7}, \p{0.01}), \level{Skeleton} (\num{4.9 \pm 1.4}, \p{0.02}), and \level{Mesh} (\num{4.8 \pm 1.3}, \p{0.02}). No difference evidences were found for \level{None} (\num{3.5 \pm 1.8}) or \level{Tips} (\num{4.3 \pm 1.5}).
|
|
\item \response{Rating}: \level{Occlusion} (\num{2.9 \pm 1.6}) was rated lower than \level{Contour} (\num{5.2 \pm 1.9}, \p{0.02}) and \level{Skeleton} (\num{4.9 \pm 1.9}, \p{0.03}). No difference evidences were found for \level{None} (\num{3.5 \pm 2.0}),\level{Tips} (\num{3.6 \pm 2.1}) or \level{Mesh} (\num{4.6 \pm 1.6}).
|
|
\end{itemize}
|
|
|
|
In summary, \level{Occlusion} was worse than \level{Skeleton} for all questions, and worse than \level{Contour} and \level{Mesh} on 5 over 6 questions.
|
|
Results of \response{Difficulty}, \response{Performance}, and \response{Precision} questions are consistent in that way.
|
|
Moreover, having no visible visual \factor{Hand} rendering was felt by users fatiguing and less precise than having one.
|
|
Surprisingly, no clear consensus was found on \response{Rating}.
|
|
Each visual hand augmentation, except for \level{Occlusion}, had simultaneously received the minimum and maximum possible notes.
|
|
|
|
\begin{subfigs}{results_questions}{Boxplots of the questionnaire results for each visual hand augmentation.}[
|
|
Pairwise Wilcoxon signed-rank tests with Holm-Bonferroni adjustment: ** is \pinf{0.01} and * is \pinf{0.05}.
|
|
Lower is better for \textbf{(a)} difficulty and \textbf{(b)} fatigue.
|
|
Higher is better for \textbf{(d)} performance, \textbf{(d)} precision, \textbf{(e)} efficiency, and \textbf{(f)} rating.
|
|
]
|
|
\subfig[0.32]{results/Question-Difficulty}
|
|
\subfig[0.32]{results/Question-Fatigue}
|
|
\subfig[0.32]{results/Question-Precision}
|
|
\par
|
|
\subfig[0.32]{results/Question-Performance}
|
|
\subfig[0.32]{results/Question-Efficiency}
|
|
\subfig[0.32]{results/Question-Rating}
|
|
\end{subfigs}
|