week 7 Flashcards
(32 cards)
Why is data important in HCI?
Because HCI is a natural science involving measurement and evaluation across safety, comfort, effectiveness, efficiency, and adaptability.
Why is data presentation important in HCI?
It allows researchers to recreate their findings inside someone else’s mind, making results interpretable and impactful.
What types of data are common in HCI?
Quantitative (numerical, measurable) and qualitative (descriptive, categorical), though quantitative is dominant.
What are examples of qualitative and quantitative data in HCI?
Qualitative: user opinions. Quantitative: neurophysiological signals.
In which HCI contexts are qualitative and quantitative data more balanced?
Comfort, Effectiveness, and Adaptability studies.
What are nominal data in HCI?
Categorical data without order—e.g., interaction modes, headset brands.
What are ordinal data in HCI?
Ordered categories where intervals aren’t necessarily equal—e.g., satisfaction levels, experience levels.
What are discrete numerical data in HCI?
Ratings or values that can be counted and averaged—e.g., cybersickness rating scales.
What are continuous numerical data in HCI?
Smooth, measurable data like cognitive load or pupil diameter that can be averaged.
What are ratio scales in HCI?
Normalized values that account for baseline variation, e.g., percentage accuracy improvements.
What are parametric statistical methods?
Statistical tools assuming normal distribution, e.g., t-test, ANOVA, Pearson correlation.
What are non-parametric statistical methods?
Methods that do not assume a normal distribution, e.g., Chi-square, permutation tests.
Why combine statistics with AI/ML in HCI?
To bridge understanding gaps, increase believability, and make results more accessible to varied audiences.
What is the traditional way to present data?
2D visualizations like bar charts, boxplots, and violin plots to help interpret numeric trends.
What are key differences between bar charts, boxplots, and violin plots?
Bar charts show means, boxplots show distributions and outliers, violin plots show full data distributions.
What is a state-of-the-art method for data presentation?
Immersive visualization using VR/AR and 3D models to recreate data from a first-person perspective.
What is an example of 3D data presentation?
The ‘glass brain’ model from UCSF, showing brain activity spatially and temporally.
What are the technical steps to present data in VR/AR?
3D modeling, integration into Unity/Unreal, rendering, and deployment on VR/AR platforms.
What is a violin plot?
A data visualization that shows the distribution of data across different levels using kernel density estimation, combining boxplot and density plot.
What do the shapes in a violin plot represent?
The width of each ‘violin’ indicates the frequency or density of data points at different values.
What is the state-of-the-art in data presentation?
Immersive 3D visualizations using tools like Unity, AR/VR, and real-time data rendering, exemplified by the Glass Brain.
What was the goal of the deep learning VR attention study?
To assess how VR versus 2D display affects attention levels using EEG measurements (specifically P3b component).
What type of study design was used in the VR attention study?
A within-subject design with counterbalanced display modes (VR and 2D) and randomized target/distractor conditions.
What were the independent variables in the VR study?
Display mode (VR vs 2D) and stimulus type (target vs distractor).
