Graphical Tests for Power Comparison of Competing Designs

Area of Science:

• Data Visualization
• Statistical Inference
• Human-Computer Interaction

Background:

• Lineups serve as visual testing tools analogous to statistical inference tests.
• They offer objective evaluation of graphical findings with fewer distributional assumptions than traditional tests.
• Lineups are versatile for situations where standard statistical tests are not feasible.

Purpose of the Study:

• To introduce lineups as a method for evaluating the efficacy of competing graphical designs.
• To assess the theoretical properties of lineups in graphical design evaluation.
• To compare the performance of different graphical designs using objective measures.

Main Methods:

• Two simulation studies were conducted to evaluate graphical designs under controlled conditions.
• Amazon Mechanical Turk (MTurk) was utilized for participant recruitment and data collection.
• Accuracy and speed of evaluation were used as metrics to assess design effectiveness.

Main Results:

• Cartesian coordinates were found to be superior to polar coordinates for fast and accurate pattern spotting.
• The study demonstrated the effectiveness of lineups in objectively evaluating plot designs.
• Performance differences between designs were highlighted by pushing perceptual limits.

Conclusions:

• Lineups provide a robust framework for the objective assessment of graphical visualization designs.
• The choice of coordinate system significantly impacts the efficiency and accuracy of data interpretation.
• Further research can leverage lineups to optimize graphical representations for various data analysis tasks.