Rubywong123/UI_Human-Study
0
Welcome to Streamlit!
Edit /app.py to customize this app to your heart's desire. :heart:
If you have any questions, checkout our documentation and community forums.
Metric definitions
- Realism of Task:
- Definition: Whether the task resembles something a real user would encounter in everyday app usage.
- Example: "Search for a product and add it to the cart" is realistic; "Click random buttons" is not.
- State Reasonability:
- Definition: Whether the UI states and their transitions are reasonable given the app's typical structure and context.
- Example: A "checkout" button inside a map application is unreasonable.
- Action Validity:
- Definition: Whether each action logically corresponds to the intended next state.
- Example: Clicking “Submit” should lead to submission, not a homepage redirect.
- Logical Consistency (Thoughts):
- Definition: Whether any explanatory comments or inferred logic are coherent and free of contradictions.
- Example: “User clicks search to find item” followed by “user wants to delete profile” is inconsistent.
- Task Completion:
- Definition: Whether the trajectory ends with the task's goal fully achieved.
- Example: If the goal is “send a message,” is the message actually sent by the final step?
- Trajectory Consistency:
- Definition: Whether actions and transitions form a coherent flow, with no contradictions or unexpected diversions.
- Example: The trajectory should not flip between unrelated tasks or change context without reason.
- Effective Ratio:
- Definition: Ratio of meaningful steps to total steps; low ratio indicates inefficiency or redundant actions.
- Example: A 10-step trajectory with 7 relevant steps → 70% effective ratio.
- Topic Abstraction:
- Definition: Whether the task is generalized and meaningful, not just low-level UI manipulation.
- Example: “Complete login” is abstracted; “Click input, type name, click button” is not.
