AnSci DataLab

Model-Dependency Bias: Netlogo & AI tutor

How it Works

  • Animals forage for food in two habitats with different nutritional levels.
  • If they don’t find food in time, they may die, with lighter animals at higher risk.

Your Task

You were hired to study the foraging behavior of animals. Run the simulation and analyze the system step by step, and answer using the following questions:

    Step 1: Start the simulation by clicking Setup and then Go. Then, examine only the Average Weight plot.

    Step 2: Activate the Total Population Weight plot in the Additional Information section. Restart the simulation by clicking Setup and then Go.

    Step 3: Activate the Number of Dead Animals plot in the Additional Information section. Restart the simulation by clicking Setup and then Go.

    Case Study:

    • A renowned scientist, Dr. Jekyll, collected data from these two habitats and concluded that Habitat A produces heavier animals, suggesting that this environment is more suitable for foraging.

    • Another famous scientist, Mr. Hyde, also observed this ecosystem and concluded that the animal population in Habitat B was heavier, possibly due to increased nutrient consumption. He hypothesized that Habitat B might be contributing to animal obesity.

    Step 4: Inactivate the button Repeatable-Run and try running the simulation at least 10 times, each time focusing on the Individual Average Weight plot.

    Final Reflection:

    Why it Matters

    This simulation highlights how data interpretation can be misleading when only one variable is explored. Different metrics may hide important aspects of the system. For example, analyzing averages may conceal disparities in the number of animals across different groups, potentially leading to survivor bias. It is recommended to study multiple metrics and perspectives to reach more accurate conclusions in animal systems like livestock production.

    Livestock Sustainability Group – University of Minnesota

    AnSci DataLab

    Copyright © 2025 ·

    AnSci DataLab

    · All rights reserved