Day 13: Bias in Data

Learning Objectives

• DAT-2.E: Explain how programs can be used to gain insight and knowledge from data.

Essential Questions

• How does bias occur in data collection and analysis?
• What are the consequences of biased data?
• How can we identify and mitigate bias in datasets?

Materials Needed

• Presentation slides on data bias
• Sample biased datasets
• Case study handouts
• Bias identification worksheet
• Computers for dataset analysis

Vocabulary

• Bias
• Sampling bias
• Selection bias
• Confirmation bias
• Representation
• Data equity
• Algorithmic bias
• Data cleaning
• Fairness

Procedure (50 minutes)

Opening (8 minutes)

1. Review and Connection (3 minutes)

   • Review data privacy concepts from previous lesson
   • Connect to today's focus on bias in data

2. Warm-up Activity (5 minutes)

   • Show students a clearly biased survey question (e.g., "Don't you agree that...")
   • Ask: "What's wrong with this question? How might it affect the data collected?"
   • Discuss how question design can influence results

Main Activities (32 minutes)

3. Lecture: How Bias Occurs in Data Collection and Analysis (12 minutes)

   • Define bias in data context: systematic errors that create unfair outcomes
   • Explain common sources of bias:
     • Sampling bias: non-representative selection of data points
     • Measurement bias: flawed data collection methods
     • Confirmation bias: seeking data that confirms existing beliefs
     • Exclusion bias: leaving out certain groups or variables
     • Historical bias: perpetuating past inequities in new data
   • Discuss consequences of biased data:
     • Reinforcing stereotypes and discrimination
     • Making incorrect predictions or recommendations
     • Leading to unfair resource allocation
     • Causing harm to underrepresented groups
   • Explain approaches to identifying and mitigating bias:
     • Diverse data collection methods
     • Representative sampling
     • Awareness of historical context
     • Testing for disparate outcomes
     • Transparency in methods and limitations

4. Case Studies: Examples of Biased Datasets and Their Impacts (10 minutes)

   • Present 2-3 real-world examples of bias in data and algorithms
     • Facial recognition accuracy disparities
     • Hiring algorithm bias
     • Medical research data gaps
     • Biased training data for machine learning
   • For each case, discuss:
     • How the bias occurred
     • What impacts it had
     • How it was or could be addressed

5. Activity: Identify Potential Sources of Bias in Sample Datasets (10 minutes)

   • Students work in small groups
   • Each group receives a dataset or description of a data collection method
   • Groups analyze and identify:
     • Potential sources of bias
     • Groups that might be underrepresented or misrepresented
     • How the bias might affect conclusions drawn from the data
     • Suggestions for improving the data collection or analysis
   • Groups document their findings

Closing (10 minutes)

6. Group Sharing and Discussion (5 minutes)

   • Groups briefly share their bias analyses
   • Discuss common themes and insights
   • Reflect on the challenge of creating truly unbiased datasets

7. Assessment and Preview (5 minutes)

   • Students complete a worksheet analyzing a dataset for potential biases
   • Preview that next class will begin the data project

Assessment

• Formative: Quality of bias identification during group activity
• Bias Analysis Worksheet: Thoroughness of bias identification and quality of improvement suggestions

Differentiation

For Advanced Students

• Analyze more complex datasets with subtle biases
• Research techniques for algorithmic fairness
• Explore statistical methods for detecting bias

For Struggling Students

• Provide datasets with more obvious biases
• Offer a structured template for bias analysis
• Use visual aids to illustrate bias concepts

Homework/Extension

• Find an example of potentially biased data in news media or research
• Design an unbiased data collection method for a given research question
• Research how a company or organization is addressing bias in their data

Teacher Notes

• Approach the topic with sensitivity, as discussions of bias can touch on personal experiences
• Use diverse examples that highlight different types of bias
• Make connections to broader social and ethical issues in computing
• Emphasize that identifying bias is not about assigning blame but improving data quality
• Consider discussing how awareness of bias relates to responsible data use