← Back to course
2/6
Week 11 Β· Bias, Access and Many Voices

Tuesday

Where bias hides
// Whose voices are in our data, and whose are missing?
⏱ about 20 min

Tuesday: Where Bias Hides

Comet has a fix. "If we want fair data, let's just collect 40 more answers on the screen. More data, less bias!"

Wren tilts his head. "What do you notice about where those 40 would come from?"

Comet thinks it over. "The same screen. The same people who already use it."

"Then the same people would still be missing," Wren says.

Nova hovers between them. "Would you like a hint?" she asks. "Look at where the data come from, not only how much there is."

Comet sighs, then grins. "Okay. Where else could bias be hiding?"

Wren opens the crew's very first plan. "Let's check every step, data analyst."

More data is not the fix

Bias often comes from the type or source of the data collected.

Collecting more data does not remove that bias. Forty more screen answers would come from the same source.

Bias at every level

Computing innovations can reflect human biases, through their algorithms or through their data.

Bias can enter at every level of building something, and programmers should act to reduce it.

One kind of bias is a wrong assumption that developers make about their users.

Equity deficits include little exposure to computing, less access to education, and fewer training opportunities.

Crew step (made-up, from the Commons Count story)Hidden assumptionWho it could miss
Put the survey on the front-desk screenEvery member uses the screenMembers who walk past or come when the desk is closed
Run the survey for one weekThat week is a normal weekMembers who were away that week
Ask only in writingEveryone reads the survey easilyMembers who need the questions read aloud
Collect only while the desk is openMembers visit when the desk is openEvening and weekend members after hours
SPOT THE BIAS
  • Read the question.
  • Tap your answer.
Comet wants 40 more answers from the screen. Will that remove the bias?
The crew assumed every member uses the screen. What kind of bias is that?
Little exposure to computing is an example of what?
StatementTrue or false?
Collecting more data always removes bias.?
Bias can come from the source of the data.?
Bias can enter at every level of building something.?
Programmers should act to reduce bias.?
WHY THIS EXERCISEKnowing where bias comes from tells you where to look for it.
Bias often comes from the type or ____ of the data collected.
A wrong ____ about users can build bias into a tool.
HUNT FOR HIDDEN BIAS
  • ?List each step the crew took.
  • ?Write the assumption behind each step.
  • ?Ask who that assumption could leave out.
  • ?Plan a check or a fix for that group.
WHY THIS EXERCISEWalking step by step keeps a bias hunt from skipping any level.
On paper, add one new row to the table: a crew step, its hidden assumption and who it could miss.

Sharp detective work. Tomorrow in Count Lab you will run the bias audit with real numbers from the story.

← Monday