1 of 23

2 of 23

3 of 23

Would You Eat This?

The Hidden Power of One Word

4 of 23

Response Bias

RESPOND TO THESE QUESTIONS IN YOUR SLIDESHOW

Definition of response bias

Why understanding response bias is important in surveys

5 of 23

Are people less likely to eat vegan pizza?

6 of 23

Unbiased

Surveys

7 of 23

Surveys

Biased

8 of 23

Hypothesis

Unbiased

Biased

9 of 23

Methodology & Participants

Volunteers surveyed in lobby

Coin flip →

  • Heads = unbiased survey
  • Tails = biased survey

Used my phone to enter their response.

Anonymous and I didn’t watch

10 of 23

Principles of the Experiment

Replication → 50 volunteers used for the survey

Random Assignment → Flipped a coin to see which survey

Control → Unbiased survey

Measurement → Recorded their responses on Google Form

Comparison → Compare % for each category

11 of 23

Results

Google Sheets

n = 50

(25 biased, 25 unbiased)

YES!

YES!

No.

No.

Vegan

Pizza

12 of 23

YES!

YES!

No.

No.

Results

Nearly twice would eat regular pizza (80%) compared to 48% who would eat “vegan pizza.

Vegan

Pizza

13 of 23

YES!

YES!

No.

No.

Results

Nearly twice would eat regular pizza (80%) compared to 48% who would eat “vegan pizza.

52% wouldn’t eat vegan pizza, nearly double the 20% for regular pizza.

Vegan

Pizza

14 of 23

YES!

YES!

No.

No.

Results

Nearly twice would eat regular pizza (80%) compared to 48% who would eat “vegan pizza.

52% wouldn’t eat vegan pizza, nearly double the 20% for regular pizza.

More people rejected vegan pizza.

Vegan

Pizza

15 of 23

YES!

YES!

No.

No.

Results

Nearly twice would eat regular pizza (80%) compared to 48% who would eat “vegan pizza.

52% wouldn’t eat vegan pizza, nearly double the 20% for regular pizza.

More people rejected vegan pizza.

More people said yes to the

regular pizza.

Vegan

Pizza

16 of 23

Simulation

% yes to pizza = % yes to vegan pizza

if “vegan” doesn’t matter

YES!

YES!

No.

No.

17 of 23

Simulation

% yes to pizza = % yes to vegan pizza

if “vegan” doesn’t matter

Could our difference of 32% (80%-48%) easily happen by chance?

YES!

YES!

No.

No.

Vegan

Pizza

32%

18 of 23

19 of 23

20 of 23

21 of 23

22 of 23

Conclusion

  • There was a 1.7% chance that this difference would have occurred by chance.
  • This is VERY unlikely.
  • It is likely that the survey design created response bias.

23 of 23

Wording matters!

A single word can change how people perceive a question.

Labels, like "vegan," carry different associations that can influence decisions.

Response bias can shape survey outcomes.

(important in marketing, politics, and social research)

Has vegan negativity changed over time?

Will it change in the future?