Key Lines:
"We demonstrate that neural responses to persuasive messages can predict variability in behavior change in the subsequent week."
"our data suggests that the brain’s real-time response to persuasive communications may be able to complement self-report measures and provide additional predictive capacity."
The Experiment:
Sunscreen messaging to increased use of sunscreen.
Results:
a) 99.3% were >0, suggesting that the correlation between actual behavior change and predicted behavior change of independent subsets is significantly likely to be positive (α < 0.05, two-tailed)
b) This cross-validation analysis suggests that on average, approximately one quarter of the variability in behavior change scores in a new sample of 10 individuals can be inferred based on a regression equation constructed from the MPFC activity in the first 10 individuals, above and beyond their self-reported attitudes and intentions.
Sample size: 20
Recruited how?:
were recruited from the UCLA subject pool and through mass emails and posted fliers
Issues with this study:
the cross-validation predictive model.
the small sample size
the potential claim
Problems we need to solve:
Sam proposed the idea of doing a study with the student population.
Using a cross-validation model we may be able to predict and generalize our approach to the all of the student market. To predict to see what schools and decisions students make after seeing marketing material.
Citation:
Falk, E. B., Berkman, E. T., Mann, T., Harrison, B., & Lieberman, M. D. (2010). Predicting Persuasion-Induced Behavior Change from the Brain. Journal of Neuroscience, 30(25), 8421-8424. doi:10.1523/jneurosci.0063-10.2010
"We demonstrate that neural responses to persuasive messages can predict variability in behavior change in the subsequent week."
"our data suggests that the brain’s real-time response to persuasive communications may be able to complement self-report measures and provide additional predictive capacity."
- Neural responses can predict persuasive messages behavior change.
- ROI: MPFC = behavior change.
- "Neural signal can predict complex real world behavior days in advance."
- FMRI study.
The Experiment:
Sunscreen messaging to increased use of sunscreen.
Results:
a) 99.3% were >0, suggesting that the correlation between actual behavior change and predicted behavior change of independent subsets is significantly likely to be positive (α < 0.05, two-tailed)
b) This cross-validation analysis suggests that on average, approximately one quarter of the variability in behavior change scores in a new sample of 10 individuals can be inferred based on a regression equation constructed from the MPFC activity in the first 10 individuals, above and beyond their self-reported attitudes and intentions.
Sample size: 20
Recruited how?:
were recruited from the UCLA subject pool and through mass emails and posted fliers
Issues with this study:
the cross-validation predictive model.
the small sample size
the potential claim
Problems we need to solve:
- How do we increase the sample size for an FMRI study?
- How do we increase the speed? As they didn't have a chance to test the copy.
Sam proposed the idea of doing a study with the student population.
Using a cross-validation model we may be able to predict and generalize our approach to the all of the student market. To predict to see what schools and decisions students make after seeing marketing material.
Citation:
Falk, E. B., Berkman, E. T., Mann, T., Harrison, B., & Lieberman, M. D. (2010). Predicting Persuasion-Induced Behavior Change from the Brain. Journal of Neuroscience, 30(25), 8421-8424. doi:10.1523/jneurosci.0063-10.2010