Combining Experimental Methods with Biometric Tools to Analyze Food-Related Behavior
Date
2018-08-05Metadata
Show full item recordAbstract
This dissertation uses experimental economics methods and biometric tools to test for the
consistency of individual preferences and analyze how such preferences are affected by states of
cognitive impairment and resource scarcity. Across the study, emphasis is made on the effect of
hunger on food choices and intertemporal decisions, along with the implementation of health-related
intervention programs tailored to individuals with different health characteristics. The
behavioral findings are supported by eye tracking data, which provide insightful information on
how visual attention and arousal impact final food choices. The methodologies used to measure
individual preferences are hypothetical and non-hypothetical, and the statistical tools used to
analyze this data include econometric models for categorical and limited dependent variables in
preference space and in willingness-to-pay (WTP) space.
The first essay tests the consistency of individual preferences over the same repeated choice
experiment. Results based on a within-subjects design indicate that after changing the position of
the same alternatives in the choice set, participants were consistent with their choices 69% of the
time. Moreover, after reverting back to the identical original positions of the alternatives but
randomizing the order of the choice sets, individuals’ choices were consistent 67% of the time.
The robustness of these results was further demonstrated by using random parameters models with
flexible mixing distributions to calculate WTP for the products attributes. Importantly, none of the
attributes followed a normal distribution, which highlights the importance of considering more
flexible forms such as polynomials when estimating the distribution of random parameters.
The second essay tests for the presence of an anticipatory food reward effect and examines
whether this effect is ubiquitous or if there are differential effects by body mass index (BMI). In a
controlled laboratory experiment, participants performed a cognitive test and a food choice task in
randomized order. The results showed that overweight and obese individuals exhibited an
anticipatory food reward effect, which enhanced their cognitive capacity after merely choosing a
food snack that would be consumed at the end of the experimental session. This cognitive
impairment induced by hunger only affected the food choices of obese individuals, who were more
likely to make unhealthy food choices. This finding was complemented by eye tracking data,
which indicated that the obese exhibited more arousal or engagement towards the food products
under a low cognitive capacity.
Finally, the third essay consists of a laboratory experiment implemented to investigate
whether inducing health related thoughts and future self-image representations influence the food
choices and intertemporal decisions of overweight, obese and normal weight individuals. The
results indicate that providing information about the immediate consequences associated with
healthy/unhealthy habits increased the number of healthy food choices and patience level of
overweight and obese individuals. However, when the obese interacted with their potential future
healthier and unhealthier selves, the opposite effect was uncovered. This effect might be due to
the fact that obese individuals look at the reward of becoming healthier as somewhat unattainable
in the short-run, requiring more tangible or plausible immediate rewards. The behavioral findings
of this essay are supported by eye tracking data, which revealed how the temptation towards
unhealthy food snacks exhibited by overweight, obese, and normal weight individuals translated
to their final food choices.
Citation
Segovia Coronel, Michelle Stefania (2018). Combining Experimental Methods with Biometric Tools to Analyze Food-Related Behavior. Doctoral dissertation, Texas A & M University. Available electronically from https : / /hdl .handle .net /1969 .1 /174105.