Idea:
Since I want to learn something about myself from this task, as well as exhibit my life for others, I decided to collect additional data that might answer a question I didn't know the answer to. I considered what quantifiable data would be readily available at any point in the day, but not under my conscious control. Counting how many people smile at me, for instance, could easily be affected by cues I gave them, and would be very dull for the private actions, to boot.

Thus, my task will answer the question: Does BetZ exhibit higher anxiety in public situations than in private situations? I will use pulse as a proxy for social stress, controlling for potential physical stressors exertion and food consumption.

Methodology:
Roughly every half hour, and during discrete actions, I will take my pulse at the radial artery over 15 seconds, and note the following variables:
Action in process (text)
Public or private (binary)
Level of exertion (text, possibly converted to a scale upon completion)
Number of minutes since last food or drink

During the study, I added variables to cover other potentially relevant factors, such as whether music was involved, whether music was slow or fast, and whether I was singing.

Data Collection:
There seems to be a social taboo against checking one's pulse, outside of the gym or doctor's office. It's also difficult to scribble notes on a paper in public without appearing crazy. Nonetheless, I was able to record most of my actions, with slight delay on some. My pulse-taking and watch-watching technique may have improved slightly over the course of the day.

Results:
Pulse was higher in public situations than private, even controlling for level of exertion; this effect is statistically significant at the 1%, 5%, and 10% levels in some regressions, but not provably different from zero in others, depending on the specification.  The number of minutes since undertaking heavy exertion was also important to control for;  using a logarithmic value of number of minutes since heavy exertion strengthened the effect and created a better fit for the entire regression model.

Listening to music, fast or slow, did not affect pulse, but singing did.  Regression 16 seems to be the most explanatory regression, taking into account the effects of being in public, various activity level indicators, the log of the time since heavy exertion, and whether I'm singing.  However, the next best fit is regression 18, which drops the original variable of interest (being in public or private), but keeps the singing dummy.  Regression 13, which does not include the singing indicator, but does include the public/private variable, is the next best fit, according to adjusted R-squared.

Being late did not raise pulse, after controlling for the exertion of running for a bus, etc.

Discussion:
Although the results are not conclusive, they suggest that celebrities and the homeless may be at risk for a heart attack related to high pulse, due to constantly being in the public sphere, coupled with likely cocaine use (not a variable covered in this study).  Recording artists and buskers may face additional risk because they engage in the high-pulse activity of singing, as well;  some should consider the risk management approach Ashlee Simpson and Milli Vanilli have so famously employed (i.e., not actually singing).  However, further study is needed.

Bibliography:
http://en.wikipedia.org/wiki/Pulse