Health Benefits of Recreational Veganism: Empirical Evidence

Despite the growing body of evidence featured on this blog and elsewhere, and the near consensus of the recreationally vegan-economist community, there still remains a small body of skeptics who deny the positive effect of the vegan diet on human health. This post presents research conducted by the Recreational Vegan’s Quantitative Strategies Branch (QSB) in an attempt to further promulgate the scientific truth. This is done by quantifying the effect of veganism on health using the tools of econometrics.

A random sample consisting of 8 observations on individuals’ percent veganism and number of sick days in 2006 was collected. To counter a possible error in variables problem due to individuals reporting inaccurate measures of their veganism, a detailed schedule of veganism levels was supplied to survey respondents. As shown in the following table, there is substantial variation in the sample.

Percent Veganism Sick Days
Mean 61.8 3.3
Min 25 0
Max 99 7
Std. Dev. 32.4 2.6

The following regression was performed using ordinary least squares

SickDaysi = β0 + β1 · PercentVegani + εi

where i indexes individuals. The results are shown in the table below. The R-squared is 0.76 and a White test does not reject the null of homoskedasticity, indicating that there are no obvious specification errors. It would be possible to include more variables, such as dummy variables for gender or parenthood. However, these factors are unlikely to be correlated with percent veganism and as a result leaving them out will not cause an omitted variable bias.

Coefficient Std. Error P-value
Constant 7.68 1.12 0.001
Percent Vegan -0.07 0.01 0.005

Veganism has the expected sign and is significant at the 1% level. The results suggest that increasing veganism by 10 percentage points will reduce the number of sick days per year by 0.7. This has clear policy implications, suggesting that employers should offer incentives to employees to increase their veganism levels, possibly by offering free vegan lunches.


8 Responses to Health Benefits of Recreational Veganism: Empirical Evidence

  1. Darren says:

    Allow me to offer myself up as an additional data point to augment your awesome 8 person sample.

    Veganism score: near 0 (Lunch on Friday: steak sandwich, cold shrimp, and french fries)
    Sick days in the last 1.5 years: 0

    Bye bye, statistically significant coefficient!


  2. vegetable95 says:


  3. vegetable95 says:

    As a more substantial rebuttal, I should point out that a fundamental principle of statistics is that inferences can be made about a population without collecting data on all of its members. QSB believes that our sample is sufficiently random and representative to perform well. The small sample size is not a serious concern. I direct you to

    for an example of similar work in the literature.

  4. Consider a workplace where nobody takes sick days. Therefore, I would suggest that it’s not that you have “0 sick days”, it’s more “N/A”. QED.

  5. Darren says:

    1. The individual reporting 99% veganism is not an outlier?
    2. I have in fact heard of drawing inferences about a population from a sample of its members, however I would submit to you that a sample size of 8, seven of which are no doubt self-professed vegans or vegan-sympathizers, does not constitute a representative sample. Out-of-sample problems will abound when trying to generalize this research.
    3. Have you conducted a specification search? One theoretically important variable which should have been included: The increased vitality, joy, and meaningfulness of private sector employment improves morale, which reduces sick days substantially compared to the ennui and lethargy endemic to the public sector 😉

  6. Darren says:

    “increased vitality, joy, and meaningfulness of private sector employment…”

    Did I just write that? I think I’m going to be sick.

    First sick day tomorrow!

  7. The Guerilla Economist says:

    To Whom It May Concern:

    Your quantitative analysis is incorrect to an absurd degree. Specifically:

    1. Not even my colleagues in Psychology would argue that a sample size of 8 observations would represent a population. Hence, your analysis suffers from small sample bias, and your estimates are grossly misrepresentative.

    2. Although you admit to possible errors in variables, you fail to overcome the potential correlation between your regressor (% vegan) and your error terms. An appropriate method to resolve this problem is, of course, instrumental variables estimation. Might I suggest that, as opposed to percentage veganism, you might measure an individual’s muscle mass-weight ratio. It is a well-known fact that, due to their diet, vegans are unable to accumulate any muscle mass, and remain remarkably thin and bony. Thus, such a measure may provide a possible instrument for percentage veganism, and may alleviate the problem of errors-in-variables (although I am skeptical this resolves your problems entirely – see point 3).

    3. Point 2 underscores a greater problem: percentage veganism may be correlated with error terms due to endogeneity, which in turn causes more estimation problems. In particular, vegans tend to be (as mentioned above) thin and bony, which may result in less resilient immune systems, which may result in an increased probability of illness (ceteris paribus), which may in turn result in an increased number of sick days (ceteris paribus). Another possibility is that, due to their vegan whimpishness, vegans may themselves be more adverse to pain (since they go so far as to boo-hoo animal slaughter for caloric intake), and thus, ceteris paribus, vegans may be more likely to take a sick day relative to a carnivore. Unable to account for this leaves the skeptical reader’s qualms unsatisfied.

    In summary, this fleeting counteranalysis has proven the empirical study presented by this blog to be grossly misrepresentative.

    Again: watch thy step, Rec Veg,

  8. GE: I believe what you are saying, if I may clarify this for the general public, is



Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: