Conventional Wisdom and Cause vs Correlation

September 5, 2005

I recently listened to the audiobook “[Freakonomics](http://www.freakonomics.com/ “Freakonomics”)” by Steven D. Levitt and Stephen Dubner, and as a result I’ve been on a conventional-wisdom busting crusade. I was having a discussion with [Chris Cowan](http://www.uhduh.com/ “UHDUH”) about the importance of a college degree on predicting the likely success of a person in business. (Interestingly we had wildly differing opinions of success to begin with. But I digress.)

The debate was heated and somewhat surprising. I took a very risky position and after all was said and done, I turned to Steven Levitt for some perspective. The following are some excepts from an email exchange with the brains behind the book. (Italics = my email, bold + italics = his response).

… His position* was that having a college degree illustrates “an ability to make long term goals and complete them,” and that this indicates likely success. “The most important thing about it is that you are ending a milestone and it allows you to reflect upon your achievement,” he said. He used the “US census” and an extrapolated statement “that college grads earn more money then non-grads” as an indicator of success in that subset of the population. And he argued that it was good practice to eliminate candidates for hiring based on whether or not they had a degree.

My position was that the degree of success that one achieves in life is more likely the result of the factors in their life that make them more likely to go to college in the first place such as affluence, stable home life, etc., rather than the mere fact of them successfully navigating our severely lacking higher education system.

Also, I argued, the census includes people who have no means of going to college and have vastly different life paths than those people who are likely to make it through college, and thus skews the numbers. “It is possible that given a more specialized subset, one that was equal in every way accept college degree, may in fact prove vastly different results,” I expounded.

I am not discounting the fact that “conventional wisdom in our society does use a college degree as a primary (and sometimes sole) criteria for hiring,” but that it is “unfairly weighted.” The fact that some “underprivileged” bootstrap themselves into college and thus into a “degree only” job, could merely be a reflection of the determination and perseverance that would have made them successful despite a college degree.

Ok enough of the background, is there any basis for my argument? Or am I grasping as straws? I tried to find supporting documentation online, but I’m not having much success. Do you know of any study combatting the conventional wisdom here?

The evidence is strong that more education increases earnings by about 8 percent a year. That appears to be a causal impact, not just correlation. so if you take a random person and give them more education, their earnings tend to rise.

So that seems to argue against your position.

Note that the rise in earnings may not be because the workers are more productive, only that employers treat them as more productive. That gives your position some hope.

… You said “the rise in earnings may not be because the workers are more productive, only that employers treat them as more productive,” which is precisely my point. But it would seem that my position is nearly impossible to prove (or disprove), because the importance of the college degree is so ingrained in our society. As a result, it is possible that those who may have the drive and initiative to self-study in a particular higher-level skill, and become experts without the higher education system, may be dissuaded from even attempting that life path because they believe that the stigma of not having a college education will hold them back.

Given today’s increasing availability of information, it is becoming more likely that self-taught individuals (specifically in technology related fields) can be more skilled at graduation age than many college grads. Not proven by any means, but only possible.

So maybe there is more to college and graduation than just the degree … right? The big caveat in my argument is that the life lessons learned in college, such as independence and higher-level social problem solving, are immeasurable. These are frequently used in counterpoint arguments against me. But I’m sure there are very likely other life-lesson education scenarios that could occur in the absence of college life.

But the question remains “why do we give the college degree such an importance in our society?” As you stated “more education increases earnings by about 8 percent per year,” but is that a result of the actual impact of the person given their education or just our belief that their education somehow makes them a better and more productive employee? And, back to my first point, since we are dissuading all of the non-grads early on in life, is it possible that we manufactured the “cause” relationship through hundreds of years of conventional wisdom?

At the heart of my argument I do believe in the power of education, knowledge and applied information. I just don’t believe that having a piece of paper ensures proper application of said knowledge. There are a lot of educated idiots in the world…

So far it’s not looking to good for me, but I thought I would post it up here and see if anyone else has an opinion on it. If I get any response from Mr. Levitt I’ll post it.

6 Responses to “Conventional Wisdom and Cause vs Correlation”

  1. Christine Says:

    I am in a very difficult niche. Not only am I not college educated, but I am also female. So I am, I suppose, a statistical anomoly in myself. The story and reasons why college never happened for me are irrelevant. I will use myself as an example. I have wanted to know how to make websites pretty much from the first time I visited one. My opportunity to learn came in 1997 when I worked for a small company and updated the website. From then until the summer of 2000, I built HTML sites on Geocities. Then, in 2000, I was working as a volunteer on a tech support site and I wanted to help out. I was instructed to learn PHP, and I did so, using the owner’s existing code as a basis. The rest is history. My point on this is… if I went to college for 4 years, I would come out with a degree in hand, a big student loan debt, and no in-depth knowledge of any particular programming language. At which point I would have to learn it myself anyway. Isn’t it better than I used the same amount of time to specifically learn and use the particular things that I want to do in my career? And why should we discredit someone who didn’t make the “committment” to college when they are doing what they love? After all, wouldn’t someone who’s doing what they enjoy doing be more productive and have a high morale? And wouldn’t someone who is passionate about what they do be more likely to stay educated with new developments and changes in the industry and be self-motivated to constantly improve and learn?

  2. Chris Cowan Says:

    The lesson’s learned in college are more then just vocational knowledge. I agree in the IT world there are many many people who fall outside the norm. It’s almost common practice to hire people without formal education. When I was job hunting for my current position almost every ad required a four year degree in a IT related field.

    Even if you feel like a college degree is a total sham there are still many companies in our industry that require it. When it comes to job hunting it’s all about marketability. I’ve never been denied a job interview for having a college degree. On my resume it makes me more attractive. Even more attractive is that my degree is in something different then Computer Science which implies that I also have enough drive to be a successful self-taught developer. So I’m covering both the self-taught drive and the college grad. And when it comes to unemployment you need every advantage you can get to procure employment.


  3. I am not debating the fact that our society does in fact give great credibility to the college degree, thus making a candidate “more attractive” in the eyes of the majority of the population. But the question isn’t so much what we believe to be true, but what is in fact true. Are the things that we say about people with a college degree actually true? Are they more productive, better prepared, smarter, more determined … or do we just assume that about them because they have the degree?

    It is a fact that people with more education make more money. It is a fact that the majority of professional-level job postings state a college degree as a requirement. It is also a fact that hundreds of years of convetional wisdom can be proven wrong. Think of how we used to treat African-American people in this country.� It was conventional wisdom of those times that blacks we not human.� I think we all know how that turned out.


  4. The following is an excerpt from our chat on the topic.� Sorry for the bad writing … chat style you know…

    “Levitt is right; there is a definitive, causal relationship between education and earnings, proved over and over … so what?

    It’s about the knowledge. Regression only calculates a “regression to the mean” — a crude way of saying it’s a sophisticated averaging technique, holding all other things equal.

    That said, there are the tails of the “bell curve” distribution. Meaning, there are plenty of rich 10th grade educated people, and plenty of poor PHDs.

    Much comes down to productivity, desire, motivation, self-direction, passion, etc.

    Also, not related to education, but the Goleman work on emotional intelligence which suggests that EQ [seems to be] a stronger predictor of success than IQ, which again, is not education per se.

    Hence, don’t storm out of meetings pissed off and manage your emotions effectively in the moment; be empathetic; be aware of how your emotions drive you, etc.

    See Daniel Goleman: Emotional Intelligence (his first book).”

  5. Bryan Says:

    The difficulty here is that you are trying to measure two variables at the same time. One is the effect of a college degree and the other is the effect of the drive and determination it takes to achieve a college degree. Since, for the most part, these are found in the same individuals, it’s hard to determine whether they are treated differently than non-grads due to the status of the college degree or the underlying qualities that helped them achieve their degree. What you need are control groups.

    One possibility would be people who actually graduated college but for whatever reason no longer have a degree. Unfortunately, I doubt you will find people in that category. Nor will you find many people who will be willing to pretend they don’t have a degree to see how much less they will be paid during their lifetime.

    However, you might have a little bit more luck with the converse. People who have not gone through the effort of obtaining a real degree but pretend that they have one. Your theory predicts that they ought to achieve greater success since it is based on the status of graduating college rather than any real “drive and determination” related to completing college.

    There are a few public examples of people who have been caught doing this. I believe that some of the people were quite successful too. Unfortunately, you only find out about these cases when they are caught, so statistically you may not be able to draw conclusions from these individual cases. However they provide great anecdotal support for your case and are hard to argue against.


  6. Bryan, thanks for the insight. I think that proving or disproving my theory whould take so many years of study and the number of people you would have to convice to buck the convention in order to get a control group would be nearly impossible.

    The one thing I will say for sure tho (which should prove to me much more provable) is that applied knowledge coupled with determination/motivation are the building blocks of success in both groups. A college degree is a subjective indicator at best.


Leave a Reply