New study on the long-term value of liberal arts education

[Wahoo Redux]

Lawyers might still be lawyers but the ROI of law school has become appallingly low for many. Automation is also wreaking havoc with equity and bond trading, as well as numerous other professions. I don't need a human sales clerk at Filene's or Bloomingdale's to help me select a shirt I can buy off the internet. And I wonder if the number of librarian positions at universities has held steady now that people aren't combing through physical books and journals like they used to.

Repeatedly we've said the technology has had an effect.

But we still trade stocks, we still buy shirts, we still have librarians. 

The culture is pretty much the same only infused with technology.

[ciao_yall]

Lawyers might still be lawyers but the ROI of law school has become appallingly low for many. Automation is also wreaking havoc with equity and bond trading, as well as numerous other professions. I don't need a human sales clerk at Filene's or Bloomingdale's to help me select a shirt I can buy off the internet. And I wonder if the number of librarian positions at universities has held steady now that people aren't combing through physical books and journals like they used to.

Repeatedly we've said the technology has had an effect.

But we still trade stocks, we still buy shirts, we still have librarians. 

The culture is pretty much the same only infused with technology.

Law - though databases are automated, that really affected paralegals other service providers.

Stocks and bonds - trading is automated and the jobs are really in programming, not so much writing slips and running around the trading floor.

Shopping, travel and other fields - customers have become their own service workers. Again, jobs are in programming.

There is a barbell effect as industries mature. Good example is retail - The middle is squeezed (Macy's, Sears) because the price conscious consumer will tolerate less service to save a few bucks (Wal-Mart, Target). The high-end customer who values unique merchandise and high quality will pay the margins to support Saks, Nordstrom, and others.

So firms cluster around the low-end and high-end. Banking, gyms, salons, hotels, you-name-it.

Anyway, the same seems to be happening with jobs. For people with low skills, we'll always need janitors and medical assistants. For people with high skills, we will always need doctors and computer programmers.

For the middle, it's more complicated. Think of the "3 D's" - if there is an aspect of a job that is Dirty, Dangerous, or Dull, it is at high risk of being automated.

[ciao_yall]

Lawyers might still be lawyers but the ROI of law school has become appallingly low for many. Automation is also wreaking havoc with equity and bond trading, as well as numerous other professions. I don't need a human sales clerk at Filene's or Bloomingdale's to help me select a shirt I can buy off the internet. And I wonder if the number of librarian positions at universities has held steady now that people aren't combing through physical books and journals like they used to.

That said, to get you to buy a shirt online there is a lot of consumer psychology and branding strategy. So while a sales clerk, retail manager and merchandiser isn't available, these are being replaced by a marketeer and a computer programmer. On the low end, it creates jobs for warehouse workers and delivery drivers.

[Wahoo Redux]

The point about generations and the validity of the study, folks, is that while some things have changed regarding technology the basic skill sets taught in college are very much the same then as now and are still applicable now as they were then.  Careers and business (since that seems to be the measure of education) are still largely the same, just mediated by technology.  We had consumer psychology and advertising classes in the '80s which got people ready to go into advertising, and while the term "branding" was not used a lot, it was still very much a thing produced by advertising agencies.  Look at any magazine from the '70s or '80s for verification.  We can be pedantic about the minutia all we like.

[marshwiggle]

Here's a link to the original Georgetown study.

An odd thing is that it's comparing institutions rather than programs, as evidenced by some of the sections of the report.

Here are some of the headings in the report:

• The ROIs at the most selective liberal arts institutions are very high. (emphasis added)
• High graduation rates are associated with high ROIs at liberal arts colleges.
• Liberal arts institutions with smaller shares of low-income students have higher ROIs.
• ROI is high for institutions that have a high share of STEM majors.
• Geography is an important factor in ROI. (It further states that "Colleges in regions where per capita earnings are greater have higher median ROIs than colleges in regions where per capita incomes are lower.")

And in other news, water is wet.

[Caracal]

The point about generations and the validity of the study, folks, is that while some things have changed regarding technology the basic skill sets taught in college are very much the same then as now and are still applicable now as they were then.  Careers and business (since that seems to be the measure of education) are still largely the same, just mediated by technology.  We had consumer psychology and advertising classes in the '80s which got people ready to go into advertising, and while the term "branding" was not used a lot, it was still very much a thing produced by advertising agencies.  Look at any magazine from the '70s or '80s for verification.  We can be pedantic about the minutia all we like.

And I think the broader point is that more information being readily available has actually made being able to analyze and understand that information ever more important. My job as a historian is not to know what happened in the Black Hawk War off the top of my head, any more than it is your job as a Chemist (or something else?) to know the molecular structure of Glutamine. That we can find these things without going to the library or pulling a book off our shelves is convenient, but it hasn't made us obsolete because our job isn't to know facts but to interpret and analyze. Has it changed what we do? Sure, I do kinds of research now would have been basically impossible 25 years ago, but the basic work of understanding and analyzing the data hasn't changed that much.

[Kron3007]

I can tell you for certain that doing a literature review now compared to when I was an undergrad is day and night.  In the 80s, running a simple regression would require punch cards and access to a central computer, now I can (and am expected to) do much more complex analyses from my office computer.  This alone has changes how we design experiments and look at data.  It is a pretty big shift in my world. 

My point exactly.  You do exactly the same things you used to, just speed and information access have become much better.  We simply have more and better tools than we had before.

Not only does college force people into using these tools and these skills, but the overall skill-set that was taught 40 years ago is still applicable today.

I think the answer to calculating the ROI on a college education is that we shouldn't, or at least not to this level.  I have the same argument when people say that college grads in general make X amount more than non-college grads.  Correlation does not mean causation, and people who choose (and are able) to go to college tend to have a completely different outlook than those who dont and may have earned more regardless of this education.  Obviously there is likely a benefit to going to college, but you cant use this data to quantify it.   

Firstly, education is not simply about financial return on investment.

Secondly, why can't we use this information to quantify the benefits of going to college?  I hear that business about cultural capital, personal choice, family connections, etc. a lot---yet proof of these factors is itself pretty hard to quantify unless we can look into an alternate dimension to see our exact doubles in a mirror world without their college degrees.   

Thirdly, the numbers indicate tremendous value to a college education, as with this study.

I completely agree that the ROI of education is not all monetary.  I personally think that one of the best reasons to go to university beyond personal fulfillment is to increase the chances of getting a job that you actually enjoy.  Hopefully it pays well too, but enjoying work is a huge factor IMO. 

As for why we cant use this information to quantify the economic benefits of college, I think I have outlined my main concerns but in brief it is because it is fundamentally flawed to draw these conclusions from the data at hand.  The data do not meet the assumptions needed to do a statistical analysis and doing so is prone to all sorts of false conclusions.  This is like when people show the correlation between the use of roundup and autism rates in America.  There is a strong correlation, but that does not mean there is a causal link.  You could just as easily show a correlation between organic food sales and autism rates...         

[Wahoo Redux]

I completely agree that the ROI of education is not all monetary.  I personally think that one of the best reasons to go to university beyond personal fulfillment is to increase the chances of getting a job that you actually enjoy.  Hopefully it pays well too, but enjoying work is a huge factor IMO. 

As for why we cant use this information to quantify the economic benefits of college, I think I have outlined my main concerns but in brief it is because it is fundamentally flawed to draw these conclusions from the data at hand.  The data do not meet the assumptions needed to do a statistical analysis and doing so is prone to all sorts of false conclusions.  This is like when people show the correlation between the use of roundup and autism rates in America.  There is a strong correlation, but that does not mean there is a causal link.  You could just as easily show a correlation between organic food sales and autism rates...       

The viability of the study is a different debate.  I have not studied it and I'm not in a data-quantifying field, but someone who is presumably knowledgeable legitimized the study.  Could you be more specific as to why the data is inadequate? 

And while I totally agree that a job one enjoys is one of the prime benefits of college training (what I did with my life), I was hoping to move the discussion away from the idea that employment is the only reason for college and not the only way to evaluate college education.

[Wahoo Redux]

The point about generations and the validity of the study, folks, is that while some things have changed regarding technology the basic skill sets taught in college are very much the same then as now and are still applicable now as they were then.  Careers and business (since that seems to be the measure of education) are still largely the same, just mediated by technology.  We had consumer psychology and advertising classes in the '80s which got people ready to go into advertising, and while the term "branding" was not used a lot, it was still very much a thing produced by advertising agencies.  Look at any magazine from the '70s or '80s for verification.  We can be pedantic about the minutia all we like.

And I think the broader point is that more information being readily available has actually made being able to analyze and understand that information ever more important. My job as a historian is not to know what happened in the Black Hawk War off the top of my head, any more than it is your job as a Chemist (or something else?) to know the molecular structure of Glutamine. That we can find these things without going to the library or pulling a book off our shelves is convenient, but it hasn't made us obsolete because our job isn't to know facts but to interpret and analyze. Has it changed what we do? Sure, I do kinds of research now would have been basically impossible 25 years ago, but the basic work of understanding and analyzing the data hasn't changed that much.

Agreed.

But again, what we do hasn't really changed, we simply have better tools. 

[Kron3007]

I completely agree that the ROI of education is not all monetary.  I personally think that one of the best reasons to go to university beyond personal fulfillment is to increase the chances of getting a job that you actually enjoy.  Hopefully it pays well too, but enjoying work is a huge factor IMO. 

As for why we cant use this information to quantify the economic benefits of college, I think I have outlined my main concerns but in brief it is because it is fundamentally flawed to draw these conclusions from the data at hand.  The data do not meet the assumptions needed to do a statistical analysis and doing so is prone to all sorts of false conclusions.  This is like when people show the correlation between the use of roundup and autism rates in America.  There is a strong correlation, but that does not mean there is a causal link.  You could just as easily show a correlation between organic food sales and autism rates...       

The viability of the study is a different debate.  I have not studied it and I'm not in a data-quantifying field, but someone who is presumably knowledgeable legitimized the study.  Could you be more specific as to why the data is inadequate? 

And while I totally agree that a job one enjoys is one of the prime benefits of college training (what I did with my life), I was hoping to move the discussion away from the idea that employment is the only reason for college and not the only way to evaluate college education.

The reason the data is not sound is that it is not random.  To draw this conclusion, the students would need to be randomly assigned to different types of education and then we could evaluate the outcomes.  As mentioned, people who choose a liberal arts education are a subset of the population with a different outlook on life, education, and likely socioeconomic background.  It is quite likely that this subset of people would have earned more than most regardless of the degree/program they enrolled in due to various factors   This alone makes the conclusions, as presented, flawed.  Often, papers like this address these limitations in the discussion, but this gets lost when the media summarizes the findings and in the end it leads to false conclusions, or at least overstates the evidence.

A good analogy is in the medical world.  If you do a survey study that finds people who take a specific drug have lower incidence of a disease, it suggests that this drug may have an effect but people who take that drug may also do a lot of things that are contributing (diet, excersize, etc).  Before you could ever legally make such a claim for a drug, you would need to do a randomized (and blinded) clinical trial for this exact reason.

Obviously this is impossible in the case, bit it doesn't change the fact that we cannot really draw this conclusion based on this information.  This isn't to say there is not a benefit of a liberal arts education, just that we cannot really make such a strong claim, and especially put a dollar value on it.

[spork]

[. . .]

A good analogy is in the medical world.  If you do a survey study that finds people who take a specific drug have lower incidence of a disease, it suggests that this drug may have an effect but people who take that drug may also do a lot of things that are contributing (diet, excersize, etc).  Before you could ever legally make such a claim for a drug, you would need to do a randomized (and blinded) clinical trial for this exact reason.

Obviously this is impossible in the case, bit it doesn't change the fact that we cannot really draw this conclusion based on this information.  This isn't to say there is not a benefit of a liberal arts education, just that we cannot really make such a strong claim, and especially put a dollar value on it.

I'll make a related, even more specific analogy: a survey study indicates a correlation between a new, very expensive drug and health outcome. Yet the study fails to account for the fact that people wealthy enough to afford the drug (e.g., employed with health insurance benefits) generally receive better medical care and overall are healthier than average.

[marshwiggle]

I'll make a related, even more specific analogy: a survey study indicates a correlation between a new, very expensive drug and health outcome. Yet the study fails to account for the fact that people wealthy enough to afford the drug (e.g., employed with health insurance benefits) generally receive better medical care and overall are healthier than average.

To summarize the findings from the report:

Here are some of the headings in the report:

• The ROIs at the most selective liberal arts institutions are very high. (emphasis added)
• High graduation rates are associated with high ROIs at liberal arts colleges.
• Liberal arts institutions with smaller shares of low-income students have higher ROIs.
• ROI is high for institutions that have a high share of STEM majors.
• Geography is an important factor in ROI. (It further states that "Colleges in regions where per capita earnings are greater have higher median ROIs than colleges in regions where per capita incomes are lower.")

I strongly urge eye protection before proceeding to my summary of the report's findings.


Summary:
High income students in areas with high income going to selective liberal arts colleges with a high percentage of STEM majors and high graduation rates have higher ROIs than low income students in areas of low income attending non-selective liberal arts colleges with a low percentage of STEM majors and low graduation rates.

I sincerley hope the eye protection prevented damage due to the blinding flash of the obvious.

[ciao_yall]

[. . .]

A good analogy is in the medical world.  If you do a survey study that finds people who take a specific drug have lower incidence of a disease, it suggests that this drug may have an effect but people who take that drug may also do a lot of things that are contributing (diet, excersize, etc).  Before you could ever legally make such a claim for a drug, you would need to do a randomized (and blinded) clinical trial for this exact reason.

Obviously this is impossible in the case, bit it doesn't change the fact that we cannot really draw this conclusion based on this information.  This isn't to say there is not a benefit of a liberal arts education, just that we cannot really make such a strong claim, and especially put a dollar value on it.

I'll make a related, even more specific analogy: a survey study indicates a correlation between a new, very expensive drug and health outcome. Yet the study fails to account for the fact that people wealthy enough to afford the drug (e.g., employed with health insurance benefits) generally receive better medical care and overall are healthier than average.

So is the answer to improve equity by denying everyone a liberal arts education? Don't raise the bridge, lower the water...

[spork]

I never stated or implied that.

[Wahoo Redux]

To draw this conclusion, the students would need to be randomly assigned to different types of education and then we could evaluate the outcomes [...] It is quite likely that this subset of people would have earned more than most regardless of the degree/program they enrolled in due to various factors   

I hear what you are saying.  However, one cannot do the kind of study you are suggesting.  As I posted earlier, we could fashion a control group if we had a mirror universe with which to experiment.

I also keep seeing the "It is quite likely" that this group of students would do about as well no matter what they studied in college----but this is a pretty unproven assertion. 

What we do have are numbers that indicate that the worst charges against a liberal arts education are simply untrue.

That's what this study shows us, if nothing else.

There is an odd resistance to any good news or factual correction regarding the lib arts.