This week newspapers have been covering how (some) black students at Duke are mad about a study done by a couple of professors. one of them an economist at Duke. This is a dangerous topic, because no matter what one says on this issue, someone is going to be offended. So, here I go!
Basically, a survey of Duke students' majors found that black students at Duke are more likely to switch to an "easier" major after they get there than white students are. And so, black students at Duke are mad. At what exactly, I haven't been able to divine from the reports. Most articles quote Nana Asante, president of the black Student Alliance: (from dukecheck.com):
“The implications and intentions of this research at the hands of our very own prestigious faculty, seemingly without a genuine concern for proactively furthering the well-being of the black community, (are) hurtful and alienating.” She asked the researchers and the university administration “what image has this … report portrayed to the rest of the country, namely our peer institutions, about Duke and its black students? Asante said the research paper does not “account for the societal, complex and institutional factors that must be considered in any attempt to delineate trends in racial differences in grade point averages and major choices, in a scholarly manner.
The implication is that because Duke and many other schools have affirmative action policies to give more minorities the opportunity to attend, that some may compensate by choosing less difficult majors at Duke. They found the same thing with respect to others who get more "opportunity" to get in with more leeway on scores, namely legacies (kids of alumni). It makes sense to me! I actually saw a few cases where it appeared to me that a legacy, athlete, or minority just wasn't as prepared for the Duke Juggernaut. Look, very few people are! Why are people mad?
It is well-known and commonly discussed that on average, Hispanics and blacks attend poorer schools, and might be less well prepared for college. Similarly, kids from rural areas (like me) went to school systems where we didn't have the greatest teachers, either, with a few exceptions, of course. For example, the calculus teacher in my high school did not understand even the basics of calculus. (I also had a French teacher in 9th grade who knew absolutely no French.) Yes, we had AP classes, but when discussing a poem, my AP English teacher told my good friend (who had the 3rd highest GPA in a large school):
Teacher: "In this line he is saying this."
Friend: "I think he could be saying that."
Teacher<enraged>: "You are not here to think. I am the only one paid to think here."
Additionally, I dealt with the same kind of anti-intellectualism you read about in poor, urban high schools. Having spent 6 years at Duke getting my Ph.D., let me tell you: My rural high school did not prepare me to survive at Duke, and I would not have survived had I gone there for Undergrad. I did a lot of growing and catching up in my 5 years at Appalachian State, and was therefore well-prepared for my Ph.D. work. Again, why are the students mad? I find this data very enlightening, and do not find it demeaning to anyone.
Years ago, I was concerned about a similar possibility: I wondered if it was possible that, if a student is let in to a school in some preferential way, that they might be possibly less prepared, and less likely to graduate-- it just stands to reason (some suggest this about George W. Bush's record). I am pleased to say, that most research has not found this to be the case for black students at most highly ranked universities. The currently-criticized study actually may help to partially explain why. I am intimately aware of how preparation, both academic and in terms of maturity and attitude, can affect chances of graduation. Also, any professor/advisor has seen students who move up to a more difficult or down to a less difficult major, oftentimes due to that student having high or low preparation compared to his peers.
If students or advocates want to get mad, do you:
Get mad at the data?
Get mad at the researchers for collecting the data?
Get mad at the researchers for writing the report?
Get mad at Duke for their affirmative action?
Well, please don't get mad at me! Yes, I get it... they got mad because they think it makes them "look bad"? Sorry, I disagree. And, it is juvenile to get mad at data. To kill the messenger is also juvenile. I personally know many of the Duke Economics faculty since I went there, but also because I have worked with them for years in a program for minorities who want to get their Ph.D. in economics. These people voluntarily worked for many years on this program, and had a great impact on a lot of very smart kids. These guys are not trying to make anyone look bad, they are uncovering facts.
What burns me up most is this: Any time someone is determined to get offended when data doesn't show that they are made of solid gold, you put a chilling effect on such research. It is already very difficult to report unpopular results on race, gender, sexual orientation, and the like. By yelling each and every time, you guarantee that fewer people will do research on your group, and we will understand less, and therefore be less able to uncover when there are genuine concerns. If you are outraged, I need to ask you one final, crucial question: "Having done the research, and found out something unpopular, do you really want people to shred the report at that point? Will that make you better off?" If you believe so, then you are an ostrich, and yes, should not be at an institution such as Duke. Please, please Duke students: Thank these researchers. Apologize to them. And then let them help you to think about solutions to any problems you think exist.
Monday, January 23, 2012
Friday, January 6, 2012
December's Unemployment Rates
Link: WFMY Interview on Youtube:
Well, things are finally looking up for the US economy. The unemployment rate (first estimate, they will revise it soon) for December was 8.5%, down from an estimated 8.7% in November, 2011. This, along with other good news about hiring and unemployment benefit filings all seem to confirm that we are making good progress toward recovery.
In this blog I want to briefly explain how unemployment data is collected and unemployment rates are calculated, since nearly every week someone in the media or government gets it wrong. I'll also discus some weaknesses of how it is measured, and why we should or shouldn't care.
The Bureau of labor Statistics in Washington, DC is in charge of surveying over 100,000 people each month in order to determine the unemployment rate. To simplify, suppose we want to estimate the unemployment rate for a small town. We would want to get a list of all of the households, and randomly select say, 100 of them. Suppose that a total of 200 people live in these households. The procedure for determining the unemployment rate is as follows:
First we ask all 200 people if they are at least 16 years old. Suppose 50 of our 200 are under 16; we throw them out of the survey.
Next, we ask the 150 remaining people "Did you work for pay last week?" The "for pay" eliminates housework, homework, and volunteer work. Suppose 90 of our 150 did.
Next, we ask the 60 (150-90) people remaining who said "No, I did not work for pay last week" one more question. We ask "Did you actively look for a job last week?" Actively means that you did something, such as call around, mail our your resume, fill out an application, or go into a business as ask if they are accepting applications. An example of something that is not active would be browsing through the "Help Wanted" ads in the newspaper, without doing anything. Heck, even I do that just for fun sometimes. Suppose 10 of our 60 said that they did actively look for work. The other 50 might be retired, full time students, house-husbands, or discouraged workers who have given up the search.
Our unemployment estimate comes from adding the 10 "active lookers" to the 90 employed people, to get 100 people in the labor market (working or actively seeking employment). 10 of the 100 are unemployed, for a 10 percent unemployment rate. This is currently a little higher than the 8.5% US rate and the 9.5% rate in my home state of North Carolina.
Some people don't like the method we use to calculate the unemployment rate, and I have my own problems with it. Some people don't like the fact that discouraged workers are not counted. Some people don;t like that the question asks "Did you work for pay", but not how many hours or for how much pay. I agree that this measure is not perfect, but as long as you know how it is calculated, then you know what it can and what it cannot tell us. [As a side note, we do collect data on underemployment (working PT when you want FT) and discouraged worker status]
The most important use is simply for us to be able to see if things are getting better or getting worse. And, thankfully, things are getting better. The US unemployment rate topped out at close to 10.1% two years ago. We still have a long way to go, but the trend is still slow, somewhat steady improvement, though there have been some bumps in the road. I am not an economic forecaster, but I am finally feeling more confident that we are on the right path. However, problems in the Euro zone could always come to bite us! Let's hope they can get themselves sorted out...
Additionally, the way we define the unemployment rate is pretty close to how most countries do it, so we can compare what is going on in the US to other countries. Just for fun, here are some other recent underemployment figures from other countries (from mid-to-late 2011, from various sources...)
Canada: 7.47%
France: 9.8%
S. Korea: 3.40%
Spain: 22.9%
Portugal 12.9%
Ireland: 14.1%
Brazil: 5.2%
Puerto Rico: 15.7%
Iceland: 7.1%
Japan: 4.5%
Well, things are finally looking up for the US economy. The unemployment rate (first estimate, they will revise it soon) for December was 8.5%, down from an estimated 8.7% in November, 2011. This, along with other good news about hiring and unemployment benefit filings all seem to confirm that we are making good progress toward recovery.
In this blog I want to briefly explain how unemployment data is collected and unemployment rates are calculated, since nearly every week someone in the media or government gets it wrong. I'll also discus some weaknesses of how it is measured, and why we should or shouldn't care.
The Bureau of labor Statistics in Washington, DC is in charge of surveying over 100,000 people each month in order to determine the unemployment rate. To simplify, suppose we want to estimate the unemployment rate for a small town. We would want to get a list of all of the households, and randomly select say, 100 of them. Suppose that a total of 200 people live in these households. The procedure for determining the unemployment rate is as follows:
First we ask all 200 people if they are at least 16 years old. Suppose 50 of our 200 are under 16; we throw them out of the survey.
Next, we ask the 150 remaining people "Did you work for pay last week?" The "for pay" eliminates housework, homework, and volunteer work. Suppose 90 of our 150 did.
Next, we ask the 60 (150-90) people remaining who said "No, I did not work for pay last week" one more question. We ask "Did you actively look for a job last week?" Actively means that you did something, such as call around, mail our your resume, fill out an application, or go into a business as ask if they are accepting applications. An example of something that is not active would be browsing through the "Help Wanted" ads in the newspaper, without doing anything. Heck, even I do that just for fun sometimes. Suppose 10 of our 60 said that they did actively look for work. The other 50 might be retired, full time students, house-husbands, or discouraged workers who have given up the search.
Our unemployment estimate comes from adding the 10 "active lookers" to the 90 employed people, to get 100 people in the labor market (working or actively seeking employment). 10 of the 100 are unemployed, for a 10 percent unemployment rate. This is currently a little higher than the 8.5% US rate and the 9.5% rate in my home state of North Carolina.
Some people don't like the method we use to calculate the unemployment rate, and I have my own problems with it. Some people don't like the fact that discouraged workers are not counted. Some people don;t like that the question asks "Did you work for pay", but not how many hours or for how much pay. I agree that this measure is not perfect, but as long as you know how it is calculated, then you know what it can and what it cannot tell us. [As a side note, we do collect data on underemployment (working PT when you want FT) and discouraged worker status]
The most important use is simply for us to be able to see if things are getting better or getting worse. And, thankfully, things are getting better. The US unemployment rate topped out at close to 10.1% two years ago. We still have a long way to go, but the trend is still slow, somewhat steady improvement, though there have been some bumps in the road. I am not an economic forecaster, but I am finally feeling more confident that we are on the right path. However, problems in the Euro zone could always come to bite us! Let's hope they can get themselves sorted out...
Additionally, the way we define the unemployment rate is pretty close to how most countries do it, so we can compare what is going on in the US to other countries. Just for fun, here are some other recent underemployment figures from other countries (from mid-to-late 2011, from various sources...)
Canada: 7.47%
France: 9.8%
S. Korea: 3.40%
Spain: 22.9%
Portugal 12.9%
Ireland: 14.1%
Brazil: 5.2%
Puerto Rico: 15.7%
Iceland: 7.1%
Japan: 4.5%
Friday, December 9, 2011
Burkey Reviews Stuff: The Trans-Siberian Orchestra Winter Tour, 2011 Is it just a bad Nativity story, or a bad rock concert? I can’t decide… Neither Trans-Siberian, nor an Orchestra… what is it?
Note: If you already love TSO or hate TSO, this review will
not help you. However, if you are like
me, and are vaguely familiar with them and considering going to a show, read
on.
I had very high hopes for the Trans-Siberian Orchestra show,
since I was familiar with a few songs, but not really a fan. Indeed, my wife and I used one of their songs
in our wedding. Before the show I
Googled information about it, only wanting to know the length so I could plan
for daycare and my bladder. In addition
to learning that the show was around 2 ½ hours with no break, I recall reading
a couple of comments by other reviewers:
1) Though it is partly a Christmas show, it does not leave
you feeling “Christmasey”
2) The Christmas part of the show has some sort of narration
in between the songs, which simply DOES
NOT WORK
I couldn’t agree more with these sentiments. Additionally, the show was tiresome and overall
not very enjoyable. While the musical
and voice talents were adequate, and a lot of effort was obviously put into the
show, it simply didn’t work. I have
thought long and hard about why this show left such a bad taste in my
mouth. My best answer is that every
single thing seemed to be overly contrived—melodramatic is the word -- just not
honest. The performers were working
hard—working WAY too hard, and most of the audience wasn’t quite able to buy
into it. Don’t get me wrong, there were
a few enjoyable spots, but these would quickly be overshadowed by something
inane. Let me provide some specific
examples.
The narrated story seemed overly dramatic, and the narrator
went overboard to use a low, breathy, shaky voice and exaggerated hand gestures
(think of a combination of Charlton Heston’s Moses and a bad TV preacher). Similarly, some of the singers used fakey, gravely
voices that didn’t work. In one “song”,
a guy dressed as a beggar pseudo sings an overly long story-more a spoken word
piece than a song. I couldn’t really
follow what he was rambling on about, but it felt like a sermon of some type
accompanied by a simple, repetitious guitar riff. I could feel a sense of relief from the
audience when it was finally over.
The general performance style of all of the band members
involved a similar melodramatic, attention-whoring style that was unnatural and
needy. One female singer tore at her hair as if recalling a recent genocide of
her people. She moved unnaturally-- the
musicians and singers in general tried to move in an exaggerated fashion as if
they were really into it. Or, they stuck
to contrived moves like well-timed head banging, throwing their TSO-required
long hair back and forth, or the electric violinist swinging his arm around in
circles over and over again, mimicking a juvenile electric guitar move. Look- I love it when musicians really get
into what they are doing, move and sway in weird ways, and even make funny faces, but it has to be real.
Ray Charles and Stevie Wonder do it, and you know that they didn’t learn
it from watching others—they are genuinely feeling it. With the TSO performance, I was led to
believe that this was all an over-choreographed act.
This show was what a precocious 12 year old, who is confused
about his musical identity, would design as his perfect show—throwing in a lot
of everything imaginable. Or perhaps, it
is what aliens who don’t understand humans would construct after carefully
reading about musical performances. Just
as a casserole does not benefit from adding a tablespoon of every spice known
to man, a show can actually benefit from subtlety. The TSO has laser lights, strobes, fog
machines, copious amounts of propane burning, fake snow, giant monster heads, a floating
three-axis stage above the crowd, silly uniforms, pretentious quotes from Reagan,
JFK, and Churchill, long blonde hair, and an overly-loud sound system. And, they actually do have some good music (I
am listening to it as I type this).
However, the show doesn’t work.
It is over thought and under felt.
If they took everything, and I mean EVERYTHING, and toned it down about
50%, I think the show could be much, much better.
I hate to be repetitive, so I will reiterate. ☺ The critical flaw for me was that the performers
made every effort to let us know that the show was about THEM, not about
US. They were jumping up and down on the
stage yelling “LOOK AT ME!!!!!” and pointing at themselves. Yes, literally POINTING at themselves. When you can’t get applause because your show
deserves it, you jump up and down pointing and telling people to yell. Also,
FOUR TIMES the band forced the audience to stand up. By forced, I mean that they asked the people in
the front row to stand, which causes a chain reaction. How impressive is it to beg for a standing
ovation… how about earning it? By the
fourth time, about 20% of the audience had had enough and just sat there
shaking their heads.
This quote from the TSO website sums up the kind of hubris
that created this monster. Actually,
created TWO monsters—since there are actually two versions of TSO touring
simultaneously (apparently the group I saw was the “original” one- I wonder if
the other group pulls it off better?).
When
Paul O'Neill first conceived Trans-Siberian Orchestra, his goal was as
straightforward as it was ambitious. "The whole idea," he explains,
"was to create a progressive rock band that would push the boundaries
further than any group before, following in the footsteps of Emerson, Lake
& Palmer, Pink Floyd, the Who...but take it way, way further."
Wow. Would a “real”
band put the following on their website?
With
more than 7 million albums sold TSO has inspired generations of fans to
rediscover the multi-dimensional art form of the rock opera. On the road,
meanwhile, they have become one of the world's top acts including a recent
mention in Billboard magazine as one of the top touring artists of the past
decade. With a $20 million-plus production that has played to over 7 million
people in 80+ cities, selling more than $280 million worth of tickets.
More “look at me!”, indeed.
Stop telling us how you are great, and just BE great! Now, when I started analyzing what was wrong
with the TSO show, I couldn’t help but compare it to the Blue Man Group. After all, the BMG has multiple groups as
well, and their show is certainly contrived and overdone. And, how can I possibly complain about the
overuse of stage props like propane when the BMG uses tons of toilet-paper-like
streamers? The first time I saw the BMG
in New York, I left laughing and crying, simply amazed. I didn’t want it to end. One of the Blue Men put a little blue paint
on my face and I wore it on the subway back to the flat where I was
staying. In contrast, I was begging for
the TSO pain to end about 45 minutes in.
What is the difference?
You see, the big difference is that the BMG is actually
making fun of TSO—The following two YouTube clips sum everything up beautifully.
Blue ManGroup (1 minute) (also watch the first minute or so of this one: BMGFloppy the Clown)
While TSO is religiously exploiting the stereotypes (even
Rock Concert Movement #37), BMG is lampooning them— all listed in their fictitious
“Rock Concert Instruction Manual”, outlined
here.
Hey, I know that performing is hard work and that it is easy to throw stones. All in all, I would recommend buying TSO’s albums, and of
course their show must be great for some people. I wanted to be wowed. But don’t go if you expect a Christmas show, or even real artistry. But, if you want a whiz-bang light show and the planetarium's Pink Floyd laser show is sold out, then this is a good bet. If you want to try a Christmas event that is a
little different and more than likely Free, how about going to see the Tuba Christmas in your local
area? A brass choir with only Tubas,
Euphoniums, Baritones, and the occasional Sousaphone. Oh, give it a try at
least on YouTube… Link 1 Link 2
Cheers-
-Dr. B
PS- I do understand that TSO is supposedly trying to be
socially conscious and give us a message.
Great. But the way they are doing
it is boring and incoherent. Here is a
clip where you can see what I am talking about in action (including blinding
the audience by pointing spotlights directly into their eyes, just as they did in the show I went to). (TSO Clip). If you want to deliver a message, there are better ways to do it (BMGClip).
Wednesday, November 16, 2011
The Case of Christmas Creep and Shopping on Thanksgiving
Reasons why it happens:
This is a general case of a "prisoner's dilemma" or "race to the bottom". Stores believe that if they don't do it, everyone else will, and they will lose out. They also believe that even if everyone else doesn't, they will win by being first. We are seeing the same kind of leap-frogging in the presidential primaries-- we all agree that this is the wrong outcome, but individual, selfish behavior makes it rational for individuals, but irrational for the collective result. By the way, consumers are just as much to blame- enough of us play the game and will show up whenever the stores tell us we might get a good deal and save $20-- even if that means freezing and getting sick at 3am rather than snoozing in your nice, warm bed. Not me, however-- but if you see a good deal, grab me one since you're already there. ☺
Additionally, since the Christmas season accounts for as much as 40% of yearly sales, retailers make sure to order their merchandise well ahead of time, so that there are no supply-chain mix-ups. I believe that this adds pressure as well. If you have the merchandise on hand, why not put it out?
Reasons why it is a bad idea
However, there are a few reasons why I think that this trend is a bad idea. First, I don't really believe that putting Christmas items out for twice as long will lead to twice as much spending. Plus, all of that Christmas merch takes up a lot of space-- there are lots of things that I'd like to buy that stores won't be carrying this week, because of all of the Christmas stuff they have out. A concrete example is that Walmart clears out most of their garden section for their Christmas items-- they do this so early now that you can't buy supplies for your Fall lime/seeding that one does in order to hope for a few blades of grass to grow the following year.
Secondly, as this NY Times article mentions, some retailers are thinking of opening just a little earlier for Black Friday-- Target, Khols, and others on Midnight Thanksgiving. Walmart has one-upped them by saying that they will open at 10PM on Thanksgiving. In some people's opinions, this is a bad idea. I agree! Why? Sadly, the best answer I can give is that it violates my sense of tradition. Christmas largely has its power in today's secular west because of "Christmas Traditions", after all. Black Friday has become a tradition, as well, and I firmly believe that forcing people out of stores, and giving an almost universal day off for Thanksgiving is an important part of the season. Pretty soon, Thanksgiving and Black Friday will both seem like "just another day"-- how long will it be before Christmas is just another workday, as well? (And then where will 40% of your revenues come from, eh Walmart?)
Cheers-
-Dr. B
PS: Another interesting issue I'd like to know about is, with Christmas goods out for so long, how do prices fluctuate during the season? Economists often discuss this issue in terms of the so-called "Durable Goods Problem". This phenomenon suggests that we all know that producers and sellers often charge higher prices for goods when they are first put out, but lower prices later if you wait. Think of iPhones, day after Christmas sales, the latest model of car, computer, or clothes. The simplistic version of the durable goods problem suggests that this won't work if everyone is rational, because we will ALL wait instead of paying the high price today. In reality, there are (at least) three types of people: Early adopters, "normal" people, and budget-conscious planners. So, some people get a thrill from blowing their money on the latest gadget, some buy it after the first price drop, and some people buy the used, older version from the early adopters on Craigslist! Long story short, I wonder if prices on Christmas goods follow this same pattern?
Monday, November 14, 2011
Income by Degree for Undergraduate Majors
Addendum to this article: I contacted the CEW about these concerns, and information they shared with me is in RED below.
The Wall Street Journal (WSJ) has a database of incomes by undergraduate major here. The WSJ got the data from the Center on Education and the Workforce (CEW) here, which they compiled from Federal data sources. I have a few of problems with this data as it is presented.
First, the numbers between the CEW and WSJ sites do not match up.
The WSJ Numbers were different because they were based on 2010 ACS Survey data rather than 2009 ACS data.
Second, the WSJ says that the data used is from the 2010 US Census, which has to be false. The Census Bureau did not collect any data on earnings or education in the 2010 Census. Yes, they used to, and their failure to collect this data in 2010 will be seen as a huge mistake in the future, because small area studies relying on Census data can no longer be done. By small I mean the Census Tract or Block Group Level. In any case, the study used the 2009 "American Community Survey", which is administered by the Census Bureau.
As above, they used 2010 ACS data, which should be carefully distinguished from 2010 Census Data.
Third, and most important, is that no effort is made to express the accuracy of these numbers. Let me demonstrate why this is so important. In order to demonstrate this, I am relying on combining two sources: The data at the CEW, and also a table from the ACS 2009 with summary statistics AND margins of error. Here is a link to it, but be forewarned it is pretty big. According to the 2009 ACS, 35,494,367 (+/- 120,221) people in the US have a 4 year degree as their highest education level, out of 201,952,383 people over 25 years old. Looking here at computers/math majors in the CEW report, they say that 7,829 (+/- not given) people have degrees in "Math and Computer Science" (I am assuming a double major?).
Here comes my point: The ACS in 2009 surveyed 1,917,748 households, and let us suppose that translates into around 3,000,000 people who are 25 and older (around 1.5 people over 25 per household). That means that surveys covered around 3 (million) out of every 202 (million) people, or 1.5%. If the ACS surveyed approximately 1.5% of the 7,829 Math and Computer Science majors, the quartiles for income given in the report are based on surveys of around .015*7,829≈117 people. Of course, we don't know how many people they actually surveyed with this major, but this seems small. How large might the standard error be for the $98,000 median? Here come some back-of-the-envelope calculations! Watch out for LOTS of assumptions!
To simplify things, let us assume that the distribution of salaries for the majors is a normal distribution-- of course this is not true, but this is a rough calculation, after all (and doing a better job requires having the raw data!). Then, the $75,000 and $134,000 1st and 3rd quartiles would be around 0.67 standard deviations above and below the mean. These two numbers are 23 and 36 thousand dollars away from the median, respectively (providing evidence of non-normality!), so lets guess that 0.67 standard deviations is in the neighborhood of 30,000, making 1 standard deviation around 30/0.67 ≈$45,000.
Givens:
Sample size around n≈117
s.d. around $45,000
Calculations:
standard error for a mean= 45,000/sqrt(117)≈$4,160
standard error for a median= 1.25*standard error for mean≈$5,200
95%confidence interval for the median earnings = $98,000+/-$10,192 (1.96*5,200)
Now, given that the distribution is not normally distributed, I would bet that the confidence interval would be even wider. I think that reports like this should give you a clue that some of their estimates have confidence intervals that are over $20,000 wide!
I was told that in a Future, updated report which will merge the 2009 and 2010 ACS data, the CEW plans to issue some guidance about standard errors in an appendix. I think that reporting sample sizes for each major would largely satisfy my concerns.
Now, to be fair, I am picking on one of the less frequent majors and highest variation majors in the tables... most of them will be better than this. However, these reports should still at least MENTION that this is sample data, and the numbers are only estimates, and that comparing numbers across majors might be unwise due to errors. Additionally, these survey forms are typically filled out by one person in the household for all household members-- this introduces other forms of survey error into the mix. Ask yourself-- how accurately do I know what my spouse makes in a year? How many thousands might you be off if I asked you on the spot? For that matter, how well do you know your own? Or, how accurately would you report your own income? (Copies of the Survey forms can be found here)
Cheers-
Dr. B
Photo Credit: digitalart / FreeDigitalPhotos.net
The Wall Street Journal (WSJ) has a database of incomes by undergraduate major here. The WSJ got the data from the Center on Education and the Workforce (CEW) here, which they compiled from Federal data sources. I have a few of problems with this data as it is presented.First, the numbers between the CEW and WSJ sites do not match up.
The WSJ Numbers were different because they were based on 2010 ACS Survey data rather than 2009 ACS data.
Second, the WSJ says that the data used is from the 2010 US Census, which has to be false. The Census Bureau did not collect any data on earnings or education in the 2010 Census. Yes, they used to, and their failure to collect this data in 2010 will be seen as a huge mistake in the future, because small area studies relying on Census data can no longer be done. By small I mean the Census Tract or Block Group Level. In any case, the study used the 2009 "American Community Survey", which is administered by the Census Bureau.
As above, they used 2010 ACS data, which should be carefully distinguished from 2010 Census Data.
Third, and most important, is that no effort is made to express the accuracy of these numbers. Let me demonstrate why this is so important. In order to demonstrate this, I am relying on combining two sources: The data at the CEW, and also a table from the ACS 2009 with summary statistics AND margins of error. Here is a link to it, but be forewarned it is pretty big. According to the 2009 ACS, 35,494,367 (+/- 120,221) people in the US have a 4 year degree as their highest education level, out of 201,952,383 people over 25 years old. Looking here at computers/math majors in the CEW report, they say that 7,829 (+/- not given) people have degrees in "Math and Computer Science" (I am assuming a double major?).
Here comes my point: The ACS in 2009 surveyed 1,917,748 households, and let us suppose that translates into around 3,000,000 people who are 25 and older (around 1.5 people over 25 per household). That means that surveys covered around 3 (million) out of every 202 (million) people, or 1.5%. If the ACS surveyed approximately 1.5% of the 7,829 Math and Computer Science majors, the quartiles for income given in the report are based on surveys of around .015*7,829≈117 people. Of course, we don't know how many people they actually surveyed with this major, but this seems small. How large might the standard error be for the $98,000 median? Here come some back-of-the-envelope calculations! Watch out for LOTS of assumptions!
To simplify things, let us assume that the distribution of salaries for the majors is a normal distribution-- of course this is not true, but this is a rough calculation, after all (and doing a better job requires having the raw data!). Then, the $75,000 and $134,000 1st and 3rd quartiles would be around 0.67 standard deviations above and below the mean. These two numbers are 23 and 36 thousand dollars away from the median, respectively (providing evidence of non-normality!), so lets guess that 0.67 standard deviations is in the neighborhood of 30,000, making 1 standard deviation around 30/0.67 ≈$45,000.
Givens:
Sample size around n≈117
s.d. around $45,000
Calculations:
standard error for a mean= 45,000/sqrt(117)≈$4,160
standard error for a median= 1.25*standard error for mean≈$5,200
95%confidence interval for the median earnings = $98,000+/-$10,192 (1.96*5,200)
Now, given that the distribution is not normally distributed, I would bet that the confidence interval would be even wider. I think that reports like this should give you a clue that some of their estimates have confidence intervals that are over $20,000 wide!
I was told that in a Future, updated report which will merge the 2009 and 2010 ACS data, the CEW plans to issue some guidance about standard errors in an appendix. I think that reporting sample sizes for each major would largely satisfy my concerns.
Now, to be fair, I am picking on one of the less frequent majors and highest variation majors in the tables... most of them will be better than this. However, these reports should still at least MENTION that this is sample data, and the numbers are only estimates, and that comparing numbers across majors might be unwise due to errors. Additionally, these survey forms are typically filled out by one person in the household for all household members-- this introduces other forms of survey error into the mix. Ask yourself-- how accurately do I know what my spouse makes in a year? How many thousands might you be off if I asked you on the spot? For that matter, how well do you know your own? Or, how accurately would you report your own income? (Copies of the Survey forms can be found here)
Cheers-
Dr. B
Photo Credit: digitalart / FreeDigitalPhotos.net
Subscribe to:
Posts (Atom)