Why are there so few physics majors?

[ZapperZ]

Personal experience. If it's different for you, then maybe you know different people.



Those survey numbers are so wildly at variance with my person experience, that I question the validity. This is one of those, who do you believe some random survey or your own eyes? There are enough crap surveys out there that you have to go through a ton of convince to convince my that they have any validity, and even then if survey says X and I see Y, then there is still something to be explained.

I'm surprise to see such things coming from you. Are you telling us that you cannot tell the difference between ONE data point versus MANY data points? A survey tries to sample a lot of data points, the more the better. You survey ONE data point. And you're asking why your data point is different than a collection of MANY data point? You honestly cannot see the difference?

Why indeed? People are starting to ask that question, and those departments are getting their funding cut. Let me ask another question. Why to physics professors make more money then history professors, and why do business/finance professors make more money than physics professors?

Supply and demand, opportunities, and money generated!

History vs. physics prof: A physics professor tends to being in MORE external funding. A professor that can consistently obtain such external funding not only is highly prized at a university, he/she is also highly sought-after at other institutions! Not only that, depending on that person's area of specialization, he/she is also desirable at private companies. You want to keep such a person, you pay him/her more!

Physics versus business/finance: See last part of my argument above. A finance professor has more opportunities to go elsewhere.

Zz.

 

[twofish-quant]

I'm surprise to see such things coming from you. Are you telling us that you cannot tell the difference between ONE data point versus MANY data points? A survey tries to sample a lot of data points, the more the better.

No. It's not. My wife has a Ph.D. in education, so they have to deal with this all of the time, and I learn a lot from reading her books.

People aren't rocks so the methodology in dealing with people are different from those dealing with rocks. If you are dealing with a uniform sample and you are interested in a specific question that can easily be quantified (i.e. scores on a test), then you can statistically calculate how many samples you need to get a good sample, and doing more than that is a waste of time. The statistical error goes as the square root of the number of samples which means that once you have a large number of samples, then adding more isn't going to help, and if there is a non-sampling bias in the survey that's larger than sampling error you are spending a lot of effort for nothing.

Now if you are dealing with non-uniform samples, then statistics sometimes doesn't work, and then you have to go into interviews and other methods. The thing about interviews is that you can go really deep. For her Ph.D. dissertation, she ended up using interviews of four people, and I've known people that have done dissertations using *one* person as a sample. (What you can do with one person that you can't do with several hundred is to follow them around for two months and watch what they do.)

It's ***not*** statistically representative, but you are often dealing with in which you are interested in the deep details of one person rather than very shallow information about 1000. Also with interviews, you can do things like figure out someone's deep psychology, which is something that you can't do with a survey. The other thing with people is that everyone is different, which means that statistical representativeness isn't sometimes useful.

The other thing is to conduct a valid survey is amazingly tough. Asking the right question can be tough. For example, if you want to ask if college is too expensive it seems to make more sense to ask college freshmen parents than college freshmen.

And you're asking why your data point is different than a collection of MANY data point? You honestly cannot see the difference?

I can see the difference, and I don't see why I should trust my own eyes less than a group of people that I've never heard of. For that matter, one of the reasons I may be more useful than a survey done by people you've never met is that I'm here and you can ask me questions.

History vs. physics prof: A physics professor tends to being in MORE external funding.

Sure, but why are people willing to find physics professors more than history professors? I think it has something to do with the fact that physics profs can discover things with more obvious economic benefit than history profs.

Physics versus business/finance: See last part of my argument above. A finance professor has more opportunities to go elsewhere..

Right, it's all about the money.

 

[twofish-quant]

Also I've seen so many bogus college surveys that the amount of trust that I put into them is quite low. If you want to see what schools actual employment rate is like, don't trust a survey. Sit down with a few alumni and talk to them over lunch.

 

[ZapperZ]

Also I've seen so many bogus college surveys that the amount of trust that I put into them is quite low. If you want to see what schools actual employment rate is like, don't trust a survey. Sit down with a few alumni and talk to them over lunch.

And using that methodology, you should also believe in supernatural phenomena and other pseudosciences.

Zz.

 

[atyy]

Are there really people who consider college not a trade school in the broad sense - ie. it makes it more likely that one will have a job that pays reasonably? I personally know very few who thought that college was not a trade school, and they have all changed their minds from 20 years ago.

Not sure about the beer part though

Maybe relevant: Employment status of the civilian population 25 years and over by educational attainment

 

[twofish-quant]

And using that methodology, you should also believe in supernatural phenomena and other pseudosciences.

I don't see why. Note that a typical social scientist isn't interested in researching whether there are UFO's or not, they are interested in researching what people *believe* about UFO's, and the research methods for figuring out what people *believe* are quite different from the research methods for figuring physics problems. There are some interesting philosophical issues here, and if you really think that qualitative social science methodology leads to pseudoscience, put on your asbestos suit, and there are people that can argue the issue better than I can.

Just picking some books off my wife's bookshelf. Bogdan and Bilken's Qualitiative Research Methods for Education. Strauss and Corbin's Basics of Qualitative Research. If you believe that "anything not statistical is pseudoscience" you can argue with them.

(And if social science is pseudoscience, then why are we teaching it in universities? For that matter, universities often have theology and religion departments in which you have pretty smart people that *do* believe in the supernatural. Should we get rid of those?)

The other thing is that it's common in astrophysics to "forget statistics." A lot of what we believe about stars is based on one sample (the sun) and a lot of what we believe about supernova is also based on one sample (1987A). It would be nice if we had 100% data on all supernova and all stars, but we don't, and the fact that we really have no reason to think that neutrino emissions in the sun are typical of other stars doesn't mean that HR diagrams are "better." I use my life to figure out what is going on with other people in the same way that I use the sun to figure out other stars. Other stars are different, but they are just too far away for me to get good data.

Also one thing that I like about money is that it makes things more objective. One reason that I believe that there aren't many people willing to pay cash money for pure enjoyment of astrophysics is that if there *were* such a market demand, I'd quit my job tomorrow and become a free lance tutor of astrophysics. What I'd do is to take the courses that I've taught at the University of Phoenix or University of Texas at Austin, and if people were willing to give me the same money that they give UoP or UT Austin and I were able to spend the same amount of time doing it, then that would be more than enough to lead a comfortable life.

The problem with that plan is that I can't give them the piece of paper that they can use to get cash money, and if they can't get cash money, they can't pay me cash money, and the number of people that are willing to pay me cash money for astrophysics as a hobby isn't enough to make things work. One other data point is Fathom and similar efforts. Around 2000, a number of major universities started online initiatives to teach enrichment courses and all of these were shut down after massive losses. It seems that people won't pay money for these courses if there isn't money to be made. By contrast there are tons of online MBA's, of various quality.

There's also the experiment I did at UT Austin. I taught a one unit class, which a lot of seniors used to get the one credit they needed to graduate. So what happened was that on the first day of class, I told everyone that if they just wanted to take the class for the credit, this would be the minimum amount of work that they needed to do, and they could finish it all in one month. If they wanted to learn more stuff they could stick around after that one month. Typically, the enrollment numbers went from 200 to about 10 after that month passed. (Also UT Austin is still teaching the class, so you scan ask people there for current numbers.)

About reality, we can argue a lot about the nature of reality and evidence. What matters is when you go up to the edge of a cliff and are willing to jump onto a dry lake bed to avoid a herd of stampeding buffalo. I'm not because I think gravity exists, and if I do it, I will die. Similarly, if you could convince that most university students were interested in pure learning then there is water at the bottom of the cliff, and I'd quit my job tomorrow and go into free lance tutoring. (Seriously I would.) Now, personally, I think it would be economic disaster for me if I did this, but if you convince me that gravity doesn't exist, then I'll jump. (Also, I know of a number of Ph.D.'s that make a good living teaching physics free lance. They don't live in the US but rather work for cram schools in East Asia.)

Now, I'll jump eventually. If they problem is that people want to turn my tutoring into cash that they can pay me with, I suspect that someone will figure out how to do that in the next five to ten years. If the US goes to an East Asian type testing system, then everything blows open.

Once someone figures out the cash part, then then concrete at the bottom of the cliff becomes water, and there's a herd of stampeding buffalo running toward me, so at that point reality says to jump.

Now I'm crazy. Sometimes people think I'm delusional. Other times, people think I'm visionary. I think I'm both, but one thing I care a lot about is the nature of truth and reality. Because, if you are not seeing the truth, you end up dying, either figurative or literally.

I don't think we can keep doing things the way that we've always done then. We can't go back to the 1950's because in the 1950's, the internet didn't exist. I happen to think that at some point there will be water at the bottom of the cliff, and the buffalo are going to come, so I'm preparing to jump, just not now. If that's not your reality, then go with what you see, and we'll see what happens.

 

[twofish-quant]

The other thing, if you *do* believe college satisfaction and employment survey data at face value, then this *really* is the financial equivalent of jumping off the cliff.

One reason I trust face to face interviews when it comes to figuring out someone's beliefs more than I do surveys, is that people find it easy to lie on surveys. Someone asks you if you are employed and how much money you make, people just don't feel bad about checking the wrong box. People find it much, much harder to fudge the truth, when you look them in the eye, and even if do it, there are a lot of unconscious cues. (There are a few pathological liars that can control their behavior, but they are extremely rare.)

So if I ask "are you satisfied with your education and your job" and it's a survey, I don't know what to make of it. If I'm eating lunch with them, their mouth may say "great!" but their eyes may tell a different story.

 

[twofish-quant]

One other excellent book is Gall, Borg, and Gall Educational Research: An Introduction. Chapter one is an excellent introduction into the relationship between quantitative and qualitative research and has a section on the nature of reality (seriously). Chapter 15 has a deep discussion on the various qualititative research traditions. There is a whole field called "life history research."

Also, one reason that I like the "university as a trade school" philosophy is that you find all of this stuff *incredibly* useful when looking for, getting, and keeping a job. The problem with whatever the alternative is, is that it forces you to become very silo'ed, whereas "trade school" opened my mind to very deep ideas.

 

[ZapperZ]

I don't see why. Note that a typical social scientist isn't interested in researching whether there are UFO's or not, they are interested in researching what people *believe* about UFO's, and the research methods for figuring out what people *believe* are quite different from the research methods for figuring physics problems. There are some interesting philosophical issues here, and if you really think that qualitative social science methodology leads to pseudoscience, put on your asbestos suit, and there are people that can argue the issue better than I can.

There is a difference between not accepting something in principle, versus acknowledging that while the principle is sound, how it is applied could be suspect. You appear to reject ANY kind of statistics wholesale. And what's worse, you appear to think that you anecdotal evidence trump over everything else!

That is why I said that using your methodology, one could justify all kinds of pseudoscience phenomena, because after all, a person only needs to "see it with his own eyes" and accept it, no matter whether is valid or not.

I can easily describe scenarios where your anecdotal situation is really not reflective of the general trend. Your sampling of one single data point could easily be the exception rather than the rule. After all, if one asks all the "alumni" of students of Phil Anderson, one can easily see that all of them ended up with exceptional careers in prestigious institutions. Does that mean that a career in theoretical physics has that bright of a future? Absolutely not! Yet, using your "logic", this is perfectly valid and trumps over all other statistics and reality of such a career!

So if you think you can find fault in statistics, I can easily counter that by finding faults to your methodology as well!

Zz.

 

[atyy]

I can easily describe scenarios where your anecdotal situation is really not reflective of the general trend. Your sampling of one single data point could easily be the exception rather than the rule. After all, if one asks all the "alumni" of students of Phil Anderson, one can easily see that all of them ended up with exceptional careers in prestigious institutions. Does that mean that a career in theoretical physics has that bright of a future? Absolutely not! Yet, using your "logic", this is perfectly valid and trumps over all other statistics and reality of such a career!

But if Phil Anderson's students' students did not end up with exceptional careers in prestigious institutions, that would show your counterexample false. Do you have evidence that Phil Anderson's students' students ended up where you think they did?

 

[ZapperZ]

But if Phil Anderson's students' students did not end up with exceptional careers in prestigious institutions, that would show your counterexample false. Do you have evidence that Phil Anderson's students' students ended up where you think they did?

They sure do! All of them!

But even if I find ONE that didn't, what does it tell you? Is that ONE example enough to nullify all the rest? Besides, isn't this also doing a statistical analysis, albeit at a smaller sampling?

I find it frightening that some of us think we have a grasp of our world based on a SINGLE data point AND use that to advise others! I find that to be highly irresponsible!

Zz.

 

[Dembadon]

... My question is: why are there so few physics majors? ...

It has been my experience that the average student at my university prefers the path of least resistance. Science and engineering programs take a great deal of effort and I believe a lot of students will be happy to just make it through their college experience with good enough scores to please family, friends, and potential employers, no matter what their major is. At least that's the impression I've been getting for the past couple years.

That said, physics is certainly not one of the paths within the realm of least resistance!

 

[atyy]

They sure do! All of them!

But even if I find ONE that didn't, what does it tell you? Is that ONE example enough to nullify all the rest? Besides, isn't this also doing a statistical analysis, albeit at a smaller sampling?

I find it frightening that some of us think we have a grasp of our world based on a SINGLE data point AND use that to advise others! I find that to be highly irresponsible!

Zz.

How do you know this?

But anyway, yes, I agree that good statistics are better than anecdotal evidence. However, two-fish's experience that people who go to university are extremely concerned about jobs is easily replicated. So if a survey comes up with a contradictory result, then one might ask for evidence that its methods were good, rather than giving it the benefit of the doubt.

 

[ModusPwnd]

I think there are few physics majors because people don't like doing physics and math for the most part. Also, the major is best for people who want to go to grad school and most people don't want to go to grad school.

 

[FroChro]

T
I can easily describe scenarios where your anecdotal situation is really not reflective of the general trend. Your sampling of one single data point could easily be the exception rather than the rule. After all, if one asks all the "alumni" of students of Phil Anderson, one can easily see that all of them ended up with exceptional careers in prestigious institutions. Does that mean that a career in theoretical physics has that bright of a future? Absolutely not! Yet, using your "logic", this is perfectly valid and trumps over all other statistics and reality of such a career!

So if you think you can find fault in statistics, I can easily counter that by finding faults to your methodology as well!

Zz.

You could interview those guys, find out they had same advisor, and then get more information about him - it is about making sense out of it.

You can't use statistical approach to non-statistical data.

If a measurement is incompatible with accepted theory, first question the experiment. If statistic is inconsistent with what I know about the world, first I question the statistic.

 

[FroChro]

I find it frightening that some of us think we have a grasp of our world based on a SINGLE data point AND use that to advise others! I find that to be highly irresponsible!

Zz.

If you accept an advice is viable, you have to accept the world is comprehensible. To be so, there need to be some ''patterns''. Using them you can extrapolate from ''single'' data. (Not to say that word ''single'' is highly inappropriate.)

In other words, anecdotal evidence may be very deep and throughout -> you can analyze it and make a sense out of it. Statistics, while clearly extremely useful, may be very tricky to interpret.

To me it is more importatnt whether an evidence make sense than what kind of evidence it is. (Althought once you have a enough evidence that doesn't make sense you should maybe change a definition of sensible.)

 

[ParticleGrl]

ZapperZ, let's refocus this idea of statistics back to the number in question:

Note that concern about paying for college has fallen to an all-time low of under 12%

Does this really seem like it can possibly be a valid number? In your personal estimation, how much weight should that number be given?

Most of my friends from high school went to college, and made the decision of where to go based on cost. Most of my friends in college were nervous about the debt they were taking on. Many of my friends in graduate school were struggling mightily to pay down private loans they took out to pay for undergrad, etc. The students I taught who couldn't find jobs after undergrad were tremendously worried about defaulting on their loans. One of my coworkers wants to be a stay-at-home mom, but has been bartending for the last three years to pay off her college degree. I can't imagine how life could have provided me with a sample with a number SO different (80%-90%) from the survey.

 

[ParticleGrl]

If college is supposed to be a trade school with more beer, why are there history majors? Women's studies majors? And, to pick on the usual punching bag, art history majors?

Generally, I'd guess bad advising. If you grab an art-history major and ask them why they are doing it, you'll get something like "well, it doesn't really matter what your college degree is in, just that you have one. So you might as well study something that interests you." My high school and college academic counselors told me the sheepskin was the thing, not the specifics of the degree.

The problem I find is that employers want to treat college like a trade school, so the students who treat it like one as well seem to have a leg up on those who don't. My friends in engineering programs nearly all got jobs in science/tech/engineering. My friends in science programs went on to get phds, and few of them have the sort of jobs they wanted.

 

[ZapperZ]

You could interview those guys, find out they had same advisor, and then get more information about him - it is about making sense out of it.

But this is my point! You can't simply accept some anecdotal evidence, and then generalized it! The fact that Anderson is a Nobel laureate, he gets to choose the top of the top of the crop for students, and the fact that his reputation alone is sufficient to land his graduates serious consideration (what I call pedigree) for many positions are ALL FACTORS that make this to NOT be the norm! Ignoring such a thing will give a seriously distorted view of the world!

When one looks at only ONE data point, one can't tell if such a situation is due to some exception or the rule, or a unique case. When you interview someone, was he/she describing the situation at that time?? Or is the situation still valid now? Do you get an idea of the TREND over a period of time to give you some indication of where things are heading?

If you accept an advice is viable, you have to accept the world is comprehensible. To be so, there need to be some ''patterns''. Using them you can extrapolate from ''single'' data. (Not to say that word ''single'' is highly inappropriate.)

In other words, anecdotal evidence may be very deep and throughout -> you can analyze it and make a sense out of it. Statistics, while clearly extremely useful, may be very tricky to interpret.

To me it is more importatnt whether an evidence make sense than what kind of evidence it is. (Althought once you have a enough evidence that doesn't make sense you should maybe change a definition of sensible.)

See above.

I've been in this field for quite a while, and I think I'm perceptive enough to know many things about it. Still, I always look at many of the statistics that characterize different aspects of the field, and I continue to be surprised by many. I'm not saying that all of them are accurate, but whether one buys them or not, these still provide some amount of snapshot of a larger sampling than anyone of us can do. As someone who advices students, I do not want my preconceive ideas of how I think things are going to trump over a series of consistent statistical results. Things change over a period of time, and often, they change very quickly, as we have seen the past couple of years.

At this point, I think people are misinterpreting my objection here. I object to putting anecdotal, single-data-point "evidence" ahead of statistics, or even ignoring statistics completely. I didn't say to ignore anecdotal evidence! I question the principle that no statistics can be trusted, and that ONLY anecdotal evidence is valid! It is why I asked for this clarification from twofish quant. There is a difference between dismissing all statistics as a matter of principle, versus stating that certain ones are no good due to various problems. If one believe the former, than one has to also dismiss all experimental high energy physics.

ZapperZ, let's refocus this idea of statistics back to the number in question:
Does this really seem like it can possibly be a valid number? In your personal estimation, how much weight should that number be given?

I wouldn't know. I have no knowledge of this number. If I were to go by MY situation, then I'd say all students are in great shape, because I graduated with ZERO debt. Now, how about THAT anecdotal evidence?

Zz.

 

[atyy]

The Numbers That Really Intrigue Physics Majors
The median annual income for a physicist is $94,240. The middle 50% earn between $72,910 and $117,080. The lowest ten percent earned less than $52,070, and the highest ten percent earn $143,570. The average starting salary offer to Physics doctoral degree candidates is $52,460.

 

[ParticleGrl]

ZapperZ, if I remember correctly, you teach at an undergrad institution. And surely, regardless of occupation,you interact with some recent college grads. Does your experience suggest college is getting easier or harder to pay for? Generalizing from one point is bad, but so is trusting a number that seems in obvious conflict with other facts. College costs are rising, and aid is being cut. A survey shows that worries about paying for college are at an all time low? Skepticism is surely warranted.

 

[FroChro]

Thank you for answer, I quite agree with you and arguing over details would be pointless here.
Just one objection:

When one looks at only ONE data point, one can't tell if such a situation is due to some exception or the rule, or a unique case. When you interview someone, was he/she describing the situation at that time?? Or is the situation still valid now? Do you get an idea of the TREND over a period of time to give you some indication of where things are heading?

Zz.

I think properly done analysis of anecdotal evidence should be able to take (some of) those things mentioned above into consideration. In the example you provided I believe I would realize that sample I am interviewing is quite exceptional.
Of course, I agree that most time a conclusion wouldn't be perfectly right, but statistics are neither.

 

[ZapperZ]

ZapperZ, if I remember correctly, you teach at an undergrad institution. And surely, regardless of occupation,you interact with some recent college grads. Does your experience suggest college is getting easier or harder to pay for? Generalizing from one point is bad, but so is trusting a number that seems in obvious conflict with other facts. College costs are rising, and aid is being cut. A survey shows that worries about paying for college are at an all time low? Skepticism is surely warranted.

I'm not teaching at an undergrad school. I work at a US Nat'l Lab.

Zz.

 

[ParticleGrl]

The Numbers That Really Intrigue Physics Majors
The median annual income for a physicist is $94,240. The middle 50% earn between $72,910 and $117,080. The lowest ten percent earned less than $52,070, and the highest ten percent earn $143,570. The average starting salary offer to Physics doctoral degree candidates is $52,460.

And here we have an excellent example of how to mislead with numbers. Every single one of these numbers is totally irrelevant to a physics major with no intention of getting a phd in physics- why? Because physics bachelors don't work as physicists! Its not only irrelevant, its misleading. This seems to suggest that study physics -> work in physics is as normal as something like study engineering -> work in engineering.

Now, for those who DO plan on going to graduate school, many of these still aren't relevant. Many phd holding physicists are never able to get a job as a physicist!

The number that is of some relevance is the average starting salary offer to physics doctoral degree candidates. The average starting salary for an bachelors engineer is between 50 and 60k, depending on the type of engineering. So an engineer makes as much with 4 years of school as a phd physicist does with 10.

Of course, an engineer with an extra 6 years of experience is probably making between 70-80k. By the time the physicist has gotten his job offer, he could have saved an extra 240k had he been an engineer. Even after he gets hired with his phd he is making 20k-30k a year less than he would have been with an engineering degree.

 

[DrummingAtom]

The Numbers That Really Intrigue Physics Majors
The median annual income for a physicist is $94,240. The middle 50% earn between $72,910 and $117,080. The lowest ten percent earned less than $52,070, and the highest ten percent earn $143,570. The average starting salary offer to Physics doctoral degree candidates is $52,460.

This reminded me of a story. I told my friend a few years back that I was considering going back to school for physics and he immediately said "Oh physicists make like $90k/year, don't they?" I didn't really know what their salary was but I asked him how he would know that since he went to school for chemistry. He said he remembered a physics professor saying it in his first physics class lecture. This lecture had about 400 students and the university is one of the largest in the nation. After reading on here, I later found out that most physics majors/PhD's struggle to ever become physicists.. I'm sure the professor left that part out.

I completely agree with ParticleGrl, it's very misleading to use numbers that aren't relevant to positions these physics majors/PhD's ever get.

 

[twofish-quant]

There is a difference between not accepting something in principle, versus acknowledging that while the principle is sound, how it is applied could be suspect. You appear to reject ANY kind of statistics wholesale. And what's worse, you appear to think that you anecdotal evidence trump over everything else!

What ever gave you that idea?

In the astrophysics forums, I talk a lot about galaxy counts as evidence for baryon concentration. Also, if we were talking about double-blind randomized drug trials, I'd take that over anecdotal evidence for the effectiveness of a drug.

But we aren't talking about galaxy count data and double-blind randomized drug trials.

*IN THIS PARTICULAR SITUATION*, I've found that anecdotal data is far, far more useful than survey data. The problem with using survey data is that in general *FOR THE TOPIC UNDER DISCUSSION*, I've found that the surveys tend to be extremely poorly set up so that I already have huge amounts of suspicion, and the surveys rarely provide enough data so that you can critique the methodology. Also, even when done right, surveys usually don't provide useful information, and what you find out from them is often irrelevant.

One thing about anecdotal data, you can usually get enough deep information to cross-check for reliability. In situations where "professionals" do surveys, you can also get that sort of information, but a lot of surveys are done by amateurs, who invariably don't provide enough information about how the survey was done, and even when you have a good survey, it may not contain the information that you want.

For example, I want to figure out how my wife or my boss will react if I do something, I do that based on anecdotal evidence. Now I could do a survey of 1000 wives or 1000 bosses, and *even if it was a valid survey* that information would be totally useless to me.

That is why I said that using your methodology, one could justify all kinds of pseudoscience phenomena, because after all, a person only needs to "see it with his own eyes" and accept it, no matter whether is valid or not.

No that's not how it works. There are a ton of techniques to make sure that when doing qualitative research, what you are doing is both reliable (i.e. reproduciable) and valid (i.e. measuring what you think you are measuring). For example, I gave a transcript of this thread to a social scientist, they would start "coding" it. They would break it up into themes, and then count the number of times I mention something, Then, they would give it to some other person and then would code it. At that point, you look at the results, and if two people read what I say and come to different conclusions about what I meant, then it's not reliable, whereas if you have two people read what I say and then code it in the same way then there is something there.

Also, if a social scientist talks to a young Earth creationist, he is interested in their **beliefs** about geology, not geology itself. I've found those techniques to be useful, because if I just tell them that the Earth is 6000 years old and they are an idiot for believing that, then I'm not going to convince them, but if I can get inside their mind, and understand *why* they think the world is 6000 years old, then maybe I can convince them otherwise.

[QUOTE[I can easily describe scenarios where your anecdotal situation is really not reflective of the general trend. Your sampling of one single data point could easily be the exception rather than the rule.[/QUOTE]

Sure, but sometimes that's what you want. If you have an bad inner city school, and 99.9% of the people in it end up in menial jobs and one person makes it to Harvard, that's may be the one person you want to research *BECAUSE* he is the exception.

Similiarly, if I wanted to a study on careers of astrophysics Ph.D.'s, I probably want to interview Brian May. It's not that it's typical for astrophysics Ph.D. to turn into rock band drummers, but *because* Brian May is unusual, he is worth studying.

One thing about people is that everyone is exceptional. People are not electrons. Every electron is exactly the same as every other electron, but every person is different from every other person. This is a big headache for educational studies. What you want to do is to have two classrooms that are exactly the same except for one variable. You *can* do this to some extent with drug trials, but this turns out to be *impossible* for classrooms or workplaces. And even if you could do this, there are hundreds maybe thousands of variables that interact in complex ways.

After all, if one asks all the "alumni" of students of Phil Anderson, one can easily see that all of them ended up with exceptional careers in prestigious institutions. Does that mean that a career in theoretical physics has that bright of a future? Absolutely not! Yet, using your "logic", this is perfectly valid and trumps over all other statistics and reality of such a career!

No it doesn't. If I ask all of the alumni of Phil Anderson, I learn something about the alumni of Phil Anderson, and that could be useful because I'm curious *why* alumni of Phil Anderson end up with exceptional careers. Is it because he selects people ahead of time? Is it because he is particularly good with political connections? Is it because he does something "magical", and if so is it something that can be reproduced? Is it just dumb luck?

I suspect that by asking these sorts of questions, I'm going to learn 100x more about the Ph.D. system that if I send out a survey to 100 Ph.D.'s.

[QUOTE[So if you think you can find fault in statistics, I can easily counter that by finding faults to your methodology as well![/QUOTE]

I think that statistics *in this particular situation* tend to be unreliable or useless. One reason I'm a stickler about this is that I'm married to an educational researcher so I know what a proper survey looks like.

Also, a lot involves making do with what you have. I'd *love* to get statistical distributions of neutrino emissions from supernova or seismology results from stars, but I can't. If I had a wayback machine, I'd love to rerun the big bang 1000 times and see what comes out, or go back in time, change my life, and see what happens.

But that's not available so I have to make do with what I have.

 

[twofish-quant]

Besides, isn't this also doing a statistical analysis, albeit at a smaller sampling?

No it isn't. One thing about Ph.D. outcomes is that the numbers are sometimes small enough so that you *can't* do statistical sampling. Statistical sampling involves taking a small number of a large population, taking a measurement of that sample, and then extrapolating the result. You can then run probabilities to see the sampling error.

The problem is that when you run into small numbers, everything breaks down. If you have a population size of about 20, and you sample 4, the sampling errors are so huge that you can't make any generalizations. In order to make generalizations you have to take a large enough fraction of the population, that you really aren't doing statistics.

I find it frightening that some of us think we have a grasp of our world based on a SINGLE data point AND use that to advise others! I find that to be highly irresponsible!

The problem here is that I know what I'm thinking and I know my life history. If I could read people's minds, and then insert my soul into their bodies so that I can relive their life, then maybe I could talk about someone else's life, but my telepathy machine isn't working.

Also, I try not to give any advice, and I also try to make it clear that my life isn't anyone else's life. One thing I an interested in is how relevant my experiences really are. For example, I graduated in the middle of the dot-com boom and it was trivially easy to find a job in industry. I'm curious about how people now are fairing.

The other thing is that one big problem with Ph.D. advising is that you may be talking to eight people, but if all of them are professors in major universities, that's not going to be that useful since you really don't have 8 data points, you really have more or less one. Now if you talk to someone that isn't a professor, you might learn something.

The other thing is that there are a lot of different Ph.D.'s here and you can make some inferences based on what various people are saying. For example, one thing that happened to be was that going from Ph.D. -> work was an extremely psychologically *painful* and traumatic experience. Now if I was the only one that mentioned this, then I'm weird, but you have a *lot* of people with the same sorts of anger issues, which means that it's probably common.

 

[twofish-quant]

But anyway, yes, I agree that good statistics are better than anecdotal evidence.

I don't think it's inherently better or worse. It's all pieces of the puzzle. One problem with survey data is that sometimes even if you do a good one, the results are often very superficial, and they don't answer the question that you are interested in.

Also the quality of anecdotal evidence varies. For example, if I find someone that gets cured of terminal liver cancer after eating yellow jellybeans, that really tells me nothing since it's well known that in any group of cancer sufferers some people will spontaneously get better.

 

[atyy]

I don't think it's inherently better or worse. It's all pieces of the puzzle. One problem with survey data is that sometimes even if you do a good one, the results are often very superficial, and they don't answer the question that you are interested in.

Also the quality of anecdotal evidence varies. For example, if I find someone that gets cured of terminal liver cancer after eating yellow jellybeans, that really tells me nothing since it's well known that in any group of cancer sufferers some people will spontaneously get better.

By good, I also meant relevant.

 

[twofish-quant]

But this is my point! You can't simply accept some anecdotal evidence, and then generalized it!

I'm not generalizing. My response to the survey results was that they don't match my experience. Maybe I'm weird.

The fact that Anderson is a Nobel laureate, he gets to choose the top of the top of the crop for students, and the fact that his reputation alone is sufficient to land his graduates serious consideration (what I call pedigree) for many positions are ALL FACTORS that make this to NOT be the norm! Ignoring such a thing will give a seriously distorted view of the world!

But I'd be seriously interested in which factors are most important. It may be that once you interview him and his students, you find that it's not the obvious factors, but something else. Also, you can cross check things. If it's being a Nobel laureate, then how are other Nobel laureates doing? Also what criterion *defines* cream of the crop? Does he give his prospective students an IQ test? I don't think so but I don't know. Are his students more driven than other students? Where does his "reputation" come from?

The cool thing is that if you let be sit down for an afternoon talking with him and his students, I might be able to learn something interesting. Also, a lot of this is "actionable." If you find out that his students are getting jobs because he gives them particularly good career advice, then you can do the same thing. Or you might find that they are getting jobs because of some factor you can't replicate in which case you are screwed.

When one looks at only ONE data point, one can't tell if such a situation is due to some exception or the rule, or a unique case.

Sure. That may not be a bad thing. Also one reason that I post is so cross check my experiences. I'm seeing this. If everyone else sees something different, then maybe I just have weird friends. However, it's interesting that this is *not* the reaction.

When you interview someone, was he/she describing the situation at that time?? Or is the situation still valid now? Do you get an idea of the TREND over a period of time to give you some indication of where things are heading?

Sure those are issues. Also, memory is very tricky. I can tell you want I think I remember from 1998 when I graduated, but memory is imperfect. There are also some pretty well known cognitive biases in memory. For example, people tend to put events in a narrative, and then events that don't fit, don't get remembered.

The other problem is that people are often looking for information that doesn't exist. Someone that graduates in 2017 doesn't care about the median salary in 2012, but rather the median salary in 2017, which is information that doesn't exist.

I'm not saying that all of them are accurate, but whether one buys them or not, these still provide some amount of snapshot of a larger sampling than anyone of us can do.

They often don't, and I've found salary statistics to often to be wildly misleading, and the methodology are terrible.

As someone who advices students, I do not want my preconceive ideas of how I think things are going to trump over a series of consistent statistical results.

I do, since the statistics are more often than not bogus. Personal experience again plays a factor, because when I was an undergraduate, I was presented with highly bogus statistics talking about a shortage of physicists. Fortunately, those stats smelled bad enough so that I didn't believe them.

The one thing that I would have liked to have was a talk with someone that lived through the physics crash of the 1970's, since one thing that no one told me was that there *was* a physics crash of the 1970's. I'd *still* be interested in talking with a physics Ph.D. from that era to see how their lives turned out.

One rule of thumb is to be very suspicious of statistics purporting to show something that financially benefits the person presenting the statistics to you. Also there is a bit of accountability. Suppose, you use the APS stats and they get it wrong, and screws up someone's life. Nothing bad is going to happen to them. Now if I give some student some horrible advice, they will be able to find me, and at least make me feel bad about it.

At this point, I think people are misinterpreting my objection. I object to putting anecdotal, single-data-point "evidence" ahead of statistics, or even ignoring statistics completely.

I object in certain cases, and not in others. Statistics regarding job demand and salaries are so suspect that they often do more harm than good. In this situation, people were handing out bogus projections as late as 2005, when people gave up, and I think the internet had something to do with it.

Also, I can't tell the future. One reason for talking to people is that sometimes you are looking for a piece of information that isn't obvious.

For example, I would have liked to ask someone from the 1970's, "so when did you realize that you weren't going into academia, and how did it *FEEL*" One thing about me that makes me feel old is that it was much harder to find someone with that information before the internet. People had been putting up bogus projections of a science shortage since 1975, and my guess is that it only stopped after the internet got established enough so that you couldn't get away with that.

I didn't say to ignore anecdotal evidence! I question the principle that no statistics can be trusted, and that ONLY anecdotal evidence is valid!

And since I talk about galaxy count data all of the time, I wonder what I ever said to made you think I believed that.

But we aren't talking about galaxy counts. We are talking about survey data of student motivations, in which I think that even well done surveys have serious limitations.

 

[physics girl phd]

No one goes to school in physics because they want to be a stay-at-home mom when they grow up.

Hey -- physics is sexy!

I know of at least a few "women in physics" who are now stay-at-home moms.

Note: I'm not counting myself COMPLETELY in that number either, since I get the "trailing spouse" part-time employ... ugh. But: We just got back our course evaluations from last term... and I ROCKED IT (looking at the freely-available info out there, my numbers are higher in the "calc-based EM for engineers" class than any other faculty who have taught the same course in the past three years!)

 

[redrum419_7]

I go to the University of Toledo and am an Astrophysics major. In my Physics for Scientists and Engineers I class, there are 2 physics majors (myself included), a chemistry major, a biology major, and 107 engineers.

 

[ParticleGrl]

I know of at least a few "women in physics" who are now stay-at-home moms.

Which is fine- but I bet pretty much all of them would leap at the chance for a full-time position (be it industry or academic) where they can do some physics in their geographic area. Which sort of underscores the whole "its hard to get a job where you can do physics, even with a phd" thing.

I stand by my original assertion- people generally don't do physics because most people who COULD do physics CAN do engineering, and the latter is a better chance at the sort of career physics majors/engineers want.

 

[daveyrocket]

Hey -- physics is sexy!

I know of at least a few "women in physics" who are now stay-at-home moms.

Note: I'm not counting myself COMPLETELY in that number either, since I get the "trailing spouse" part-time employ... ugh. But: We just got back our course evaluations from last term... and I ROCKED IT (looking at the freely-available info out there, my numbers are higher in the "calc-based EM for engineers" class than any other faculty who have taught the same course in the past three years!)

Note that the post I was quoting was listing a variety of jobs for physics majors as support for a claim of versatility of physicists. Anyone with *any* education can become a stay-at-home mom or dad. So it doesn't say anything about the versatility of a physicist.

 

[lisab]

Hey -- physics is sexy!

I know of at least a few "women in physics" who are now stay-at-home moms.

Note: I'm not counting myself COMPLETELY in that number either, since I get the "trailing spouse" part-time employ... ugh. But: We just got back our course evaluations from last term... and I ROCKED IT (looking at the freely-available info out there, my numbers are higher in the "calc-based EM for engineers" class than any other faculty who have taught the same course in the past three years!)

I was a stay-home mom for a while after I got my BS. Turns out to be a pretty good preparation for mommyhood, based on how my daughter turned out .