I always enjoy Shankar Vedantam's pieces on Morning Edition, but as a recipient of one of Milkman's scam emails I feel compelled to write in about this story. While I am not a social scientist, and while I don't dispute the data she has, it is unfortunate that some of the context of this story has been ignored. First, as Shankar undoubtedly knows, people in public positions are likely to receive scores or even hundreds of unsolicited email every day. This is certainly true with professors. How one responds to unsolicited email, however, is widely divergent and may have little to do with the simple racial and ethnic biases that MIlkman and Co. claim to have found. For one, some professors whom I know simply do not respond to any unsolicited email, period. Is that 'rude'? Possibly, but it cannot be considered a sign of 'bias'. Secondly, some professors (myself included) are admittedly widely inconsistent when it comes to answering unsolicited email. While it may sound self-serving, I like to think that it my case whether or not I respond to an unsolicited email has more to do with the following external factors than with subtle encodings of racial or ethnic bias: my workload at the moment I receive the email, including but not limited to deadlines that are pending and/or the amount of grading I am facing; the time of the semester (early is better than middle or later); and so forth. These factors are far more likely to determine whether or not I respond. Now, it may be that Milkman's study discusses such variables (I have not read it), but the story as presented on NPR did not consider anything like this; instead, the piece implied that professors' sole criteria was the alleged racial or ethnic identity of the alleged sender of the email as evidenced solely by the sender's 'name'. So either Milkman or Vedantam was pushing a conclusion that is far less certain that the story made it out to be. Finally, I have to note that Milkman's methodology is, at least in my case, extremely haphazard. When I received one of her 'scam' emails, it was from a student (whose name I do not recall - no, seriously!) praising my scholarship and seeking to pursue a Ph.D. under my direction. Flattering, but completely misinformed. While my department does offer a Ph.D. in History, it does so only in the field of US History, and I am not a US historian. So when I received the 'scam' email from this alleged student, I filed it away in my mind in the category of 'students who are so unserious as to not merit further consideration'; after all, had the alleged student spent any time on our department website (as all serious students do), he/she would have realized that doctoral work under my supervision was impossible. Of course my case may be exceptional. But I'm sure there are lots of comparable stories to be found among Milkman's data. After all, she is not interested in human beings and the complex motivations for their behavior; instead, she is interested only in comparing percentages of response rates. That, I'm afraid, is not very satisfying.

Hi Richard,
None of your stated factors that prevent you from responding to a student's email should correlate with that student's gender or race, so your argument doesn't make any sense. If your workload is resulting in you responding randomly to only 10% of the emails you receive from students, then no such bias against race or gender should be observed.

Unless the authors of the study decided to use all Hispanic names on one day, Anglo-Saxon names on the next, etc... But that seems such an obvious source of potential error I have difficulty believing that is the case. Richard has explained why he didn't respond in his instance, but it does nothing to explain the discrepancy in responses when you change names or across departments.

We will have to agree to disagree.

Richard,
Reading your extensive posts above and below I expected a explanatory answer here rather than a sentence.

Let's be clear. I am not going to engage in a close debate about social science statistical data methods. I am not qualified to do that. I can only comment from my perspective. I have done so fully, honestly, and in detail (see most recently the reply to Eric Ledlum). Unless you would like to accuse me of implicit, unconscious racism for failing to respond to Juanita Martinez, my point has been merely to suggest possible reasons why some responses might differ from the explanations offered by Milkman and, more significantly, by Shankar Vedantam. Of course I realize I'm one person, and not an entire study. But how many other respondents may have responded as I did? In my opinion, the initial email was so poorly conceived and crafted as to cast serious doubts about the interpretation of its results. If I had received an email that listed the qualifications and experience of the 'fictional student', if she had told me she wished to meet with me for plausible reasons other than becoming my PhD student (advice about the job market, again given some basic qualifications), and THEN I ignored it, then there might be something to my having acted out of unconcsious bias. I'd like to think that had I received such an email, I would have gladly agreed to meet with such a promising student, who actually knew something about my field. We won't know. The point has been that with such a transparently bogus email, why WOULD I respond? Do you respond to all the spam email you get?

Richard,
I am in the Bio-medical field and receive a fair amount of emails from foreign students looking for post graduate studies. My institution does a good job in filtering spam. Many write to me based on published papers (unsolicited?). Quality of writing (grammar and spelling) in most cases is poor though the very few who joined me has been excellent students. I have a standard letter to respond to them and do not pay attention to their degrees, institutions, type of email they use or praise of my research at this stage. Though few local students write to me, if I received Milkmans "spam" I would have responded with my standard email irrespective of their gender or race.

If a student didn't have the courtesy to do a literally five minute search on my research and say "your research on the effect of global warming on marine organisms" instead of "your research", why should I have the courtesy to think that they were making meeting me their "first priority"?

Shocked at how many responses they did get. I would not have responded at all if they didn't note at least a detail or two about my research, showing it was not just spamming many many professors.

"she is not interested in human beings and the complex motivations for their behavior; instead, she is interested only in comparing percentages of response rates."

Is there any other way to efficiently conduct a study with a large sample size? What you call disinterest in human beings most in academia would call "research."

I'm sorry but your whole comment reads like, "I know there's racism and racial bias out there, but it's not coming from me. Oh no, never me." I know that's not how you intended it, but if you read it again I bet you will see how someone could infer a defensive tone.

I accept that I'm defensive. My point is simply that there are lots of possible explanations for the apparent data. I am a humanist, after all, and not a social scientist. My research would not proceed in such ways, as I would want not aggregates but highly contextualized interpretations of multiple results. That is, perhaps, a fundamental difference between social scientists and humanists. I don't say one is better, nor do I say Milkman is 'wrong'; I do say that the presentation on NPR missed a lot of nuance, and ended up sounding congratulatory of this study, when in fact it has been the subject of controversy since its inception (see Chronicle of Higher Education).

What an interesting clarification! A request to meet with a potential
doctoral student at a specific date and time DOES seem very different
from a more general "I was wondering if you could meet with me to talk
about my academic future," which was the impression left by the NPR
story. A potential doctoral student is a much bigger commitment, and
likely one that would take more than 10 minutes to consider. Plus,
that's a student essentially asking for a job. While the email is not
poorly written, I wouldn't call it "PhD caliber," particularly since it
was not personalized in any way.

I do think that the massive
statistical difference in the responses to white students vs students of
color, and to male students vs female students, is indicative of some
bias, not necessarily of each (or any) individual recipient, but collectively within academia.

Either you are an academic who believes in numbers (data) or not. Trying hard to explain it
away undercuts your credentials as professor.

Interesting. I was joking with my wife this morning as I heard this story that there does need to be some follow-up as to why the profs disregarded minority and female queries more often, and said, "hey they can bring some thumbscrews: 'why did you not reply to Latoya or Juanita?'"

So its interesting that you got one of the letters and described the issue with your particular case.

I'm resonably certain there *is* some racial bias, but how that works may be innocent enough: I hazarded a guess that some professors might find it very unusual in their field to receive a letter from a Latoya or a Depak and the name itself may make them give additional scrutiny to the letter, perhaps looking up the name to see if the student is even enrolled in the school. I wonder if they used, say, eastern european names, or "hippy" names that stand out as "unusual" if that would also signal more scrutiny

But i think the main point is that while you may have a special explanation for your case, a large enough sample should even out random factors like one prof being especially busy and unable to reply. (though perhaps alternatively, profs in the most lucrative areas might in fact be the busiest, but then why would they consistently reply to the Brads but not the Patels??)

what I also found interesting is that being a minority or woman prof onesself seemingly did NOT make it any more likely that one would respond to a minority or female query.

Sadly a study like this is probably not easily repeatable: knowledge of this kind of study being done might affect the behavior of profs.

I can offer a possible explanation for the low response rate for those with Asian sounding names. Many faculty receive numerous such letters from students from India and China hoping to get into Ph.D. (or post-doc) programs and sometimes even needing help getting visas to study here. The texts read as though they were written to be general enough to be blasted out to thousands of professors. I can easily believe that any letter bearing and Indian or Chinese name and resembling one of these 'email blasts' would prompt a reflexive stab at the delete key.

Agree that very foreign sounding names, of any type, would be deleted or filed quicker because of the perception of having to obtain a visa for the student. A lot of issues with this currently in my department, the students or researchers are very aggressive about getting help, and sometimes the university does not want to sponsor them and the head of the group is put at a disadvantage.

It would have been helpful to somehow include that the student is an American citizen, to tease out bias due to that factor.

The same mechanism would also result in a large differential response rate for two vaguely Italian sounding names, John Viaggi and John Viagra.

Notice no Italian, Polish, Russian, Armenian, Arabic names.

I found the email in the paper.

The precise wording of emails received by faculty was as follows:

Subject Line: Prospective Doctoral Student (On Campus Today/[Next Monday])
Dear Professor [Surname of Professor Inserted Here],
I am writing you because I am a prospective doctoral student with considerable interest in your research. My plan is to apply to doctoral programs this coming fall, and I am eager to learn as much as I can about research opportunities in the meantime.
I will be on campus today/[next Monday], and although I know it is short notice, I was wondering if you might have 10 minutes when you would be willing to meet with me to briefly talk about your work and any possible opportunities for me to get involved in your research. Any time that would be convenient for you would be fine with me, as meeting with you is my first priority during this campus visit.
Thank you in advance for your consideration.
Sincerely,
[Student’s Full Name Inserted Here]

Thoughts on "today/next Monday"?

How did they pick? Didn't they realize that could be a factor?

You bring up a whole bunch of.... let's just politely call it irrelevant, stuff.

The content of the email hardly matters. That's being controlled for. It's the same email written by the same people. All that's changing is the name attached to it, which gives indication as to the ethnicity.

Very defensive of you.

It matters a lot.

If the predisposition is not to respond at all because the time frame is very tight (today/next Monday, with no clarification as to which was used and why), it skews the results.

Richard, that may be true for any individual professor. But if those were the reasons, then the lower response rate should be relatively equal across all races, especially if the letters are exactly the same. 6500 letters, it cannot be that so many professors just happened to be busy when a minority emailed them, not over such a wide range of professors and so many mailings.

Also, the piece did not imply the "sole criteria" was race, she was careful to not imply racism or personal bias. In fact, I'm sure almost every professor would not even consciously take race into consideration. However, the experiment was set up so race and field of study were the only variables. So those have to be the explanation, unless the letters were flawed somehow (such as: if the letters were not exactly the same for each field of study). Even if the letters were poorly written (as some suggest below) then if they were all exactly the same for that field, then the negative responses should have been both high and equal across all races. But they weren't.

Richard,
You can find the paper here
http://papers.ssrn.com/sol3...

I agree that there is potentially a lot that they are not controlling for in their study. I suppose their hope is that some of those idiosyncratic features of a person or their day is completely random across their sample -- that is, it's not the case that only History professors have issues with grading while Accounting professors have issues with some other thing. If this is the case, then these shouldn't bias the qualitative results. That said, I have not read the paper and am only offering one possible response.

Here's what looks to be an identical study from 2012 by the same authors.
http://knowledge.wharton.up...

Here's a striking comment from the last paragraph of that makes you think about all the factors that could contribute to this (emphasis mine):

The paper has already changed the way Milkman responds to prospective students who seek out her advice. “I used to just hit ‘delete’ if I didn’t know the person. I did not respond to those emails,” she says. “I have completely changed my behavior. Now that I have done this study, I know that this is where subtle discrimination can seep in. I now try to respond uniformly across the board to everyone.”

Might the same thing be happening here -- professors who don't interact with female/minority students as much as they do with their white/male counterparts, intentional or unintentional. If so, then getting closer to the root of the problem would be to see why and in what ways do professors interact with/get to know their female and minority students less often. Is it voluntary where professors just don't make the effort, or involuntary where students don't make the effort.

And this brings up another thought, which might be a long shot, could white male names have smaller variation (more instances of the same/similar name) resulting in these randomly generated names from the study getting more hits of "I recognize this student" than the randomly generated names of other groups?

And what's up with Hispanic females? Their response rates are very VERY close to those of White males (2.3% and 6.3% difference).

Fascinating study in any case.

Good responses, guys. I'm not really disputing the statistical results. I do wonder about causation and correlation, though.That is, is the conclusion reached in the study (and reported faithfully by Shankar) the cause of the discrepancy, or is that discrepancy correlative. Clearly I was annoyed with my experience, so I may be biased in other ways. I also would not dispute the possibility that such bias exists. I did think the NPR piece was a bit simplistic.

The word "discrimination" is used, well, indiscriminately as a conclusion.

Richard,
If some one writes to you referring to your research do you consider it as unsolicited? Also, why do you think one responds to unsolicited email of some but not others?

I posted the actual text from a paper on the study above. I have NEVER gotten an email soliciting a PhD or doctoral position at my university where they actually talked about my research other than "your research is great". To the students sending the emails, I feel like I would be just as much an expert in my real field as the top marine biologist in the world, or perhaps top 17th British literature expert. So I generally do not respond. The students ALWAYS say where they are now, and what they are doing as an undergrad or grad student, so the email shown DID matter because it was wayyy too general for sciency types like myself, and likely too general for business types as well. Maybe humanities professors are just waiting for someone to ask them to dance, though my friends who teach in the humanities often have funding issues.

When one is part of a study without volunteering for it, one would hope that, just as when someone films non-professionals on the street, there is a notification afterwards and follow-up.

The follow-up could be as simple as: "We have been conducting a study, which included unsolicited emails sent to professors. You received such an email on X date, and did/did not respond. A brief note on why you did/did not respond would be helpful. All responses will be confidential and no professor will be identified by name in our study." along with a few choices such as "I am in need of doctoral students", "I did not have time to respond in the time frame presented", "I do not respond to unsolicited emails", "The student did not include a resume or CV so I could decide whether to respond", "I always respond to requests to meet with prospective students" and so on.

As Richard noted, he received such an email. Was there some kind of follow-up to the professors to help determine if there was causation? The major concern I have is that the email was very very strict in terms of time - "on campus today"? Wow. Today? Really? Why not stop by one of my classes and ask me in person to meet then or later? I suppose one could argue that the time frame should be immaterial, but in reality it is not. Skewing of results can occur when there is an overriding factor such as the strict time period.

My comment to Eric Ledlum above seems to answer your question.

To be frank, your comments to Eric brings up more questions than answers to my initial queries. Thank you any way.

fine and dandy. Doubt away.

I agree with all of the cautions on drawing causation conclusions from this study. However, to me, its finding confirms the reality women and minority students tend to feel and know. When somebody said to me "if you have a hammer, you tend to see everything as nail", I thought, "if one get hammered repeatedly, it's so exhausting to prevent oneself from seeing everything as nail."

There are more scholarship opportunities based on race and gender, so it is kind of amazing that any professor would turn down someone with those characteristics. Then again, it is a hassle for professors to mentor students who are on scholarship. Is this really about perceived economic class, since the white names were "rich kid" names not generic or "poor kid" names?

All of the white names weren't just white, they were English (as in England) first and last names. What about other Western Europeans and Eastern Europeans? What if they included Kowalski and Nemerov?

If this is about names and their impressions more than race and gender, the researchers should look at non-WASP names. It's an unfair comparison.

The majority of merit and need scholarships go to white men, so I'm not sure what that has to do with responses.

There are lots of foundation scholarships for minority students.

I used to work in a college of arts and sciences at a private research university. The department received about five-hundred applications per year for its graduate program with a dozen openings, mainly from The People's Republic of China and India. A few arrived from South America, particularly Brasil, some from Korea. US citizens applied as well, predominantly for the clinical program. At that time, graduate students earned about $30,000.- a year, in part provided by the department and in part by the thesis supervisor's research grants. Non-citizens did not qualify for independent federal funding. In the medical school, the supervisor also had to pay the student's tuition.

I received e-mails like this on a number of occasions. Quickly the conversation turned to money (which I did not have) and visas (ditto). I was only a small fish in the pond. The big fish, the ones with the money, were complaining that they were inundated with such requests.

The findings of this study don't surprise me.

They mentioned knowing what level each professor was, but I did not see a table showing the rate of response based on career level.

Other responders have explained why many of these requests could go unanswered--and there are many good reasons--but few address the result that surprised Vedantam: the comparative failure to respond to students with Asian names. In my field--English--most of us visible on the Web are bombarded with all kinds of requests for help from students with Asian names. Most are seeking free assistance in learning English as a second language. The tell-tale is that the notes are generic, not really engaging with the specifics of our scholarship. The sample note in the NPR story was generic enough to seem like yet another spam message--though the part about actually being on campus would have been more likely to make me reply.

That said, I've had a few satisfying scholarly exchanges with Asian students. Recently a pair of scholars in China wrote me inquiring about the possibility of translating my work into Chinese. I put them in touch with my editor, who asked them to send him more details of their proposal. They never replied.

I've had several PhD candidates in other countries ask to be on their dissertation committees. I agreed in one case, but never heard back from her.

That's a terrible explanation and a terrible reason.

1. We're really going to start judging people just by name now? Without reading the email at all? Cool. How is that any different from judging someone by their skin color? How is that in *any way* a valid reason?

2. Even if I grant you that Asian names are more likely to be about language blah blah (which is an incredibly racist stereotype). What about black and Hispanic names? Surely they have full grasp of the English language. What's your excuse there?

How about just accepting the fact that this is systemic racism :(

The email itself was pretty bad and not many professors would consider it. They should be shocked that they got any responses.

I would venture to say that the white names given reflect wealth, not just being white, and that factor hasn't been commented on. James Smith evokes less wealth than Brad Anderson. Maybe James Smith is too ambiguous racially for the researcher? Or what if they used Cletus McCoy as a white name? Would they get the same response?

Someone with a lot of time could answer such emails. When you are limited by time, and some things have to go, then these are one of the first ones to cut. Those with some time to spare would reply. I do not see any racism from this study. Are we saying that Fine Arts people discriminate against white females and Chinese females, and pick Chinese males and Hispanic females instead? Seems like a strange discrimination. That is what this study shows.

Dong and Wong were part of two of the Asian names picked, and used by this professor (see the study report link I posted). NFN, most folks in the US have a predisposition to wonder if those names are made up (yes, a form of discrimination, probably based on 1980s movies). However, the "white" names were somewhat uncommon but identifiably white. The Hispanic and Indian names were very generic as well. I would daresay that picking less common but "white-sounding" names is biased on the part of the researcher.

It concerns me that they only had 38 people identify the races and genders based on names alone. They had 90 names to pick from and apparently picked names that were "most obvious" for the gender race combination. I find that problematic.

Asked why there is gender and race bias in faculty mentoring (more often at private schools than public; much more often in business schools than in the natural sciences and almost undetectable in the humanities) we are told more research is needed. I will propose a different answer. When you see race and gender based disparities, you can skip the "more research" phase of the investigation and conclude with confidence that the cause of the disparity is racism and sexism.

Huh? The question is why the racism/sexism is concentrated in certain fields and types of institutions. The theory offered by Shankar, that people with more money are more likely to be in business and private schools, and also less likely to interact regularly with people from diverse backgrounds, is a decent place to start (as long as we're careful of confirmation bias).

note that the email I received did not mention 'mentoring', as if we are talking about undergraduates seeking support. I received an email from an unknown student requesting to be my doctoral student. That's not exactly the same as the warm-sounding 'mentoring' mentioned in the NPR story.

The issue with audit studies is that while experimental, most are not equipped to get at the causal processes. More research is needed, because we can't simply say it's due to overt racial/gender bias and call it a day. For instance, the fact that women and minority professors showed the same behavior suggests that overt bias is not the best explanation for the study results. One possible explanation for the race and gender effects is that minority faculty members have to make sure that they're not perceived as willing to "help a brother/sister out" which may make them less likely to mentor minority or female students. More senior and powerful minority faculty members have more freedom to explicitly work to advance minority students.

Is it really so simple? If so, how do you explain the fact that female and minority professors showed the same behavior?

Is there some clever joke that I'm missing? Are you honestly proposing that people shouldn't research thoroughly in order to fix possibly faulty research just so your beliefs can be "confirmed"?

interestingly the study (see my links below) showed that there was "reverse discrimination" (their words) in the fine arts where white males were less likely to receive a response.

I'm not surprised to see a slightly lower response rate based on race. I was a little surprised to hear that being a white male vs white female made much of a difference. So, I looked up the paper. Being a white male vs white female doesn't make much of a difference - about 1% in public universities and 9% in private universities. To put it another way: for white women in public universities, this study found no discrimination relative to white males. Saying things like "we see a 25 percentage point gap in the response rate to Caucasian males vs. women and minorities." misrepesents the situation. What we're seeing is a race-based effect. We're seeing a small difference for white women, despite the fact that the article lumps white women in with all the other minorities. The allegation of racism is much better supported than the claim of sexism. (In other studies, they've found that white women asking for charity donations could pull more than twice as much donation money, an average, than other race/gender categories. So, pro-female advantages are not unheard of.)