Why so many PhDs?

[ZeeMore21]

That's why there are very few regression models that only use one explanatory variable. The signalling model doesn't use just one predictor, but weights a number of predictors accordingly (i.e., quantity and however you would like to "quantify" quality of research experience is likely to trump GPA for Ph.D. admissions). Also, there is noise within signalling models, which is why Signal Detection Theory was proposed; essentially it can become a Bayesian statistics problem as you increase the amount of trials, gaining information between variation in the predictors (i.e., the ranges of accepted individuals with various stats/grades/etc.), and their success in the program => if you notice that students with a balanced (fairly high) GPA along with a thorough research background seem to do well, but those with high GPA+no research nor high research+very low GPA don't do well, then you'll begin incorporating this in your model to find 'ideal' candidates; if there's a way to quantify SOP quality, too, for example, by assigning a grade, a binomial (good vs. bad) descriptor, or by using a Likert scale, this can, too, be translated into a regression model.

Like I've said before, there is no cut and dry formula. Regardless of the formula you propose, I'm sure many humanities admission committee are happy with the way they do things. They are qualified enough to know who belongs in their programs or not without fussing over a ridiculous formula that isn't fool proof.

[Behavioral]

Like I've said before, there is no cut and dry formula. Regardless of the formula you propose, I'm sure many humanities admission committee are happy with the way they do things. They are qualified enough to know who belongs in their programs or not without fussing over a ridiculous formula that isn't fool proof.

Who the heck said it is a discrete formula? Regressions are essentially how all estimates are made, whether you use an actual formula or something that's similar to a formula. Do you think anyone actually writes out a formula to say, "Yes, we will admit this person."? That is NOT what I proposed. A regression is a POST-HOC analysis (i.e., DONE AFTER decisions have already been made as a way to explain what predictors may have the most explanatory value in a decision). A regression can be used to test predictions if they are empirically established to see whether or not a theory holds true. Very rarely does one EVER begin to actually USE a regression model as a diagnostic for whatever was tested.

Also, regressors are ESTIMATORS. Any person in a quantitative field could only wish they could find a regression model with an r^2 value of 1.00. How a regression model is estimated is by aggregating past decision criteria along with the result and then finding some least-sqaures (variance-minimizing estimator); i.e.:

Candidate 1: 4.0 GPA, Good research experience, 1600 GRE - Result Accept

Candidate 2: 3.5 GPA, Good research experience, 1300 GRE - Result Accept

Candidate 3: 4.0 GPA, No research experience, 1500 GRE - Result Reject

...etc

This would be a case of a binomial distribution regression model, which could also be called a linear probability model (since you can only be rejected or accepted and not something in between [i.e., you're half accepted], thus 0 = 0% and 1 = 100%)

Anyway, I'm done here. If you have problems with what I said earlier, I already apologized and tried to qualify my claims by saying that I don't believe that Masters-degree holders are "second-rate" as some of you claimed I was implying. Apparently there are 0 other posters here with any kind of economics background since I seem to have to explain every one of my posts, so all I'm doing is clogging up this thread with needless tangents.

[robot_hamster]

Who the heck said it is a discrete formula? Regressions are essentially how all estimates are made, whether you use an actual formula or something that's similar to a formula. Do you think anyone actually writes out a formula to say, "Yes, we will admit this person."? That is NOT what I proposed. A regression is a POST-HOC analysis (i.e., DONE AFTER decisions have already been made as a way to explain what predictors may have the most explanatory value in a decision). A regression can be used to test predictions if they are empirically established to see whether or not a theory holds true. Very rarely does one EVER begin to actually USE a regression model as a diagnostic for whatever was tested.

Also, regressors are ESTIMATORS. Any person in a quantitative field could only wish they could find a regression model with an r^2 value of 1.00. How a regression model is estimated is by aggregating past decision criteria along with the result and then finding some least-sqaures (variance-minimizing estimator); i.e.:

Candidate 1: 4.0 GPA, Good research experience, 1600 GRE - Result Accept

Candidate 2: 3.5 GPA, Good research experience, 1300 GRE - Result Accept

Candidate 3: 4.0 GPA, No research experience, 1500 GRE - Result Reject

...etc

This would be a case of a binomial distribution regression model, which could also be called a linear probability model (since you can only be rejected or accepted and not something in between [i.e., you're half accepted], thus 0 = 0% and 1 = 100%)

Anyway, I'm done here. If you have problems with what I said earlier, I already apologized and tried to qualify my claims by saying that I don't believe that Masters-degree holders are "second-rate" as some of you claimed I was implying. Apparently there are 0 other posters here with any kind of economics background since I seem to have to explain every one of my posts, so all I'm doing is clogging up this thread with needless tangents.

I just now read through all of these latest posts. I understand what you are trying to get at.

[ktel]

I thought Behavioral's posts were mostly clear. While some of the more technical aspects were a little confusing, it's fairly obvious that admissions committees don't actually use these formulas, however these formulas may explain their decision-making process.

The whole reason why I started this thread is because I personally think the idea of launching straight into a PhD is strange. I want to have my Master's as a buffer to determine if research and academia is really the path I want to go down. I have also heard that while a Master's degree may open a lot of doors career-wise, a PhD will promptly close many of them. You simply become over-qualified for a lot of jobs. This is coming from the engineering background, which is obviously a much more practical and applied field. I have also noticed that the idea of further education past an undergrad degree is somewhat loathed by my peers. Yet I have also noticed that many of their career aspirations would be mostly unattainable without getting at least a Master's degree.

[ZeeMore21]

Mistake, see below

Edited July 3, 2011 by ZeeMore21

[ZeeMore21]

This isn't about whether Behavioral's formula makes sense or not. He/she is not in the humanities, so I do have the right to counter what has been argued. What he/she is arguing is that committees create a formula based on their admitted applicants and use this formula when the next admissions cycle comes around. I am not in the sciences or social sciences, but I can speak on behalf of humanities since I have learned from professors what they look for when looking at humanities applications.

What I am arguing, and will continue to argue is that there is no formula used to make predictions about what applicants will be chosen the next round. I am speaking strictly about humanities. If there was a formula used, then admissions results would be more predictable. However, we find that these results are not predictable. This is explained because admissions committees rotate yearly, and criteria for admissions also changes. For example, one year, an admissions committee might be looking for more American History scholars, but in another different year, they may be looking for Middle Eastern history scholars. Academic politics, whether one likes it or not, also may have something to do with who gets admitted, and this factor would not be addressed in a formula. There are way too many factors that go in the admissions process, and I see no way how a formula could be created that could be relied on to make future decisions...I actually think it would be irresponsible for a prediction formula to be used. For something arbitrary such as admissions, I disagree that prediction formulas are used, anyone who has sat on a humanities admissions committee can feel free to correct me.

Sorry ktel, that this thread turned into this type of debate. I thought that the thread was pretty informative until I came upon the post about Master students. From the humanities perspective, I believe many go into doctorate programs because a Masters degree usually isn't enough. I am in English, for example, and most English majors choose to go into PhDs because most decide to teach. I guess it depends on the field.

Edited July 3, 2011 by ZeeMore21

[ktel]

I think you and Behavioral are arguing similar things. There was no claim on his part that admissions committees actually use a physical formula to make their decisions. From what I can understand from his difficult-to-read posts (having no knowledge of economics), is that AFTER the fact it is possible to make a 'formula' based on the results that may somewhat explain what their decision making process was. The decision has already been made when this formula is constructed, and it only applies to that set of decisions. He can correct me if I've misunderstood this

[Behavioral]

I think you and Behavioral are arguing similar things. There was no claim on his part that admissions committees actually use a physical formula to make their decisions. From what I can understand from his difficult-to-read posts (having no knowledge of economics), is that AFTER the fact it is possible to make a 'formula' based on the results that may somewhat explain what their decision making process was. The decision has already been made when this formula is constructed, and it only applies to that set of decisions. He can correct me if I've misunderstood this

You're correct. There's a reason I bolded "POST-HOC" on my last post. I also reiterated that point by saying that the regression that's created isn't used as a diagnostic (meaning it isn't used AS a determining formula for admission), but rather it's a way to empirically see what variables are weighted most in successful admissions.

I think there's just a discord in technical issues. I'm talking about a fairly technical statistical technique with someone who doesn't necessarily have to know anything about this to get into graduate school (I wouldn't suspect him/her to assume I know anything about English or some methodology in the humanities that isn't employed in the social sciences, for example).

[ZeeMore21]

I think you and Behavioral are arguing similar things. There was no claim on his part that admissions committees actually use a physical formula to make their decisions. From what I can understand from his difficult-to-read posts (having no knowledge of economics), is that AFTER the fact it is possible to make a 'formula' based on the results that may somewhat explain what their decision making process was. The decision has already been made when this formula is constructed, and it only applies to that set of decisions. He can correct me if I've misunderstood this

I'm honestly not trying to be combative or belabor this subject, but I honestly just don't agree, and I should be able to do so without being voted down. And again, despite how technical Behavioral's formula sounds, at the end of the day, his/her argument is that formulas can be used to reflect on an admission commitee's practices. I still don't believe that a formula could be created given the arbitrary decisions that go in selecting candidates, I listed some of them in my last post. And here I am talking about formulas that could be used to explain how a committee chose their applicants. How could a formula be created that takes academic politics into consideration for example? From the example Behavioral used, he/she only listed GPA, good/bad research experience, and GRE scores. There are many other factors that go into admission decisions that can't be covered within a formula, and that is what I've been trying to argue all along. I don't really understand what I am arguing as being difficult, and I'm sure others would agree. I do wish there were other humanities posters who could put in their two cents because I'm sure I am not alone in my opinion.

I have no problem with Behavioral discussing a formula, perhaps other fields use this. But, I do think I can speak on behalf of humanities, that admissions decisions is nuanced, and there is no formula that could explain how and why applicants were chosen. Admissions is made on an individual basis...applications materials can be weighted differently depending on whose applications is being reviewed. With this in mind, there is no cookie cutter formula. That's all I really have to say about this subject.

And to be honest, I do not even understand how formulas came up in this thread, seems off-topic given what the OP's questions was.

Edited July 3, 2011 by ZeeMore21

[wtncffts]

I'm honestly not trying to be combative or belabor this subject, but I honestly just don't agree, and I should be able to do so without being voted down. And again, despite how technical Behavioral's formula sounds, at the end of the day, his/her argument is that formulas can be used to reflect on an admission commitee's practices. I still don't believe that a formula could be created given the arbitrary decisions that go in selecting candidates, I listed some of them in my last post. And here I am talking about formulas that could be used to explain how a committee chose their applicants. How could a formula be created that takes academic politics into consideration for example? From the example Behavioral used, he/she only listed GPA, good/bad research experience, and GRE scores. There are many other factors that go into admission decisions that can't be covered within a formula, and that is what I've been trying to argue all along. I don't really understand what I am arguing as being difficult, and I'm sure others would agree. I do wish there were other humanities posters who could put in their two cents because I'm sure I am not alone in my opinion.

I have no problem with Behavioral discussing a formula, perhaps other fields use this. But, I do think I can speak on behalf of humanities, that admissions decisions is nuanced, and there is no formula that could explain how and why applicants were chosen. Admissions is made on an individual basis...applications materials can be weighted differently depending on whose applications is being reviewed. With this in mind, there is no cookie cutter formula. That's all I really have to say about this subject.

And to be honest, I do not even understand how formulas came up in this thread, seems off-topic given what the OP's questions was.

I think this is a simple case of disagreement about the value of quantitative methods of inquiry. It's quite clear, though not in so many words, that ZeeMore simply does not believe that any regression model could capture the complexity of the admissions process enough to have any degree of predictive power. Behavioral has a much greater faith in quantitative measures, which is unsurprising given his background. Although I am generally a qualitative person in an increasingly quantitative field (political science), I do think that regression models can be tremendously helpful, and in the case of graduate admissions, I'm inclined to think a good model can be constructed (perhaps already has), which will give statistically significant and meaningful results. ZeeMore, when you talk about various things which can't be 'covered within a formula', you'd be surprised. Anything can be measured and quantified. There are, of course, all sorts of validity issues which are the stuff of endless argument in the social sciences, but, as they say, the proof of the pudding is in the eating. Quantitative models just are; they're neither good nor bad. It's up to the author to construct models which would be convincing to their colleagues.

I think you're overstating the arbitrary and individualized nature of the admissions process just a bit. You're right that, in the actual decisions process, each application is considered individually. That doesn't mean that, in the aggregate, patterns don't or won't show up. Surely, the admissions process isn't that chaotic. A regression model in political science, for example, might show that region and religious affiliation are heavily correlated with vote choice, which tells us something interesting and prompts further, perhaps qualitative, research. But that doesn't mean that any particular individual, say, a born-again Christian from Alberta (if you're Canadian, you'll know what I'm talking about), can't 'buck the trend', as it were. Statistical methods are about overall, aggregate patterns, probabilities, and likelihoods. I don't see that graduate admissions, in any discipline, would be wholly immune from such measurements.

And just to repeat his own defense, Behavioral isn't saying that admissions committees actually make decisions using formulas or regression models. In the first instance, he was simply outlining a theory from economics which might explanatorily capture the specific case of MA to PhD, etc. His point is that regression models might be used by an outside observer to analyze patterns in graduate admissions. In addition, it has nothing to do with field, except, perhaps, in so far as some fields rely on explicitly quantitative measures (GPA, GRE, LSAT) more than others, which may mean regression models which are more convincing. In this, I don't think there's a strict sciences/humanities divide; I've been told that the GRE is very important in philosophy, for instance, as is the LSAT to law school, while the somewhat vague notion of 'research experience' in the sciences isn't explicitly amenable to measurement.

[ZeeMore21]

"His point is that regression models might be used by an outside observer to analyze patterns in graduate admissions. In addition, it has nothing to do with field, except, perhaps, in so far as some fields rely on explicitly quantitative measures (GPA, GRE, LSAT) more than others, which may mean regression models which are more convincing."

I'm not sure how an outside observer would be able to fully understand how a specific admissions committee ticks unless he or she is part of the admissions committee in question. Also, the religion affiliation example you are using wtncffts doesn't really work....there are things that could be due to coincidence...how would be you create a formula based on that? Also, I'm not sure that a formula could be created for a whole field in general, departments may carry out their admissions process very differently even though they may be under the same field. Again, I am not saying that Behavioral is suggesting that formulas may be created to be used as predictors...he clarified this part so I am no longer countering that.

I think the only reason why I am passionate when it comes to this specific subject of admissions and formulas is because I rather not have people stress over trying to come up with formulas that would explain a committee's admission decisions instead of just trying to do the best they can with their applications. Regardless with how efficient some may think formulas are, they are not perfect, and tailoring one's application around a specific formula does not necessarily mean you will have success. All I've seen are proponents for admissions formulas, but no one has stated how unreliable they may be.

Edited July 3, 2011 by ZeeMore21

[Behavioral]

"His point is that regression models might be used by an outside observer to analyze patterns in graduate admissions. In addition, it has nothing to do with field, except, perhaps, in so far as some fields rely on explicitly quantitative measures (GPA, GRE, LSAT) more than others, which may mean regression models which are more convincing."

I'm not sure how an outside observer would be able to fully understand how a specific admissions committee ticks unless he or she is part of the admissions committee in question. Also, the religion affiliation example you are using wtncffts doesn't really work....there are things that could be due to coincidence...how would be you create a formula based on that? Also, I'm not sure that a formula could be created for a whole field in general, departments may carry out their admissions process very differently even though they may be under the same field. Again, I am not saying that Behavioral is suggesting that formulas may be created to be used as predictors...he clarified this part so I am no longer countering that.

I think the only reason why I am passionate when it comes to this specific subject of admissions and formulas is because I rather not have people stress over trying to come up with formulas that would explain a committee's admission decisions instead of just trying to do the best they can with their applications. Regardless with how efficient some may think formulas are, they are not perfect, and tailoring one's application around a specific formula does not necessarily mean you will have success. All I've seen are proponents for admissions formulas, but no one has stated how unreliable they may be.

I, nor anyone else, has stated that they're flawless. I even qualified my support for regression models by saying there is quite a bit of variability (i.e., not all decisions are based on some homogeneous criterion) and noise (imperfect information between applicant and admissions committee).

There might still be quite a bit of misunderstanding about what a regression is; it's not something that uses intuitively-simple numbers to explain data; the most common use for them (which is how I would posit one would use for admissions) is to assign relative weights (i.e., relative importance) between independent variables onto a dependent variable. If one can come up with a face-valid way of assigning some quantitative value to a qualitative/subjective criteria (let's say a writing sample) and say that one could rate/grade it on a scale from 1 to 5, and we see that those scoring a 5 almost always receive admission regardless of other merits (GPA, GRE, etc.), we would see a large weight attached to writing sample, for instance.

Taken in concert, there is no one variable that determines admission--nor two, nor three. However, if you can mine relevant data about factors that affect admissions (major, statement of purpose, GPA, GRE, quality of letters of rec, work experience, age, research experience, publications, presentations/talks, graduate degrees, etc.; it can go on for a while if you can argue there's a meaningful correlation) and see the differentials associated with each and their impact on decision outcomes. This isn't perfect. This isn't even causal. It's quite reasonable to say that a high GPA doesn't cause an adcomm to take more notice of your application and subsequently admit you. That's not what's being implied by the use of a regression model. What a regression (in this case) tries to do is unearth trends and patterns over an aggregate of past outcomes--if those patterns and trends don't hold true in the present or future, an updated regression will show insignificant results (i.e., there will be no weights attached to the independent variables).

It's very true that statistics could be used to lie or manipulate information. If you wanted to market favorable results, you can easily choose to omit counterexamples and outliers from your dataset and get entirely different results. I'm choosing to omit that possibility since we're talking in rhetoric about the 'possible' use of a regression model if such data were available. Aside from trying to only use GPA and GRE from the results page on this site, I doubt there's a compendium of data and results from any university that would allow for such an analysis. If there were, I'm sure the results would be interesting, no matter what the conclusion is.

[wtncffts]

I'm not sure how an outside observer would be able to fully understand how a specific admissions committee ticks unless he or she is part of the admissions committee in question. Also, the religion affiliation example you are using wtncffts doesn't really work....there are things that could be due to coincidence...how would be you create a formula based on that? Also, I'm not sure that a formula could be created for a whole field in general, departments may carry out their admissions process very differently even though they may be under the same field. Again, I am not saying that Behavioral is suggesting that formulas may be created to be used as predictors...he clarified this part so I am no longer countering that.

I think the only reason why I am passionate when it comes to this specific subject of admissions and formulas is because I rather not have people stress over trying to come up with formulas that would explain a committee's admission decisions instead of just trying to do the best they can with their applications. Regardless with how efficient some may think formulas are, they are not perfect, and tailoring one's application around a specific formula does not necessarily mean you will have success. All I've seen are proponents for admissions formulas, but no one has stated how unreliable they may be.

I'm not sure why your posts keep getting voted down; there's nothing objectionable in them. But that's another matter.

The first point: that's up to the observer. In many cases, he won't even want to 'fully understand' in the sense I think you mean. In my field and in other social sciences, there are 'large-N' and 'small-N' studies. The former takes a large number of observations and tests a particular hypothesis through the use of statistical methods such as regression. 'Small-N' case studies try to examine in close detail only a few cases, looking for the actual causal mechanisms. To an observer who is examining this question of determinants of graduate admissions success, he can do either, or preferably, both.

As far as the religious affiliation example, I'm not sure what you mean by 'doesn't really work'. It's about as typical an empirical result in a whole field of political science, 'political behaviour', as any. What do you mean by coincidence? The point of a regression model, as with other statistics, is to minimize the probability that the results would occur by chance. When they report opinion polls, as you probably know, they say something like "the results are accurate plus or minus 3 percent, 19 times out of 20". Similar thing. Believe me, I'm not trying to belittle you, but perhaps you could at least look into basic statistics and simple linear regression, which isn't all that difficult to learn. I'm not a quantitative whiz either, that's for sure.

You may indeed be right that a model which seems to apply well to one situation will not in another. That happens all the time. If that's considered a problem, then it's with the model. As I said, the proof of the success of a model depends on what the modeller wants to do with it and whether colleagues are convinced or persuaded by its validity and usefulness. It's not a matter of coming up with a 'cookie cutter formula' and pronouncing that it solves all problems; it's more that one constructs a model, tests hypotheses with it, writes up the results, and gets feedback. It's a dialogue.

The second point: I absolutely agree that applicants needn't and oughtn't to spend their time coming up with a statistical model of the admissions process, unless, perhaps, that actually happens to be their research interest and they're applying in a suitable field (wouldn't that be an interesting writing sample?). This would be something that professional academics would do, for curiosity's sake, perhaps with real-world implications. I think the data-gathering for such a study would be one of the more difficult aspects, but it's certainly possible.

[ZeeMore21]

I have my opinion on who is voting them down but it doesn't matter. Anyway, thanks wtncffts for clarifying what is going on here...it's a little off-putting when someone is purposefully trying to run circles around you with overly technical material instead of trying to be understood or answer your questions. And also, thanks for seeing where I am going with my posts. I have had college level statistics, so I do know, on a general basis, what you and other posters are getting at. I think I am more focused on the real-life applications/consequences of a regression model and I am not really focused on the technical aspects. It seems like some people are trying to attack me on the technical side more than seeing that I am coming from more of a humanistic perspective.

No one has really explained what these regression models would be used for. Wtncffts, you did point out that maybe college administrators would look at these regression models...for some reason, I just have a hard time supporting that. This is probably because I am coming from a liberal education...I tend to want to judge things on a qualitative and individual basis. I am also in English, a field that judges things qualitatively. So again, I'm not trying to be combative just to be combative...I really do have my reasons, and they are just as valuable.

Edited July 4, 2011 by ZeeMore21

[Behavioral]

I have my opinion on who is voting them down but it doesn't matter. Anyway, thanks wtncffts for clarifying what is going on here...it's a little off-putting when someone is trying to run circles around you with overly technical material instead of trying to be understood. And also, thanks for seeing where I am going with my posts. I have had college level statistics, so I do know, on a general basis, what you and other posters are getting at. I think I am more focused on the real-life applications/consequences of a regression model and I am not really focused on the technical aspects. It seems like some people are trying to attack me on the technical side more than seeing that I am coming from more of a humanistic perspective.

No one has really explained what these regression models would be used for. Wtncffts, you did point out that maybe college administrators would look at these regression models...for some reason, I just have a hard time supporting that. This is probably because I am coming from a liberal education...I tend to want to judge thimgs on a qualitative and individual basis. I am also in English, a field that judges things qualitatively. So again, I'm not trying to be combative just to be combative...I really do have my reasons, and they are just as valuable.

I like how you dismiss my posts entirely. And about my being clear -- a number of posters came in unsolicited stating that they understood my argument just fine. If you're trying to create a straw man, it isn't working.

There might still be quite a bit of misunderstanding about what a regression is; it's not something that uses intuitively-simple numbers to explain data; the most common use for them (which is how I would posit one would use for admissions) is to assign relative weights (i.e., relative importance) between independent variables onto a dependent variable. If one can come up with a face-valid way of assigning some quantitative value to a qualitative/subjective criteria (let's say a writing sample) and say that one could rate/grade it on a scale from 1 to 5, and we see that those scoring a 5 almost always receive admission regardless of other merits (GPA, GRE, etc.), we would see a large weight attached to writing sample, for instance.

Also I tried making my posts less technical by providing less technical means to show what I was talking about (example data points). If I wanted to be LESS clear, I could have thrown Mathematica and LaTeX notation in there instead of trying to include examples and explanations.

And for someone entering school for an English Ph.D., you aren't that careful a reader. wtncffts didn't mention "college administrators"--he/she mentioned professional academics (i.e., professors, researchers, scientists). There is a rich tradition in economists looking at trends in education over the past half-century and this question is something that many economists would probably like to address, but haven't the access to data to be able to write a meaningful paper about.

[ZeeMore21]

But I do apologize to the OP, I'm basically just responding to tangents. Hopefully we can get back to the question about the quantity of PhDs and what may attribute to that.

[ktel]

I really could care less that this has gone off-topic. That's how conversations work. I think much of the reasoning behind the differences between Bachelor's to PhD vs Bachelor's to Master's to PhD has been explained as mostly an American thing and more common in humanities fields.

[ZeeMore21]

I like how you dismiss my posts entirely. And about my being clear -- a number of posters came in unsolicited stating that they understood my argument just fine. If you're trying to create a straw man, it isn't working.

Also I tried making my posts less technical by providing less technical means to show what I was talking about (example data points). If I wanted to be LESS clear, I could have thrown Mathematica and LaTeX notation in there instead of trying to include examples and explanations.

And for someone entering school for an English Ph.D., you aren't that careful a reader. wtncffts didn't mention "college administrators"--he/she mentioned professional academics (i.e., professors, researchers, scientists). There is a rich tradition in economists looking at trends in education over the past half-century and this question is something that many economists would probably like to address, but haven't the access to data to be able to write a meaningful paper about.

Instead of trying to understand what point I'm trying to make, you have often chosen just to combat me with more technical information. I can respect the way posters like wtncffts have responded to my posts.

[ZeeMore21]

I like how you dismiss my posts entirely. And about my being clear -- a number of posters came in unsolicited stating that they understood my argument just fine. If you're trying to create a straw man, it isn't working.

Also I tried making my posts less technical by providing less technical means to show what I was talking about (example data points). If I wanted to be LESS clear, I could have thrown Mathematica and LaTeX notation in there instead of trying to include examples and explanations.

And for someone entering school for an English Ph.D., you aren't that careful a reader. wtncffts didn't mention "college administrators"--he/she mentioned professional academics (i.e., professors, researchers, scientists). There is a rich tradition in economists looking at trends in education over the past half-century and this question is something that many economists would probably like to address, but haven't the access to data to be able to write a meaningful paper about.

I misspoke when I said college administrators, and I do apologize for that. Please, however, don't question my abilities as an English PhD.....that is uncalled for. I have yet to attack your credentials as far as your field is concerned. I genuinely would like to know how these regression models would be used by professional academics. If they wouldn't be used to predict admissions for future academics, how else would these models be used? I do know that everyone is basing there arguments on hypotheticals and I guess my concern lies with the consequences of such models for future applicants.

[Behavioral]

Instead of trying to understand what point I'm trying to make, you have often chosen just to combat me with more technical information. I can respect the way posters like wtncffts have responded to my posts.

I put in words what a regression tries to measure. You seemed to have missed that.

You say there's no merit to using quantitative measures in a setting such as academic admissions -- I argue there is.

Math is a tool that helps one be concise. If I had to explain what regression analysis is every time I bring it up, I'd be inclined to never post about it again. The fact that it seems that most people understand what I'm talking about reinforces me to continue on a set number of assumptions (one being that most people already have a good working knowledge of regressions, and those who don't -- I already wrote in a few posts what the basic logic and reasoning behind the use of regression analysis is).

Also, this is the Internet. If you think this is combative, then you haven't been to many forums.

[Behavioral]

I misspoke when I said college administrators, and I do apologize for that. Please, however, don't question my abilities as an English PhD.....that is uncalled for. I have yet to attack your credentials as far as your field is concerned. I genuinely would like to know how these regression models would be used by professional academics. If they wouldn't be used to predict admissions for future academics, how else would these models be used? I do know that everyone is basing there arguments on hypotheticals and I guess my concern lies with the consequences of such models for future applicants.

Here are a few examples that I can come up with:

On a sociological standpoint:

Differences of weighting structures in academic admissions processes. i.e., Is there a difference in the "ideal" applicant in the US vs. elsewhere? If so, why? Culture? History/Precedence?

On an economic standpoint:

Differences in threshold/acceptance rates between public and private institutions. i.e., Does more funding (i.e., higher capacity to admit students) decrease/mediate the weights in admissions?

On a meta-academic standpoint:

Why do we weight things such as cumulative GPA if non-relevant courses are not good indicators of graduate success? i.e., What are the normative standards a committee should acknowledge to combat increasing fail/dropout/attrition rates?

[ZeeMore21]

I put in words what a regression tries to measure. You seemed to have missed that.

You say there's no merit to using quantitative measures in a setting such as academic admissions -- I argue there is.

Math is a tool that helps one be concise. If I had to explain what regression analysis is every time I bring it up, I'd be inclined to never post about it again. The fact that it seems that most people understand what I'm talking about reinforces me to continue on a set number of assumptions (one being that most people already have a good working knowledge of regressions, and those who don't -- I already wrote in a few posts what the basic logic and reasoning behind the use of regression analysis is).

Also, this is the Internet. If you think this is combative, then you haven't been to many forums.

Again, like I have argued in other posts, you are free to argue for formulas and the quantitative. No one is disputing that. Yet, when I choose to argue against the use of the quantitative as far as admissions decisions is concerned, I am seen as not understanding regression models and what they are. Like I said before, I have studied statistics...it is not like I am an alien to this field. You have to understand that people like me may have serious qualms about quantitative research without assuming that they are just ignorant.

[Behavioral]

Again, like I have argued in other posts, you are free to argue for formulas and the quantitative. No one is disputing that. Yet, when I choose to argue against the use of the quantitative as far as admissions decisions is concerned, I am seen as not understanding regression models and what they are. Like I said before, I have studied statistics...it is not like I am an alien to this field. You have to understand that people like me may have serious qualms about quantitative research without assuming that they are just ignorant.

We can agree, then.

I'm headed into a more qualitative field than what my initial trajectory was going towards because I do agree that knowledge isn't just a mathematical animal. There are nuances at play with everything that we encounter. Within a subject that is interested in people and their decision-making (my research), I know that people are dynamic creatures and they don't behave in a predictable way at all times.

If you have reservations about the usefulness of regressions in applied settings, that's fine--but as someone who does have respect for their potential use, I will try to the best of my knowledge to convince you otherwise. Math doesn't tell the whole story, but it can enlighten new information and support theories that may encapsulate that story.

[ZeeMore21]

Also, perhaps I just will take a break from this post and see what other people have to say. Behavioral, sorry for being harsh with my posts, I am probably just still on the defensive after your earlier post on the ability of M.A students as I took it very personally. Though I genuinely do have my qualms about solely relying on the quantitative, I'm sure I could have argued my position less harshly. Taking a break.

[Behavioral]

Also, perhaps I just will take a break from this post and see what other people have to say. Behavioral, sorry for being harsh with my posts, I am probably just still on the defensive after your earlier post on the ability of M.A students as I took it very personally. Though I genuinely do have my qualms about solely relying on the quantitative, I'm sure I could have argued my position less harshly. Taking a break.

Same. I think we both deserve to spend a little more time away from the computer for the time being haha