Why the science of intelligence matters: affirmative action for college admission but not for graduation

[bilby]

Correlations everywhere.

The hallmark of pseudoscience - the belief that correlation implies causation.

Correlation is not causation. But, an essential PART of every case for causation is correlation. Without correlations, there is no causation. It is not just a part of pseudoscience but a part of ALL science. If you don't believe me, then try making a rational case for a causal link between two variables without correlations. Maybe you think correlations should be omitted from science completely, in which case you would be merely a philosophical postmodernist, where all theoretical thought has nothing to do with patterns of observation.

Correlation is a part of all science; but it appears to be all of this part of science. Making it pseudoscience.

I don't have a problem with arguments containing correlations at all. I do have a problem when an argument is all correlations.

If you have something else, present it.

 

[ApostateAbe]

Correlation is not causation. But, an essential PART of every case for causation is correlation. Without correlations, there is no causation. It is not just a part of pseudoscience but a part of ALL science. If you don't believe me, then try making a rational case for a causal link between two variables without correlations. Maybe you think correlations should be omitted from science completely, in which case you would be merely a philosophical postmodernist, where all theoretical thought has nothing to do with patterns of observation.

Correlation is a part of all science; but it appears to be all of this part of science. Making it pseudoscience.

I don't have a problem with arguments containing correlations at all. I do have a problem when an argument is all correlations.

If you have something else, present it.

My argument really does not involve any causal link, and the causes of the racial intelligence gaps don't matter, so correlations really are all that matter here. You can't make causal relationships based on correlations alone, but you can make PREDICTIONS based on correlations alone. I will summarize the argument so it is clear:

(1) Students of higher test scores and higher grades are more likely pass their courses, and conversely students of lower test scores and lower grades are less likely to pass their courses.
(2) Students of disadvantaged groups are disproportionately admitted without meeting the typical requirements of test scores and grades.
(3) Therefore, students of disadvantaged minority groups are less likely to pass their courses.
(4) This prediction is confirmed with data of graduation rates.

 

[Axulus]

Correlation is not causation. But, an essential PART of every case for causation is correlation. Without correlations, there is no causation. It is not just a part of pseudoscience but a part of ALL science. If you don't believe me, then try making a rational case for a causal link between two variables without correlations. Maybe you think correlations should be omitted from science completely, in which case you would be merely a philosophical postmodernist, where all theoretical thought has nothing to do with patterns of observation.

Correlation is a part of all science; but it appears to be all of this part of science. Making it pseudoscience.

I don't have a problem with arguments containing correlations at all. I do have a problem when an argument is all correlations.

If you have something else, present it.

So it is your belief that college admissions exams should be abolished as psuedoscience?

 

[Nice Squirrel]

If disadvantaged races were NOT more likely to drop out following from a confirmed a higher rate of failing grades, then that would actually reflect a systemic educational failure. Unless you tell me that college professors practice affirmative action when assigning grades (or whatever you prefer to call it).

I pointed out the real reasons I hear for dropping out of college and rarely do they do not involve grades, they typically involve socioeconomic and relationship factors. You claim to have data about college grades and race, please post it. Because leaving college does not mean the student left because of low grades.

- - - Updated - - -

Correlation is a part of all science; but it appears to be all of this part of science. Making it pseudoscience.

I don't have a problem with arguments containing correlations at all. I do have a problem when an argument is all correlations.

If you have something else, present it.

So it is your belief that college admissions exams should be abolished as psuedoscience?

I would like to see them replaced and the GRE abolished.

 

[bilby]

Correlation is a part of all science; but it appears to be all of this part of science. Making it pseudoscience.

I don't have a problem with arguments containing correlations at all. I do have a problem when an argument is all correlations.

If you have something else, present it.

My argument really does not involve any causal link, and the causes of the racial intelligence gaps don't matter, so correlations really are all that matter here.

Hence, Pseudoscience.

 

[bilby]

Correlation is a part of all science; but it appears to be all of this part of science. Making it pseudoscience.

I don't have a problem with arguments containing correlations at all. I do have a problem when an argument is all correlations.

If you have something else, present it.

So it is your belief that college admissions exams should be abolished as psuedoscience?

Yes.

Other countries seem to do as well (or better) testing college applicants based on their preparatory knowledge in the relevant foundation subjects, without looking at any measure of 'general intelligence' at all.

This is another area where the USA is wedded to woo, and it is so ingrained that many USAians don't even realise that it is just woo.

Lie detectors and Christianity are other examples.

 

[Axulus]

So it is your belief that college admissions exams should be abolished as psuedoscience?

Yes.

Other countries seem to do as well (or better) testing college applicants based on their preparatory knowledge in the relevant foundation subjects, without looking at any measure of 'general intelligence' at all.

This is another area where the USA is wedded to woo, and it is so ingrained that many USAians don't even realise that it is just woo.

Lie detectors and Christianity are other examples.

And how do you determine their "preparatory knowledge in relevant foundation subjects"? The SAT tests for math, critical reading, and writing skills. Do you not consider those subjects to be relevant foundation subjects?

 

[bilby]

Yes.

Other countries seem to do as well (or better) testing college applicants based on their preparatory knowledge in the relevant foundation subjects, without looking at any measure of 'general intelligence' at all.

This is another area where the USA is wedded to woo, and it is so ingrained that many USAians don't even realise that it is just woo.

Lie detectors and Christianity are other examples.

And how do you determine their "preparatory knowledge in relevant foundation subjects"? The SAT tests for math, critical reading, and writing skills. Do you not consider those subjects to be relevant foundation subjects?

Not really, no. They are too basic at the level of tertiary education admissions in the 21st century. They are (in the case of critical reading and writing skills, literally) trivial.

 

[Axulus]

And how do you determine their "preparatory knowledge in relevant foundation subjects"? The SAT tests for math, critical reading, and writing skills. Do you not consider those subjects to be relevant foundation subjects?

Not really, no. They are too basic at the level of tertiary education admissions in the 21st century. They are (in the case of critical reading and writing skills, literally) trivial.

What do you mean by "too basic"? You seem to have a bizzare standard based on personal preference: the test should be about "foundation" knowledge (which is the definition of basic knowledge), but not "too basic".

Which test method/subjects do you consider "foundation knowledge" that is not also "too basic"? Please cite your sources that this is a superior method in predicting who will graduate college.

 

[ApostateAbe]

My argument really does not involve any causal link, and the causes of the racial intelligence gaps don't matter, so correlations really are all that matter here.

Hence, Pseudoscience.

No argument with a causal link, therefore pseudoscience? How do you define "pseudoscience" exactly? I made a suggestion to the admins to merge the three Science subforums together, because "pseudoscience" seems to be not just bad science but anything that disagrees too much with the politics of a moderator.

 

[AthenaAwakened]

AP tests, by subject, are better predictors of college performance than the SATs. So are GPAs.

As are family income, family address, and quite possibly shoe size.

 

[Axulus]

AP tests, by subject, are better predictors of college performance than the SATs. So are GPAs.

As are family income, family address, and quite possibly shoe size.

Which is why SAT scores shouldn't be considered in isolation, which they aren't, but rather used as a supplemental data point.

 

[ApostateAbe]

AP tests, by subject, are better predictors of college performance than the SATs. So are GPAs.

As are family income, family address, and quite possibly shoe size.

You don't have to speak purely from your gut. The data is out there. You are right that AP test scores are better predictors than the SAT of college GPA, per Table 2 and Page 10 of this study:

http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.434.6711&rep=rep1&type=pdf

...and you are right about high-school GPAs. Family income... not so much. The Appendix on Page 14 of this study shows correlations among college GPA, high school GPA, SAT, and SES (socioeconomic status):

https://research.collegeboard.org/s...mic-status-sat-freshman-gpa-analysis-data.pdf

But you got me on shoe size. I can't find the needed correlation data for that assertion of possibility (not that I have looked).

 

[bilby]

Hence, Pseudoscience.

No argument with a causal link, therefore pseudoscience? How do you define "pseudoscience" exactly? I made a suggestion to the admins to merge the three Science subforums together, because "pseudoscience" seems to be not just bad science but anything that disagrees too much with the politics of a moderator.

It just seems odd that you are so enthusiastic about the connection between college drop-out rates and intelligence, when you should be worrying about the alarming number of bed-sheet casualties caused by cheese consumption; or the dangerous effects of sour-cream on motorcyclists.

Obviously it would be a derail to discuss those here; but if you want to start some new threads on those topics, pseudoscience might be the place for them.

If correlations are all that matter, then those life and death dairy produce issues are far more important than the rather lower correlation claimed between intelligence and college graduation rates.

 

[ApostateAbe]

No argument with a causal link, therefore pseudoscience? How do you define "pseudoscience" exactly? I made a suggestion to the admins to merge the three Science subforums together, because "pseudoscience" seems to be not just bad science but anything that disagrees too much with the politics of a moderator.

It just seems odd that you are so enthusiastic about the connection between college drop-out rates and intelligence, when you should be worrying about the alarming number of bed-sheet casualties caused by cheese consumption; or the dangerous effects of sour-cream on motorcyclists.

Obviously it would be a derail to discuss those here; but if you want to start some new threads on those topics, pseudoscience might be the place for them.

If correlations are all that matter, then those life and death dairy produce issues are far more important than the rather lower correlation claimed between intelligence and college graduation rates.

Like I said, I am not making an argument about cause. Arguments about cause are suspicious if they rely on correlations alone, especially if they are merely time-dependent, easily attributable to mere randomness. My argument is NOT either (1) time-dependent, (2) relying on correlations attributable to randomness, nor (3) an argument about cause. The logic of my argument is very direct and plain, and the main problem you have with the argument is its conclusion. It is a hopeless, immoral, racist, elitist conclusion, therefore there must be something wrong with the argument. I suggest you finally be realistic, as the people hurt the most by such political delusions are the people you are trying to help.

 

[bilby]

It just seems odd that you are so enthusiastic about the connection between college drop-out rates and intelligence, when you should be worrying about the alarming number of bed-sheet casualties caused by cheese consumption; or the dangerous effects of sour-cream on motorcyclists.

Obviously it would be a derail to discuss those here; but if you want to start some new threads on those topics, pseudoscience might be the place for them.

If correlations are all that matter, then those life and death dairy produce issues are far more important than the rather lower correlation claimed between intelligence and college graduation rates.

Like I said, I am not making an argument about cause. Arguments about cause are suspicious if they rely on correlations alone, especially if they are merely time-dependent, easily attributable to mere randomness. My argument is NOT either (1) time-dependent, (2) relying on correlations attributable to randomness, nor (3) an argument about cause. The logic of my argument is very direct and plain, and the main problem you have with the argument is its conclusion. It is a hopeless, immoral, racist, elitist conclusion, therefore there must be something wrong with the argument. I suggest you finally be realistic, as the people hurt the most by such political delusions are the people you are trying to help.

I am not trying to help anyone.

I am just pointing out that you are wrong.

For the record, you are wrong about my motivation too.

 

[ApostateAbe]

Like I said, I am not making an argument about cause. Arguments about cause are suspicious if they rely on correlations alone, especially if they are merely time-dependent, easily attributable to mere randomness. My argument is NOT either (1) time-dependent, (2) relying on correlations attributable to randomness, nor (3) an argument about cause. The logic of my argument is very direct and plain, and the main problem you have with the argument is its conclusion. It is a hopeless, immoral, racist, elitist conclusion, therefore there must be something wrong with the argument. I suggest you finally be realistic, as the people hurt the most by such political delusions are the people you are trying to help.

I am not trying to help anyone.

I am just pointing out that you are wrong.

For the record, you are wrong about my motivation too.

OK, yeah, I have no idea what your motivation is.

 

[AthenaAwakened]

AP tests, by subject, are better predictors of college performance than the SATs. So are GPAs.

As are family income, family address, and quite possibly shoe size.

You don't have to speak purely from your gut. The data is out there. You are right that AP test scores are better predictors than the SAT of college GPA, per Table 2 and Page 10 of this study:

http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.434.6711&rep=rep1&type=pdf

...and you are right about high-school GPAs. Family income... not so much. The Appendix on Page 14 of this study shows correlations among college GPA, high school GPA, SAT, and SES (socioeconomic status):

https://research.collegeboard.org/s...mic-status-sat-freshman-gpa-analysis-data.pdf

But you got me on shoe size. I can't find the needed correlation data for that assertion of possibility (not that I have looked).

I am not speaking from my gut. I am not the one holding to a number on a test as an article of faith. You say that correlation is enough only because every attempt at proving causation has gone down in flames. When other correlations are shown, some with stronger connection than test scores, you wave your hands at them and say "no no, it is enough to have this weaker correlation to the scores I believe in than to, say, GPA or income level," both of which tell you more about life outcomes, success, and intelligence than any single test can.

But these correlations can be proven or disproven as actual causes, they are indicative of observable behaviors at both the macro and micro level of human action. And they are things we can change and manipulate without regard to skin color, eye shape, or mysterious genetic cocktails.

 

[Nice Squirrel]

AP tests, by subject, are better predictors of college performance than the SATs. So are GPAs.

As are family income, family address, and quite possibly shoe size.

Which is why SAT scores shouldn't be considered in isolation, which they aren't, but rather used as a supplemental data point.

Correct.

 

[AthenaAwakened]

AP tests, by subject, are better predictors of college performance than the SATs. So are GPAs.

As are family income, family address, and quite possibly shoe size.

Which is why SAT scores shouldn't be considered in isolation, which they aren't, but rather used as a supplemental data point.

Why should they be considered at all if there are other more predictive measures available?

The bible says that the value for Pi is three. We know today that three is a very poor estimate for pi and we no longer use it. So why not do the same with the SAT?