ASA Connect

Of course you should teach statistics.  What data sets you choose to consider, how you speak about the eugenics that was a part of statistics, and how you engage with your students is all part of being more equitable.  Do you talk about the fact that what and how data is collected and experimentation is designed is ultimately influenced by our perspective?  To do so is to invite the concept that perhaps a lot of findings might have differed had people of color been involved in the process.  This invites students of color in to a space that has NOT been inviting and is NOT taught by faculty of color.

What percent of statisticians are black?  Latino?  Remember that there was never any data to show that orchestras were biased against women horn players until there was finally a blind audition.  The same experimentation has not happened for academia. It should.

This change can not happen until white people like us rethink how we teach what we teach.  Critical Race Theory is only problematic to those who are not willing to think about your own whiteness has impacted all that you have achieved and all you do. 

You can not be colorblind.

Of course you should teach statistics.  What data sets you choose to consider, how you speak about the eugenics that was a part of statistics, and how you engage with your students is all part of being more equitable.  Do you talk about the fact that what and how data is collected and experimentation is designed is ultimately influenced by our perspective?  To do so is to invite the concept that perhaps a lot of findings might have differed had people of color been involved in the process.  This invites students of color in to a space that has NOT been inviting and is NOT taught by faculty of color.

What percent of statisticians are black?  Latino?  Remember that there was never any data to show that orchestras were biased against women horn players until there was finally a blind audition.  The same experimentation has not happened for academia. It should.

This change can not happen until white people like us rethink how we teach what we teach.  Critical Race Theory is only problematic to those who are not willing to think about your own whiteness has impacted all that you have achieved and all you do. 

You can not be colorblind.

Of course you should teach statistics.  What data sets you choose to consider, how you speak about the eugenics that was a part of statistics, and how you engage with your students is all part of being more equitable.  Do you talk about the fact that what and how data is collected and experimentation is designed is ultimately influenced by our perspective?  To do so is to invite the concept that perhaps a lot of findings might have differed had people of color been involved in the process.  This invites students of color in to a space that has NOT been inviting and is NOT taught by faculty of color.

What percent of statisticians are black?  Latino?  Remember that there was never any data to show that orchestras were biased against women horn players until there was finally a blind audition.  The same experimentation has not happened for academia. It should.

This change can not happen until white people like us rethink how we teach what we teach.  Critical Race Theory is only problematic to those who are not willing to think about your own whiteness has impacted all that you have achieved and all you do. 

You can not be colorblind.

I don't know why there are relatively few Black and Latino statisticians. I hope we will see more in the future.  It would be great to both expand the available talent pool of statisticians and to expand the set of desired and accessible career paths to people from those backgrounds. When empirical research demonstrates that there specific obstacles to their advancement we can discuss and implement specific solutions.  If you know of empirical research on the subject, please share some sources. Either way, I hope more research will be done on this question. Since there are so many, say, East Asian and South Asian statisticians who excel as both teachers and students of white statisticians, I suspect that the lower numbers of Black and Latino statisticians is due to factors other than "a space that has NOT been inviting and is NOT taught by faculty of color", or "white supremacy". 

But how can we seriously address this issue through Critical Race Theory, with its gauzily defined concepts and non-falsifiable assumptions such as "systemic racism" and "whiteness", and with its deliberately divisive attribution of certain attitudes and behaviors to individuals solely on the basis of their "color"?  None of that has any more empirical foundation than other pseudosciences such as Astrology, Aromatherapy or Nazi Race Theory. Yet its adherents preach it with the same certitude and emotional fervor as the evangelists for a revealed religion.

Statements such as "not willing to think about your own whiteness has impacted all that you have achieved and all you do" have no more demonstrated effectiveness than the sandwich board appeals to "Repent. The end is nigh".

Of course you should teach statistics.  What data sets you choose to consider, how you speak about the eugenics that was a part of statistics, and how you engage with your students is all part of being more equitable.  Do you talk about the fact that what and how data is collected and experimentation is designed is ultimately influenced by our perspective?  To do so is to invite the concept that perhaps a lot of findings might have differed had people of color been involved in the process.  This invites students of color in to a space that has NOT been inviting and is NOT taught by faculty of color.

What percent of statisticians are black?  Latino?  Remember that there was never any data to show that orchestras were biased against women horn players until there was finally a blind audition.  The same experimentation has not happened for academia. It should.

This change can not happen until white people like us rethink how we teach what we teach.  Critical Race Theory is only problematic to those who are not willing to think about your own whiteness has impacted all that you have achieved and all you do. 

You can not be colorblind.

I don't know why there are relatively few Black and Latino statisticians. I hope we will see more in the future.  It would be great to both expand the available talent pool of statisticians and to expand the set of desired and accessible career paths to people from those backgrounds. When empirical research demonstrates that there specific obstacles to their advancement we can discuss and implement specific solutions.  If you know of empirical research on the subject, please share some sources. Either way, I hope more research will be done on this question. Since there are so many, say, East Asian and South Asian statisticians who excel as both teachers and students of white statisticians, I suspect that the lower numbers of Black and Latino statisticians is due to factors other than "a space that has NOT been inviting and is NOT taught by faculty of color", or "white supremacy". 

But how can we seriously address this issue through Critical Race Theory, with its gauzily defined concepts and non-falsifiable assumptions such as "systemic racism" and "whiteness", and with its deliberately divisive attribution of certain attitudes and behaviors to individuals solely on the basis of their "color"?  None of that has any more empirical foundation than other pseudosciences such as Astrology, Aromatherapy or Nazi Race Theory. Yet its adherents preach it with the same certitude and emotional fervor as the evangelists for a revealed religion.

Statements such as "not willing to think about your own whiteness has impacted all that you have achieved and all you do" have no more demonstrated effectiveness than the sandwich board appeals to "Repent. The end is nigh".

Of course you should teach statistics.  What data sets you choose to consider, how you speak about the eugenics that was a part of statistics, and how you engage with your students is all part of being more equitable.  Do you talk about the fact that what and how data is collected and experimentation is designed is ultimately influenced by our perspective?  To do so is to invite the concept that perhaps a lot of findings might have differed had people of color been involved in the process.  This invites students of color in to a space that has NOT been inviting and is NOT taught by faculty of color.

What percent of statisticians are black?  Latino?  Remember that there was never any data to show that orchestras were biased against women horn players until there was finally a blind audition.  The same experimentation has not happened for academia. It should.

This change can not happen until white people like us rethink how we teach what we teach.  Critical Race Theory is only problematic to those who are not willing to think about your own whiteness has impacted all that you have achieved and all you do. 

You can not be colorblind.

It is with respect to the issues raised by John Major that, in theory, ASA members should have special expertise.  One question members interested in the measurement of racism could usefully consider is whether a decisionmaker is more likely to be racially biased in Situation 1 (where decisions yield favorable outcome rates for whites and blacks of 80% and 63%) or in Situation 2 (where white and black rate favorable outcome rates are 95% and 87%).

It would be useful for those addressing questions like this to understand that if the outcome rates involve health or healthcare issues and Situations 1 and 2 involve earlier and later points in time, then:

(a) for purpose of evaluating progress in meeting Healthy People 2020 disparities reduction goals, the National Center for Health Statistics would appraise the situation as follows depending on whether the outcome involves a health or healthcare issue:
(i) if the outcome involves a health issue like survival from cancer or COVID-19, the disparity would be deemed to have increased (on the basis of the larger relative difference in the adverse outcome at the later point in time);
(ii) if the outcome involves a healthcare issue like receipt of vaccination, the disparity would be deemed to have decreased (on the basis of the smaller relative difference in the favorable outcome at the later point in time); 

(b) for purposes of the yearly National Healthcare Quality and Disparities Report, the Agency for Healthcare Research and Quality would appraise the situation as follows regardless of whether the outcome involves a health or healthcare issue:
(i) the disparity would be deemed to be larger at the later point in time (on the basis of the larger relative difference in the adverse outcome at the later point in time);
(ii) the disparity would be deemed to have decreased over time (on the basis of the larger percentage point decrease in the adverse outcome for blacks than for whites).

The extent to which ASA members are aware of the potential for these varying interpretations, and are able to provide insight into the utility of any them for quantifying racism or other factors causing favorable and corresponding adverse outcome rates of racial groups to differ, would reflect whether they ASA members or statisticians generally in fact have special expertise with regard to issues raised by Dr. Major.

It is with respect to the issues raised by John Major that, in theory, ASA members should have special expertise.  One question members interested in the measurement of racism could usefully consider is whether a decisionmaker is more likely to be racially biased in Situation 1 (where decisions yield favorable outcome rates for whites and blacks of 80% and 63%) or in Situation 2 (where white and black rate favorable outcome rates are 95% and 87%).

It would be useful for those addressing questions like this to understand that if the outcome rates involve health or healthcare issues and Situations 1 and 2 involve earlier and later points in time, then:

(a) for purpose of evaluating progress in meeting Healthy People 2020 disparities reduction goals, the National Center for Health Statistics would appraise the situation as follows depending on whether the outcome involves a health or healthcare issue:
(i) if the outcome involves a health issue like survival from cancer or COVID-19, the disparity would be deemed to have increased (on the basis of the larger relative difference in the adverse outcome at the later point in time);
(ii) if the outcome involves a healthcare issue like receipt of vaccination, the disparity would be deemed to have decreased (on the basis of the smaller relative difference in the favorable outcome at the later point in time); 

(b) for purposes of the yearly National Healthcare Quality and Disparities Report, the Agency for Healthcare Research and Quality would appraise the situation as follows regardless of whether the outcome involves a health or healthcare issue:
(i) the disparity would be deemed to be larger at the later point in time (on the basis of the larger relative difference in the adverse outcome at the later point in time);
(ii) the disparity would be deemed to have decreased over time (on the basis of the larger percentage point decrease in the adverse outcome for blacks than for whites).

The extent to which ASA members are aware of the potential for these varying interpretations, and are able to provide insight into the utility of any them for quantifying racism or other factors causing favorable and corresponding adverse outcome rates of racial groups to differ, would reflect whether they ASA members or statisticians generally in fact have special expertise with regard to issues raised by Dr. Major.

It is with respect to the issues raised by John Major that, in theory, ASA members should have special expertise.  One question members interested in the measurement of racism could usefully consider is whether a decisionmaker is more likely to be racially biased in Situation 1 (where decisions yield favorable outcome rates for whites and blacks of 80% and 63%) or in Situation 2 (where white and black rate favorable outcome rates are 95% and 87%).

It would be useful for those addressing questions like this to understand that if the outcome rates involve health or healthcare issues and Situations 1 and 2 involve earlier and later points in time, then:

(a) for purpose of evaluating progress in meeting Healthy People 2020 disparities reduction goals, the National Center for Health Statistics would appraise the situation as follows depending on whether the outcome involves a health or healthcare issue:
(i) if the outcome involves a health issue like survival from cancer or COVID-19, the disparity would be deemed to have increased (on the basis of the larger relative difference in the adverse outcome at the later point in time);
(ii) if the outcome involves a healthcare issue like receipt of vaccination, the disparity would be deemed to have decreased (on the basis of the smaller relative difference in the favorable outcome at the later point in time); 

(b) for purposes of the yearly National Healthcare Quality and Disparities Report, the Agency for Healthcare Research and Quality would appraise the situation as follows regardless of whether the outcome involves a health or healthcare issue:
(i) the disparity would be deemed to be larger at the later point in time (on the basis of the larger relative difference in the adverse outcome at the later point in time);
(ii) the disparity would be deemed to have decreased over time (on the basis of the larger percentage point decrease in the adverse outcome for blacks than for whites).

The extent to which ASA members are aware of the potential for these varying interpretations, and are able to provide insight into the utility of any them for quantifying racism or other factors causing favorable and corresponding adverse outcome rates of racial groups to differ, would reflect whether they ASA members or statisticians generally in fact have special expertise with regard to issues raised by Dr. Major.

It is with respect to the issues raised by John Major that, in theory, ASA members should have special expertise.  One question members interested in the measurement of racism could usefully consider is whether a decisionmaker is more likely to be racially biased in Situation 1 (where decisions yield favorable outcome rates for whites and blacks of 80% and 63%) or in Situation 2 (where white and black rate favorable outcome rates are 95% and 87%).

It would be useful for those addressing questions like this to understand that if the outcome rates involve health or healthcare issues and Situations 1 and 2 involve earlier and later points in time, then:

(a) for purpose of evaluating progress in meeting Healthy People 2020 disparities reduction goals, the National Center for Health Statistics would appraise the situation as follows depending on whether the outcome involves a health or healthcare issue:
(i) if the outcome involves a health issue like survival from cancer or COVID-19, the disparity would be deemed to have increased (on the basis of the larger relative difference in the adverse outcome at the later point in time);
(ii) if the outcome involves a healthcare issue like receipt of vaccination, the disparity would be deemed to have decreased (on the basis of the smaller relative difference in the favorable outcome at the later point in time); 

(b) for purposes of the yearly National Healthcare Quality and Disparities Report, the Agency for Healthcare Research and Quality would appraise the situation as follows regardless of whether the outcome involves a health or healthcare issue:
(i) the disparity would be deemed to be larger at the later point in time (on the basis of the larger relative difference in the adverse outcome at the later point in time);
(ii) the disparity would be deemed to have decreased over time (on the basis of the larger percentage point decrease in the adverse outcome for blacks than for whites).

The extent to which ASA members are aware of the potential for these varying interpretations, and are able to provide insight into the utility of any them for quantifying racism or other factors causing favorable and corresponding adverse outcome rates of racial groups to differ, would reflect whether they ASA members or statisticians generally in fact have special expertise with regard to issues raised by Dr. Major.

It is with respect to the issues raised by John Major that, in theory, ASA members should have special expertise.  One question members interested in the measurement of racism could usefully consider is whether a decisionmaker is more likely to be racially biased in Situation 1 (where decisions yield favorable outcome rates for whites and blacks of 80% and 63%) or in Situation 2 (where white and black rate favorable outcome rates are 95% and 87%).

It would be useful for those addressing questions like this to understand that if the outcome rates involve health or healthcare issues and Situations 1 and 2 involve earlier and later points in time, then:

(a) for purpose of evaluating progress in meeting Healthy People 2020 disparities reduction goals, the National Center for Health Statistics would appraise the situation as follows depending on whether the outcome involves a health or healthcare issue:
(i) if the outcome involves a health issue like survival from cancer or COVID-19, the disparity would be deemed to have increased (on the basis of the larger relative difference in the adverse outcome at the later point in time);
(ii) if the outcome involves a healthcare issue like receipt of vaccination, the disparity would be deemed to have decreased (on the basis of the smaller relative difference in the favorable outcome at the later point in time); 

(b) for purposes of the yearly National Healthcare Quality and Disparities Report, the Agency for Healthcare Research and Quality would appraise the situation as follows regardless of whether the outcome involves a health or healthcare issue:
(i) the disparity would be deemed to be larger at the later point in time (on the basis of the larger relative difference in the adverse outcome at the later point in time);
(ii) the disparity would be deemed to have decreased over time (on the basis of the larger percentage point decrease in the adverse outcome for blacks than for whites).

The extent to which ASA members are aware of the potential for these varying interpretations, and are able to provide insight into the utility of any them for quantifying racism or other factors causing favorable and corresponding adverse outcome rates of racial groups to differ, would reflect whether they ASA members or statisticians generally in fact have special expertise with regard to issues raised by Dr. Major.

It is with respect to the issues raised by John Major that, in theory, ASA members should have special expertise.  One question members interested in the measurement of racism could usefully consider is whether a decisionmaker is more likely to be racially biased in Situation 1 (where decisions yield favorable outcome rates for whites and blacks of 80% and 63%) or in Situation 2 (where white and black rate favorable outcome rates are 95% and 87%).

It would be useful for those addressing questions like this to understand that if the outcome rates involve health or healthcare issues and Situations 1 and 2 involve earlier and later points in time, then:

(a) for purpose of evaluating progress in meeting Healthy People 2020 disparities reduction goals, the National Center for Health Statistics would appraise the situation as follows depending on whether the outcome involves a health or healthcare issue:
(i) if the outcome involves a health issue like survival from cancer or COVID-19, the disparity would be deemed to have increased (on the basis of the larger relative difference in the adverse outcome at the later point in time);
(ii) if the outcome involves a healthcare issue like receipt of vaccination, the disparity would be deemed to have decreased (on the basis of the smaller relative difference in the favorable outcome at the later point in time); 

(b) for purposes of the yearly National Healthcare Quality and Disparities Report, the Agency for Healthcare Research and Quality would appraise the situation as follows regardless of whether the outcome involves a health or healthcare issue:
(i) the disparity would be deemed to be larger at the later point in time (on the basis of the larger relative difference in the adverse outcome at the later point in time);
(ii) the disparity would be deemed to have decreased over time (on the basis of the larger percentage point decrease in the adverse outcome for blacks than for whites).

The extent to which ASA members are aware of the potential for these varying interpretations, and are able to provide insight into the utility of any them for quantifying racism or other factors causing favorable and corresponding adverse outcome rates of racial groups to differ, would reflect whether they ASA members or statisticians generally in fact have special expertise with regard to issues raised by Dr. Major.

It is with respect to the issues raised by John Major that, in theory, ASA members should have special expertise.  One question members interested in the measurement of racism could usefully consider is whether a decisionmaker is more likely to be racially biased in Situation 1 (where decisions yield favorable outcome rates for whites and blacks of 80% and 63%) or in Situation 2 (where white and black rate favorable outcome rates are 95% and 87%).

It would be useful for those addressing questions like this to understand that if the outcome rates involve health or healthcare issues and Situations 1 and 2 involve earlier and later points in time, then:

(a) for purpose of evaluating progress in meeting Healthy People 2020 disparities reduction goals, the National Center for Health Statistics would appraise the situation as follows depending on whether the outcome involves a health or healthcare issue:
(i) if the outcome involves a health issue like survival from cancer or COVID-19, the disparity would be deemed to have increased (on the basis of the larger relative difference in the adverse outcome at the later point in time);
(ii) if the outcome involves a healthcare issue like receipt of vaccination, the disparity would be deemed to have decreased (on the basis of the smaller relative difference in the favorable outcome at the later point in time); 

(b) for purposes of the yearly National Healthcare Quality and Disparities Report, the Agency for Healthcare Research and Quality would appraise the situation as follows regardless of whether the outcome involves a health or healthcare issue:
(i) the disparity would be deemed to be larger at the later point in time (on the basis of the larger relative difference in the adverse outcome at the later point in time);
(ii) the disparity would be deemed to have decreased over time (on the basis of the larger percentage point decrease in the adverse outcome for blacks than for whites).

The extent to which ASA members are aware of the potential for these varying interpretations, and are able to provide insight into the utility of any them for quantifying racism or other factors causing favorable and corresponding adverse outcome rates of racial groups to differ, would reflect whether they ASA members or statisticians generally in fact have special expertise with regard to issues raised by Dr. Major.

This has been an interesting thread to read.  But I wonder if, at least as I am reading through it, we are confounding teaching statistics and teaching about statistics.

I would guess we could all agree that there is a difference between teaching art and teaching art history.  I think the same applies to all subjects (except maybe history itself).

I would argue that there is strong value in knowing the history of the subject area - and wish I'd had an opportunity for such a class - but I would urge those who want to pepper that into their standard courses to reconsider.

I, for one, would have been rather upset having that sprung on me in college or grad school.  I took math/stats courses because I enjoyed the logic, the clarity, the topical unity across lectures, and the sheer geekiness of professors who randomized their exams.  In fact, that is part of the reason I decided to go into those areas (after a particularly horrid semester of guess-what-the-professor-thinks in some humanities courses; it was a breath of fresh air!).

We do a disservice to students who like math and statistics if we make their math and statistics courses resemble history courses.

Again, for clarity, I think it would be wonderful to have a course on the history, personalities, and development of statistics - I just don't think it should be slipped into statistics courses, but kept as a separate course.

I agree with Mary Kwasny. I like this thread because it addresses an important topic: human bias to doing the same thing in the same way. The old way was to relegate Women's studies , race studies in other departments. Now the time has come to integrate the concepts into math, science and arts.

Many departments are doing it already, including decolonizing literature, design spaces, artificial intelligence, and certainly statistics.

Data sets have to include diverse populations and classes in order for algorithms to be generalizable or even ethically acceptable.

Objective evidence has to be questioned.
Not because objectivity is bad - but whose objectivity? Are subjective evidence considered? What about qualitative evidence and data? Already many algorithms are opening up to non-numeric data.

Decolonizing the syllabus is very appealing and to me - the new generation demands it and for rightful reasons. It's not only about pointing out the obviously racist assertions of eugenics - but also about improving current practices of modeling.

Nusrat Rabbee

​

I'm a woman and frankly, I'd rather study just about anything besides women (except maybe art history).  If women's studies is now part of math and science, then I would have had trouble going into math and science if I were in school now.  Not that there's anything wrong with women's studies, but I just vastly prefer math and science courses!  I enjoy the style, the rigor, the logic!  

That is, of course, my personal preference and I don't expect that to hold for everyone (certainly not for those who major in social studies, etc!).  I also have no problem with "cross department" type classes - I even have a degree in that sort of an area - but if math is women's studies, what is math?

I agree with Mary Kwasny. I like this thread because it addresses an important topic: human bias to doing the same thing in the same way. The old way was to relegate Women's studies , race studies in other departments. Now the time has come to integrate the concepts into math, science and arts.

Many departments are doing it already, including decolonizing literature, design spaces, artificial intelligence, and certainly statistics.

Data sets have to include diverse populations and classes in order for algorithms to be generalizable or even ethically acceptable.

Objective evidence has to be questioned.
Not because objectivity is bad - but whose objectivity? Are subjective evidence considered? What about qualitative evidence and data? Already many algorithms are opening up to non-numeric data.

Decolonizing the syllabus is very appealing and to me - the new generation demands it and for rightful reasons. It's not only about pointing out the obviously racist assertions of eugenics - but also about improving current practices of modeling.

Nusrat Rabbee

​

I'm a woman and frankly, I'd rather study just about anything besides women (except maybe art history).  If women's studies is now part of math and science, then I would have had trouble going into math and science if I were in school now.  Not that there's anything wrong with women's studies, but I just vastly prefer math and science courses!  I enjoy the style, the rigor, the logic!  

That is, of course, my personal preference and I don't expect that to hold for everyone (certainly not for those who major in social studies, etc!).  I also have no problem with "cross department" type classes - I even have a degree in that sort of an area - but if math is women's studies, what is math?

I agree with Mary Kwasny. I like this thread because it addresses an important topic: human bias to doing the same thing in the same way. The old way was to relegate Women's studies , race studies in other departments. Now the time has come to integrate the concepts into math, science and arts.

Many departments are doing it already, including decolonizing literature, design spaces, artificial intelligence, and certainly statistics.

Data sets have to include diverse populations and classes in order for algorithms to be generalizable or even ethically acceptable.

Objective evidence has to be questioned.
Not because objectivity is bad - but whose objectivity? Are subjective evidence considered? What about qualitative evidence and data? Already many algorithms are opening up to non-numeric data.

Decolonizing the syllabus is very appealing and to me - the new generation demands it and for rightful reasons. It's not only about pointing out the obviously racist assertions of eugenics - but also about improving current practices of modeling.

Nusrat Rabbee

​

I'm a woman and frankly, I'd rather study just about anything besides women (except maybe art history).  If women's studies is now part of math and science, then I would have had trouble going into math and science if I were in school now.  Not that there's anything wrong with women's studies, but I just vastly prefer math and science courses!  I enjoy the style, the rigor, the logic!  

That is, of course, my personal preference and I don't expect that to hold for everyone (certainly not for those who major in social studies, etc!).  I also have no problem with "cross department" type classes - I even have a degree in that sort of an area - but if math is women's studies, what is math?

I agree with Mary Kwasny. I like this thread because it addresses an important topic: human bias to doing the same thing in the same way. The old way was to relegate Women's studies , race studies in other departments. Now the time has come to integrate the concepts into math, science and arts.

Many departments are doing it already, including decolonizing literature, design spaces, artificial intelligence, and certainly statistics.

Data sets have to include diverse populations and classes in order for algorithms to be generalizable or even ethically acceptable.

Objective evidence has to be questioned.
Not because objectivity is bad - but whose objectivity? Are subjective evidence considered? What about qualitative evidence and data? Already many algorithms are opening up to non-numeric data.

Decolonizing the syllabus is very appealing and to me - the new generation demands it and for rightful reasons. It's not only about pointing out the obviously racist assertions of eugenics - but also about improving current practices of modeling.

Nusrat Rabbee

​

I was disappointed to read in this discussion thread that some believe that a Statistics class should only teach statistical methods.  Why?  Properly using statistical methods requires subject matter knowledge to make decisions.  Throughout the years we have unleased folks from our classes that seem to think that armed with just a P-value, they can come up with proper decisions.  Even when a power analysis is taught, many times the effect size is expressed as a "percent difference" which is also completely devoid of the necessary subject matter knowledge.  (Checking for a 5% difference means something very different among different disciplines and certainly for different objectives within the same discipline.)

More power to those using examples that make students think and help us learn to make a better society.

Jim Baldwin

Jim,

Yes, subject matter knowledge is critical and that point must surely be included in a comprehensive course of study!  However, there are as many statistical subjects as there are things to count compute statistics for - from people, diseases, widgets, factories, wildlife, to produce.  Certainly any of these can make good examples and certainly there are areas where some of the topics brought up here definitely apply (say, a course on survey statistics).  However, there is a leap between covering the concept of subject knowledge and giving examples to having a women's studies-leaning (for example) statistics course.

I was disappointed to read in this discussion thread that some believe that a Statistics class should only teach statistical methods.  Why?  Properly using statistical methods requires subject matter knowledge to make decisions.  Throughout the years we have unleased folks from our classes that seem to think that armed with just a P-value, they can come up with proper decisions.  Even when a power analysis is taught, many times the effect size is expressed as a "percent difference" which is also completely devoid of the necessary subject matter knowledge.  (Checking for a 5% difference means something very different among different disciplines and certainly for different objectives within the same discipline.)

More power to those using examples that make students think and help us learn to make a better society.

Jim Baldwin

I was disappointed to read in this discussion thread that some believe that a Statistics class should only teach statistical methods.  Why?  Properly using statistical methods requires subject matter knowledge to make decisions.  Throughout the years we have unleased folks from our classes that seem to think that armed with just a P-value, they can come up with proper decisions.  Even when a power analysis is taught, many times the effect size is expressed as a "percent difference" which is also completely devoid of the necessary subject matter knowledge.  (Checking for a 5% difference means something very different among different disciplines and certainly for different objectives within the same discipline.)

More power to those using examples that make students think and help us learn to make a better society.

Jim Baldwin

I was disappointed to read in this discussion thread that some believe that a Statistics class should only teach statistical methods.  Why?  Properly using statistical methods requires subject matter knowledge to make decisions.  Throughout the years we have unleased folks from our classes that seem to think that armed with just a P-value, they can come up with proper decisions.  Even when a power analysis is taught, many times the effect size is expressed as a "percent difference" which is also completely devoid of the necessary subject matter knowledge.  (Checking for a 5% difference means something very different among different disciplines and certainly for different objectives within the same discipline.)

More power to those using examples that make students think and help us learn to make a better society.

Jim Baldwin

I was disappointed to read in this discussion thread that some believe that a Statistics class should only teach statistical methods.  Why?  Properly using statistical methods requires subject matter knowledge to make decisions.  Throughout the years we have unleased folks from our classes that seem to think that armed with just a P-value, they can come up with proper decisions.  Even when a power analysis is taught, many times the effect size is expressed as a "percent difference" which is also completely devoid of the necessary subject matter knowledge.  (Checking for a 5% difference means something very different among different disciplines and certainly for different objectives within the same discipline.)

More power to those using examples that make students think and help us learn to make a better society.

Jim Baldwin

I was disappointed to read in this discussion thread that some believe that a Statistics class should only teach statistical methods.  Why?  Properly using statistical methods requires subject matter knowledge to make decisions.  Throughout the years we have unleased folks from our classes that seem to think that armed with just a P-value, they can come up with proper decisions.  Even when a power analysis is taught, many times the effect size is expressed as a "percent difference" which is also completely devoid of the necessary subject matter knowledge.  (Checking for a 5% difference means something very different among different disciplines and certainly for different objectives within the same discipline.)

More power to those using examples that make students think and help us learn to make a better society.

Jim Baldwin