[81alum]

More interesting are the comments to the article.  Here are some of them:

Mike Callahan, the sample size was too small to be considered adequately representative of the thousands of Native Americans in the US. Anyone who has taken high school level stats can see that. Furthermore, the organization who conducted the survey was an indigenous peoples organization. This could definitely have led to survey bias. Your assertion that 67% of NA's disagree with me is premature and factually flawed if this is all you have to go on. By the way, I also represent my family who doesnt care.

This comment in particular was ludicrous. 

The total survey was 400 people, which is in fact the standard survey size for a 95% level of confidence and a 5% margin of error.  This is the normal size poll for most political campaigns, for example.

It is true that of the 400, just 98 were verified Native Americans.  This is actually a very robust result and constitutes a brobdingnagian level of "oversampling" to improve the reliability of the sub sample.  It is common in national polls to represent the opinions of minority groups gathered randomly from within a 400 person sample and not to conduct oversampling at all.  If this had been done in this survey there may have been around 12 Native American responses instead of the 98.

Now, if we used the sample of 98 by itself as a separate survey, the reliability would decrease some.  A survey based on 98 does not attain a 5% margin of error, but it does attain a 10% margin of error.  So it might be that a mere 57% of Native Americans find the name offensive, or perhaps it is more like 77%.  While the margin for error increases, it does not invalidate the results.

Particularly bizarre in that comment was the assertion that the survey was biased because it was conducted by "an indigenous peoples organization." Well, the "organization" was a sanctioned research unit of CSU San Bernardino that focuses on indigenous peoples.  Professor Fenelon and most of his team happen to be Native Americans.  Is the commentator asserting that Native Americans should automatically be regarded as non-objective about the subject of Native American oppression?  And what of the corollary?  Did it occur to the poster that Native Americans might regard surveys conducted by white people to lack objectivity?  We really need to avoid basing our judgment of the quality of social science on the racial identity of the social scientists.  That would rapidly lead us to a repulsive conclusion.

What about the 2002 survey that Garvin and the Redskins cite?  It is entirely possible that this recent survey and the 2002 SI survey are both correct for their times.  Native Americans, like all of us, evolve their thinking about issues over time.  For example, consider how public opinion on gay marriage has changed from 2002 to 2014.

Fenelon's main point, however--both in this survey and those he conducted in Ohio--appears to be that many survey organizations rely exclusively on self-identification to determine if a respondent is Native American or not.  Once the subject of the survey becomes known, respondents could easily exaggerate or even lie about their Native American connections.  His survey apparently used independent verification of Native American identity and this seems to have made a fairly big difference.

Whenever race is a critical element to polling data, you have to expect difficulties like this.  Remember all the political surveys decades before Obama that wildly exaggerated the support for candidates like Tom Bradley or for Doug Wilder.  In those days the surveys proved to be wildly inaccurate because respondents lied about their own race or about their attitudes.  Once race gets highly charged--as it obviously is on the Redskins issue today-- it adds difficulty to conducting accurate surveys, and Fenelon seems to have taken several measures to overcome those difficulties.

[81alum]

More interesting are the comments to the article.  Here are some of them:

Mike Callahan, the sample size was too small to be considered adequately representative of the thousands of Native Americans in the US. Anyone who has taken high school level stats can see that. Furthermore, the organization who conducted the survey was an indigenous peoples organization. This could definitely have led to survey bias. Your assertion that 67% of NA's disagree with me is premature and factually flawed if this is all you have to go on. By the way, I also represent my family who doesnt care.

This comment in particular was ludicrous. 

The total survey was 400 people, which is in fact the standard survey size for a 95% level of confidence and a 5% margin of error.  This is the normal size poll for most political campaigns, for example.

It is true that of the 400, just 98 were verified Native Americans.  This is actually a very robust result and constitutes a brobdingnagian level of "oversampling" to improve the reliability of the sub sample.  It is common in national polls to represent the opinions of minority groups gathered randomly from within a 400 person sample and not to conduct oversampling at all.  If this had been done in this survey there may have been around 12 Native American responses instead of the 98.

Now, if we used the sample of 98 by itself as a separate survey, the reliability would decrease some.  A survey based on 98 does not attain a 5% margin of error, but it does attain a 10% margin of error.  So it might be that a mere 57% of Native Americans find the name offensive, or perhaps it is more like 77%.  While the margin for error increases, it does not invalidate the results.

Particularly bizarre in that comment was the assertion that the survey was biased because it was conducted by "an indigenous peoples organization." Well, the "organization" was a sanctioned research unit of CSU San Bernardino that focuses on indigenous peoples.  Professor Fenelon and most of his team happen to be Native Americans.  Is the commentator asserting that Native Americans should automatically be regarded as non-objective about the subject of Native American oppression?  And what of the corollary?  Did it occur to the poster that Native Americans might regard surveys conducted by white people to lack objectivity?  We really need to avoid basing our judgment of the quality of social science on the racial identity of the social scientists.  That would rapidly lead us to a repulsive conclusion.

What about the 2002 survey that Garvin and the Redskins cite?  It is entirely possible that this recent survey and the 2002 SI survey are both correct for their times.  Native Americans, like all of us, evolve their thinking about issues over time.  For example, consider how public opinion on gay marriage has changed from 2002 to 2014.

Fenelon's main point, however--both in this survey and those he conducted in Ohio--appears to be that many survey organizations rely exclusively on self-identification to determine if a respondent is Native American or not.  Once the subject of the survey becomes known, respondents could easily exaggerate or even lie about their Native American connections.  His survey apparently used independent verification of Native American identity and this seems to have made a fairly big difference.

Whenever race is a critical element to polling data, you have to expect difficulties like this.  Remember all the political surveys decades before Obama that wildly exaggerated the support for candidates like Tom Bradley or for Doug Wilder.  In those days the surveys proved to be wildly inaccurate because respondents lied about their own race or about their attitudes.  Once race gets highly charged--as it obviously is on the Redskins issue today-- it adds difficulty to conducting accurate surveys, and Fenelon seems to have taken several measures to overcome those difficulties.