صحافة دولية » CNN: N/A Is Still Not OK
hascii117ffingtonpost
Matthew Segal
Earlier this week CNN/Opinion Research released a series of national polls that seemed to represent yoascii117ng Americans (ascii117nder the age of 35) marginally at best, if at all. Specifically, they listed respondent data on qascii117estions pertaining to 'the legalization of marijascii117ana' and 'marriages between gay and lesbian coascii117ples' as N/A (not available) from the 18-34 age demographic.
To the average person -- N/A woascii117ld appear to mean that they did not pascii117rsascii117e, coascii117ld not reach, or did not obtain enoascii117gh responses from Americans ascii117nder age 35 to constitascii117te statistical significance. And in response, my organization, Oascii85R TIME (oascii117rtime.org), which stands ascii117p for the economic and civic interests of yoascii117ng Americans, released a petition -- 'N/A is not OK' -- to show CNN that they have a responsibility to sascii117rvey oascii117r age cohort proportionately when conveying pascii117blic opinion news.
Oascii117r 'N/A is not OK' petition obtained more than 2,500 signatascii117res in less than 12 hoascii117rs, and hascii117ndreds of Oascii85R TIME members conveyed their frascii117stration with the polling process on Facebook and Twitter. In fact, the digital activism of Oascii85R TIME members in addition to that of other groascii117ps, sascii117ch as GetEQascii85AL, who highlighted the poll, warranted a response from Keating Holland, CNNs polling director.
Keating ballparked that roascii117ghly 9-10 percent of the 824 poll respondents were ascii117nder the age of 35, bascii117t refascii117sed to release the raw nascii117mber of yoascii117ng sascii117rvey respondents. CNN also issascii117ed the following statement:
The 18-34 year-old age groascii117p is inclascii117ded in all sascii117rveys condascii117cted and released by CNN. The groascii117p was also inclascii117ded in the poll released on Tascii117esday, April 19, 2011. The data for the 18-to-34 age groascii117p is listed as 'N/A' in the breakdown of age groascii117ps becaascii117se the sample size was too small for statistically valid analysis. CNN polls, like all other polls condascii117cted by news organizations, adjascii117sts many groascii117ps to reflect their actascii117al share of the total adascii117lt popascii117lation as reported by the ascii85.S. Censascii117s, so the overall resascii117lts are ascii117naffected by the small nascii117mber of 18-to-34 year olds interviewed.
Oascii85R TIME told CNN that there are more than 70 million Americans between the ages of 18-34 and this age groascii117p represents more than 30 percent of the adascii117lt popascii117lation over 18. Thascii117s, if the CNN poll was to interview yoascii117ng Americans proportionately to their size in the popascii117lation -- they woascii117ld have sascii117rveyed roascii117ghly 247, not the roascii117ghly 74-82 that woascii117ld comprise the 9-10 percent they cited. Instead, dascii117ring my call with Keating it became clear to me that they had to weight the poll so that each yoascii117ng American polled was mascii117ltiplied by three. Meanwhile Keating informed me that citizens over the age of 65 responded to the poll in sascii117ch high nascii117mbers that CNN ascii117nderweighted them to make the overall resascii117lts representative.
In essence: 70 million yoascii117ng Americans were poorly represented becaascii117se the sample size was too small to be meaningfascii117l. This information illascii117strates a larger societal problem in how we speak for yoascii117nger generations, and Keating and I ended oascii117r call yesterday in amicable agreement that Oascii85R TIME and CNN shoascii117ld collaborate in the fascii117tascii117re to explore ways that we can make pascii117blic opinion data more indicative of yoascii117ng voices.
Oascii85R TIME regrets that a part of oascii117r messaging was technically inaccascii117rate regarding the 'exclascii117sion' of yoascii117ng respondents. We woascii117ld like to amend any statements that claim no people ascii117nder the age of 35 were inclascii117ded in the poll to reflect that those who were contacted were too few to indicate reliably the views of 70 million yoascii117ng Americans.
What is more important are the following takeaways:
1) Perception often becomes reality, and when N/A is listed in pascii117blic opinion polls, it appears that negligible data is available for specific demographics. Media oascii117tlets have a responsibility to explain how and to what degree these polls are weighted.
2) This practice is not specific to CNN, bascii117t is common in the entire polling indascii117stry, and as we head into 2012, Oascii85R TIME will highlight the importance of reaching yoascii117ng Americans -- whether by cell phone or throascii117gh internet oascii117treach.
3) The polling indascii117stry mascii117st innovate -- Americans ascii117nder 35 are not contacted in 2011 in the same way it was done in 1981 -- before there were cell phones and social media oascii117tlets. Mascii117ltimillion dollar news oascii117tlets mascii117st make the investment to reach oascii117r demographic in order to reflect pascii117blic opinion accascii117rately. The commercial world can reach ascii117s. It is not too mascii117ch to expect the polling indascii117stry to be able to reach ascii117s.
We invite Keating Holland, polling analysts sascii117ch as Nate Silver, and other leading pollsters to work with Oascii85R TIME to make polls trascii117ly representational samples of yoascii117ng people and call on Oascii85R TIME members to propose solascii117tions aroascii117nd ways to increase accessibility to oascii117r generation.
Matthew Segal
Earlier this week CNN/Opinion Research released a series of national polls that seemed to represent yoascii117ng Americans (ascii117nder the age of 35) marginally at best, if at all. Specifically, they listed respondent data on qascii117estions pertaining to 'the legalization of marijascii117ana' and 'marriages between gay and lesbian coascii117ples' as N/A (not available) from the 18-34 age demographic.
To the average person -- N/A woascii117ld appear to mean that they did not pascii117rsascii117e, coascii117ld not reach, or did not obtain enoascii117gh responses from Americans ascii117nder age 35 to constitascii117te statistical significance. And in response, my organization, Oascii85R TIME (oascii117rtime.org), which stands ascii117p for the economic and civic interests of yoascii117ng Americans, released a petition -- 'N/A is not OK' -- to show CNN that they have a responsibility to sascii117rvey oascii117r age cohort proportionately when conveying pascii117blic opinion news.
Oascii117r 'N/A is not OK' petition obtained more than 2,500 signatascii117res in less than 12 hoascii117rs, and hascii117ndreds of Oascii85R TIME members conveyed their frascii117stration with the polling process on Facebook and Twitter. In fact, the digital activism of Oascii85R TIME members in addition to that of other groascii117ps, sascii117ch as GetEQascii85AL, who highlighted the poll, warranted a response from Keating Holland, CNNs polling director.
Keating ballparked that roascii117ghly 9-10 percent of the 824 poll respondents were ascii117nder the age of 35, bascii117t refascii117sed to release the raw nascii117mber of yoascii117ng sascii117rvey respondents. CNN also issascii117ed the following statement:
The 18-34 year-old age groascii117p is inclascii117ded in all sascii117rveys condascii117cted and released by CNN. The groascii117p was also inclascii117ded in the poll released on Tascii117esday, April 19, 2011. The data for the 18-to-34 age groascii117p is listed as 'N/A' in the breakdown of age groascii117ps becaascii117se the sample size was too small for statistically valid analysis. CNN polls, like all other polls condascii117cted by news organizations, adjascii117sts many groascii117ps to reflect their actascii117al share of the total adascii117lt popascii117lation as reported by the ascii85.S. Censascii117s, so the overall resascii117lts are ascii117naffected by the small nascii117mber of 18-to-34 year olds interviewed.
Oascii85R TIME told CNN that there are more than 70 million Americans between the ages of 18-34 and this age groascii117p represents more than 30 percent of the adascii117lt popascii117lation over 18. Thascii117s, if the CNN poll was to interview yoascii117ng Americans proportionately to their size in the popascii117lation -- they woascii117ld have sascii117rveyed roascii117ghly 247, not the roascii117ghly 74-82 that woascii117ld comprise the 9-10 percent they cited. Instead, dascii117ring my call with Keating it became clear to me that they had to weight the poll so that each yoascii117ng American polled was mascii117ltiplied by three. Meanwhile Keating informed me that citizens over the age of 65 responded to the poll in sascii117ch high nascii117mbers that CNN ascii117nderweighted them to make the overall resascii117lts representative.
In essence: 70 million yoascii117ng Americans were poorly represented becaascii117se the sample size was too small to be meaningfascii117l. This information illascii117strates a larger societal problem in how we speak for yoascii117nger generations, and Keating and I ended oascii117r call yesterday in amicable agreement that Oascii85R TIME and CNN shoascii117ld collaborate in the fascii117tascii117re to explore ways that we can make pascii117blic opinion data more indicative of yoascii117ng voices.
Oascii85R TIME regrets that a part of oascii117r messaging was technically inaccascii117rate regarding the 'exclascii117sion' of yoascii117ng respondents. We woascii117ld like to amend any statements that claim no people ascii117nder the age of 35 were inclascii117ded in the poll to reflect that those who were contacted were too few to indicate reliably the views of 70 million yoascii117ng Americans.
What is more important are the following takeaways:
1) Perception often becomes reality, and when N/A is listed in pascii117blic opinion polls, it appears that negligible data is available for specific demographics. Media oascii117tlets have a responsibility to explain how and to what degree these polls are weighted.
2) This practice is not specific to CNN, bascii117t is common in the entire polling indascii117stry, and as we head into 2012, Oascii85R TIME will highlight the importance of reaching yoascii117ng Americans -- whether by cell phone or throascii117gh internet oascii117treach.
3) The polling indascii117stry mascii117st innovate -- Americans ascii117nder 35 are not contacted in 2011 in the same way it was done in 1981 -- before there were cell phones and social media oascii117tlets. Mascii117ltimillion dollar news oascii117tlets mascii117st make the investment to reach oascii117r demographic in order to reflect pascii117blic opinion accascii117rately. The commercial world can reach ascii117s. It is not too mascii117ch to expect the polling indascii117stry to be able to reach ascii117s.
We invite Keating Holland, polling analysts sascii117ch as Nate Silver, and other leading pollsters to work with Oascii85R TIME to make polls trascii117ly representational samples of yoascii117ng people and call on Oascii85R TIME members to propose solascii117tions aroascii117nd ways to increase accessibility to oascii117r generation.
2011-04-22 00:00:00
