Party Lines "Define" Statistical Affordable Care Act Data
We have an interesting statistical phenomena underway where pro-Affordable Care Act (ne: Democrats/Liberals) politicos refuse to believe that those presently covered under private insurance will lose their coverage as a result of the Act, and where anti-Affordable Care Act (ne: Republicans/Conservatives) dogmatics are splashing the media with gaudy survey-driven statistics pointing at the impending private sector coverage loss. So should we believe what our respective political parties are telling us and stick our collective heads in the sand when we hear any dissenting commentary? What if we are among the independent voters and thinkers that are trying to make sense of both debates, where we don't have the luxury of going along with whatever party leadership is feeding us?
The debate has caught fire in recent weeks, with controversial polls and consultants conflicting about whether employees will benefit or lose coverage by 2014. For example, McKinsey & Company earlier this week posted the questionnaire and methodology of an online survey it had administered that was denounced by the White House and others "pro-Act" for contending that nearly 1/3 of U.S. employers would definitely or probably drop employee coverage when provisions of the health care law hit in 2014.
The White House wasted little time responding to McKinsey's attempt at information-gathering clarification. “As we learn more, it has become clear that this one flawed study from McKinsey is truly an outlier,” explains Nancy-Ann DeParle, an Assistant to the President and Deputy Chief of Staff. Putting this statistical impasse to the side for a moment, I have one piece of advice for Nancy-Ann DeParle: "throw away your thesaurus; it has misserved you; no one knows what the hell an 'outlier' even is." Further, if you are attempting to marginalize the survey results of a respected management consulting firm, feel free, but don't dilute your argument with unknown words. It demonstrates a lack of command of the English language in a colloquial sense, and by default, demonstrates a lack of command of anything, including the Affordable Care Act's integrity.
It gets better, particularly if you are a fan of permutations of "CYA." McKinsey acknowledged that its survey was “not comparable” to the studies by the Congressional Budget Office, Urban Institute or others using economic modeling. Rather, it surveyed business owners using an online panel. McKinsey said it paid for the survey by Ipsos, a French marketing firm, “to capture the attitudes of employers,” large and small. In addition, McKinsey attempted to rebut the criticisms by the White House and others, asserting that its report was "not intended to be predictive." How it can not be considered "predictive?" A noted management consulting firm conducts and publishes a survey asking interviewees specifically if coverage will be cut in 2014, and then finds that 1/3 will "definitely" or "probably" get rid of coverage. How is this not predictive? The only way it could not be considered such is if we adjust the calendars so that all subsequent years will count down in time. So next year will be 2010, the following year will be 2009, and so on. Although a bit cumbersome, that would get McKinsey off the "predictive" hook.
Not surprisingly, McKinsey’s explanations did not satisfy Senator Max Baucus, chairman of the Senate Finance Committee, and several from the House who have inquired. “This report is filled with cherry-picked facts and slanted questions,” he said in a statement. I have a question for Senator Baucus: "what report isn't?"
But we bet that the White House liked this recent report: the Urban Institute, a Washington research center estimates (based on analyses of Census Bureau and Department of Health and Human Services surveys) that employer coverage will increase modestly for workers and their dependents in firms with 50 or fewer employees.
Over all, including all employers, analyses by a number of widely cited researchers predicted little or no change in employer-sponsored insurance in 2014. They include the Congressional Budget Office, RAND, Lewin Associates and Mercer. Senator Baucus probably wouldn't find any flaws in the design of the survey instruments used by these organizations since the results correlate with his current political position. But perhaps Senator Baucus and other pro-Act politicos should consider a July 6, 2011 USA TODAY article about huge Medicaid cuts that states: "Insurers and employers have their own concerns about the payment cuts. They say trimming the rates will prompt providers to raise their prices for patients who have private insurance." The article continues: "It's always a concern that when providers get less from Medicaid, they will shift the costs to private insurance so families and employers pay more."
It's "always a concern" because, like a bad weather forecast that ends up coming true, you know beforehand that things are going to get worse, and they usually do. So perhaps McKinsey wouldn't have gotten as much anti-Act mileage out of statistical data that had employers report: "we plan to raise costs" instead of "we will definitely or probably get rid of our group health insurance plans." The latter smacks of self-sustaining drama, while the former has been said so many times that no one even hears it anymore. That's the whole problem with survey reports -- it's tough to get press if you keep proclaiming the same things, things that everyone has heard so many times to the point of nausea. Better that you should figure out a way to rig questions so that you can have compelling, press-worthy outcomes. And if there is substantial pushback, just claim it's not meant to predict anything, which, unfortunately, might make many people wonder why you would conduct such a survey in the first place if it wasn't meant to predict anything.
Or, in the case of the White House, just "cherry pick" outcomes from two or three disparate surveys so you can cobble together helpful data that advances your case, even when it becomes necessary to use a thesaurus to help explain it. And if the anti-Act crowd questions the integrity of your researched/published results, just explain that it's not meant to be 'predictive.'
The debate has caught fire in recent weeks, with controversial polls and consultants conflicting about whether employees will benefit or lose coverage by 2014. For example, McKinsey & Company earlier this week posted the questionnaire and methodology of an online survey it had administered that was denounced by the White House and others "pro-Act" for contending that nearly 1/3 of U.S. employers would definitely or probably drop employee coverage when provisions of the health care law hit in 2014.
The White House wasted little time responding to McKinsey's attempt at information-gathering clarification. “As we learn more, it has become clear that this one flawed study from McKinsey is truly an outlier,” explains Nancy-Ann DeParle, an Assistant to the President and Deputy Chief of Staff. Putting this statistical impasse to the side for a moment, I have one piece of advice for Nancy-Ann DeParle: "throw away your thesaurus; it has misserved you; no one knows what the hell an 'outlier' even is." Further, if you are attempting to marginalize the survey results of a respected management consulting firm, feel free, but don't dilute your argument with unknown words. It demonstrates a lack of command of the English language in a colloquial sense, and by default, demonstrates a lack of command of anything, including the Affordable Care Act's integrity.
It gets better, particularly if you are a fan of permutations of "CYA." McKinsey acknowledged that its survey was “not comparable” to the studies by the Congressional Budget Office, Urban Institute or others using economic modeling. Rather, it surveyed business owners using an online panel. McKinsey said it paid for the survey by Ipsos, a French marketing firm, “to capture the attitudes of employers,” large and small. In addition, McKinsey attempted to rebut the criticisms by the White House and others, asserting that its report was "not intended to be predictive." How it can not be considered "predictive?" A noted management consulting firm conducts and publishes a survey asking interviewees specifically if coverage will be cut in 2014, and then finds that 1/3 will "definitely" or "probably" get rid of coverage. How is this not predictive? The only way it could not be considered such is if we adjust the calendars so that all subsequent years will count down in time. So next year will be 2010, the following year will be 2009, and so on. Although a bit cumbersome, that would get McKinsey off the "predictive" hook.
Not surprisingly, McKinsey’s explanations did not satisfy Senator Max Baucus, chairman of the Senate Finance Committee, and several from the House who have inquired. “This report is filled with cherry-picked facts and slanted questions,” he said in a statement. I have a question for Senator Baucus: "what report isn't?"
But we bet that the White House liked this recent report: the Urban Institute, a Washington research center estimates (based on analyses of Census Bureau and Department of Health and Human Services surveys) that employer coverage will increase modestly for workers and their dependents in firms with 50 or fewer employees.
Over all, including all employers, analyses by a number of widely cited researchers predicted little or no change in employer-sponsored insurance in 2014. They include the Congressional Budget Office, RAND, Lewin Associates and Mercer. Senator Baucus probably wouldn't find any flaws in the design of the survey instruments used by these organizations since the results correlate with his current political position. But perhaps Senator Baucus and other pro-Act politicos should consider a July 6, 2011 USA TODAY article about huge Medicaid cuts that states: "Insurers and employers have their own concerns about the payment cuts. They say trimming the rates will prompt providers to raise their prices for patients who have private insurance." The article continues: "It's always a concern that when providers get less from Medicaid, they will shift the costs to private insurance so families and employers pay more."
It's "always a concern" because, like a bad weather forecast that ends up coming true, you know beforehand that things are going to get worse, and they usually do. So perhaps McKinsey wouldn't have gotten as much anti-Act mileage out of statistical data that had employers report: "we plan to raise costs" instead of "we will definitely or probably get rid of our group health insurance plans." The latter smacks of self-sustaining drama, while the former has been said so many times that no one even hears it anymore. That's the whole problem with survey reports -- it's tough to get press if you keep proclaiming the same things, things that everyone has heard so many times to the point of nausea. Better that you should figure out a way to rig questions so that you can have compelling, press-worthy outcomes. And if there is substantial pushback, just claim it's not meant to predict anything, which, unfortunately, might make many people wonder why you would conduct such a survey in the first place if it wasn't meant to predict anything.
Or, in the case of the White House, just "cherry pick" outcomes from two or three disparate surveys so you can cobble together helpful data that advances your case, even when it becomes necessary to use a thesaurus to help explain it. And if the anti-Act crowd questions the integrity of your researched/published results, just explain that it's not meant to be 'predictive.'
Comments
Post a Comment