Methods of Research with Michael D. Tanner from the Cato Institute
On Thursday, October 23rd, Michael D. Tanner came to Penn’s campus from Washington D.C. for the WPBA Policy Cup’s first seminar. Representatives from the forty one participating teams gathered in Steinberg Hall-Dietrich Hall in the evening for Mr. Tanner’s presentation. A Senior Fellow at the Cato Institute, Mr. Tanner started the evening with a laugh. ”How do I do my research? I give it to a research assistant.” Instead of a dull summary of research techniques, Mr. Tanner organized the night around the questions that a policy-maker must ask himself as he searches for the best solutions to today’s health care problems.
The central message of the evening was one that Mr. Tanner has learned over his years at Cato: Unless (or even if) one has only a column of data, there are no unbiased sources of information. As policy-makers, do not accept what is given at face value, but strive to look at the second and third level elements of common claims.
Mr. Tanner began with the forms of bias he sees in information that crosses his desk every day. Ideological bias is evident in the left or right-leaning articles in certain health care journals. Institutional bias is the idea that people tend to present information based on the way that they are taught, or that institutions provide information that is conducive to their own survival. The most important bias is perhaps personal-actualization bias, as Mr. Tanner calls it, in which people have their own goals for the information that affects their method of presentation. Some researchers are always contrarians, for example, and will only present data that is contrary to conventional wisdom. Or if a researcher hopes to discover something new with his data, he might redo a study until the information shows a unique new twist or a conclusion worthy of publication.
As evidence to support his advice, Mr. Tanner led the students through an investigation into a few health care statistics that are commonly heard in the news. ”There are 47 million Americans without health insurance,” was the first statistic. This number shocks us. How can we deprive of 47 million of our fellow citizens of health insurance? Our emotions would urge us to vote for a proposal expanding health care coverage. When we break the broad claim into the more specific parcels of information, however, our emotional reaction is less strong. First, the real number is 46.7 million, 20 million of whom are not U.S. citizens. The number eligible for current programs is 12.1 million, suggesting that one policy response should be an awareness campaign for current programs, not an expansion of coverage. Secondly, 3.5 million of the “uninsured” have household incomes over $75,000. Perhaps they can afford health insurance under current programs, but choose not to buy it. This problem may be out of the hands of policy makers. And 45% were uninsured for four months or less, while 16% were uninsured for two years or longer. This last statistic emphasizes the problem of any survey response-based data set. The “uninsured” number is measured at one given point during the year, and tells us nothing about whether the survey population had insurance previously, is in the process of getting insurance, or has never had access to insurance. In fact, the method of data collection may argue for a graver problem. Many U.S. citizens do not have one form of health insurance that they can count on to last through job changes and moves.
The last section of the evening’s presentation was titled “Assumptions Matter: Garbage In, Garbage Out.” Mr. Tanner argued for the importance of evaluating assumptions when creating or evaluating policy. The Obama health care plan, for example, has been estimated to cost $65 billion, $110 billion, $160 billion, and $600 billion. All reputable research groups, very different numbers. The $600 billion came from Health Systems Innovation, which was paid a $50,000 consulting fee by the McCain campaign. The other three numbers differ based on the assumptions made by the evaluating party.
A question from the audience brought Mr. Tanner to his conclusion. ”So with no unbiased information and a political system that can impede the implementation of even the most logical policy proposals, what do we as policy-makers do?”
“Make sure to conduct the most responsible reading of the data possible, with assumptions that are rational and dependable,” Mr. Tanner responded. ”And then sell policy to the public by choosing to present the data in the way that, while not untrue, best supports your case.” Perhaps this was the most important lesson of all. Knowledge of the biases inherent in information sources is crucial to performing accurate research. Knowing our own biases, and playing off or to the biases of others, is what allows us to move from researching to crafting and building support for our own policy ideas.
Related posts:
Leave a Reply
You must be logged in to post a comment.
Add to Google