Supplemental to Discussion on Facts and Reason
Go here for a tool that allows you to calculate the proper sample size for a population, allowing for various error margins.
One thing to consider about arguments that present numbers--polls, statistics and the like--is that the numbers don't really tell you much if you don't have a clear idea how they were obtained. If you see the results of a poll but don't know how the poll sample size was decided and what the poll's margin of error is, you're not getting crucial information that can help you understand what the numbers really mean. A sample size that is too small produces poll results that are likely to be skewed--that is, they don't reflect the real population the poll presumed to question. A high margin of error (above 2%) can make poll results questionable. If a poll shows that 48% of people respond a certain way to a question, for example, but the margin or error is around 5%, then what appears to be a minority opinion may actually be one held by a majority.
Another thing to consider about arguments based on numbers is how they can be built to sound authoritative while distracting you from the real issue at hand. For example, the argument Lunsford quotes by Dick Armey about traffic light cameras certainly cites a series of pretty specific, authoritative-sounding numbers--how much money is earned by the cameras, how many people were cited for tickets before and after a crucial change was made to the way the cameras were set up, and so forth. Note, however, in all his authoritative number-citing, Armey left out one pretty important set of numbers: how many accidents were there at the problem intersections before the cameras were installed, and how many were there once the cameras got set up and running? Did you notice that Armey left out that set of numbers? I would think most of us would tend to allow Armey to frame his discussion as he did, simply because he sounded fairly straightforward in his choice of evidence. If we take a more critical look, though, and ask what numbers are missing from his argument, I think we might at least ask whether his numbers paint a fair picture of the problem he's presenting.
In my view, good arguments are first about working to get at the truth as fairly as you can. In a world that is utterly awash in numbers, I think we have to ask of the arguments that use numbers just how fairly are the numbers being presented. Good, solid data are indispensible parts of a responsibly drawn argument. Numbers that distract, distort or misrepresent do nothing but kill a speaker's credibility. Look thoughtfully at the numbers you are shown, but remember to question them, as well.
One thing to consider about arguments that present numbers--polls, statistics and the like--is that the numbers don't really tell you much if you don't have a clear idea how they were obtained. If you see the results of a poll but don't know how the poll sample size was decided and what the poll's margin of error is, you're not getting crucial information that can help you understand what the numbers really mean. A sample size that is too small produces poll results that are likely to be skewed--that is, they don't reflect the real population the poll presumed to question. A high margin of error (above 2%) can make poll results questionable. If a poll shows that 48% of people respond a certain way to a question, for example, but the margin or error is around 5%, then what appears to be a minority opinion may actually be one held by a majority.
Another thing to consider about arguments based on numbers is how they can be built to sound authoritative while distracting you from the real issue at hand. For example, the argument Lunsford quotes by Dick Armey about traffic light cameras certainly cites a series of pretty specific, authoritative-sounding numbers--how much money is earned by the cameras, how many people were cited for tickets before and after a crucial change was made to the way the cameras were set up, and so forth. Note, however, in all his authoritative number-citing, Armey left out one pretty important set of numbers: how many accidents were there at the problem intersections before the cameras were installed, and how many were there once the cameras got set up and running? Did you notice that Armey left out that set of numbers? I would think most of us would tend to allow Armey to frame his discussion as he did, simply because he sounded fairly straightforward in his choice of evidence. If we take a more critical look, though, and ask what numbers are missing from his argument, I think we might at least ask whether his numbers paint a fair picture of the problem he's presenting.
In my view, good arguments are first about working to get at the truth as fairly as you can. In a world that is utterly awash in numbers, I think we have to ask of the arguments that use numbers just how fairly are the numbers being presented. Good, solid data are indispensible parts of a responsibly drawn argument. Numbers that distract, distort or misrepresent do nothing but kill a speaker's credibility. Look thoughtfully at the numbers you are shown, but remember to question them, as well.
posted by Michael Saunders at 7:33 AM
0 Comments:
Post a Comment
<< Home