I know you are in a hurry. I am talking about populations of television shows not people. The samples in this case are such a large % of the whole, that I'm not even sure that sample analysis applies. It would be like doing an opinion poll of a population and interviewing 80% of the population. For most of the years of television, there weren't as many shows as today, so it was possible to watch or know about the stories of a high percentage of the whole population of shows; counting in talking to friends and acquaintences one could pretty much review all shows. A show with a new theme like cops as repressors would be reported through this grapevine pretty quickly. So, as the sample approaches 100% , your bias analysis below becomes less and less cogent.
Another way of saying it is that the sample has to be representative of the population. In this case, the population under consideration is the group of cop shows , not some group of people. The statistically correct point that I have been making, which the below is not very cogent to , is that I have a sufficiently representative and unbiased sample of cop shows to say that my generalization that cop shows emphasize "hero" over "repressor" is typical of the total population of shows. Get it ?
Watch Mr.Wizard. It's a great show.
>>> <d-m-c at worldnet.att.net> 02/03 9:59 PM >>>
>Charles: You are not reading my posts accurately. I'm saying that what I
am doing is statisitically good. I have a large sample of data on television shows. My objection is that Paul is making an unscientific objection, based on shows that might be exceptions to the generalization I am making, but some people seem to dispute Paul's facts. I don't have to watch every show. I have a big sample and other people's reports. I had not tried to speak about L&O directly, although now I have evidence on it from Paul and others, so of course I can speak about it.
Charles, since I have a deadline to meet and since I want to get Judy off the ground, this one is gonna be chalked up to coitus interruptus baby.
THREE RULES OF SCIENTIFIC SAMPLING:
1. Every member of the population must have a chance to be in the sample. This is what RANDOM means. It doesn't mean to casually and randomly just sample people on the street.
2.The chance that each member of the population will be selected must be *known* Here we use probability theory to determine precisely the likelihood that a member of the population will be picked to be in the sample.
3. Members of the sample must be *indepedent* of one another; that is, choosing one member doesn't alter the probablity that another member will be picked for the sample. (in other words, you can't pick people on the street and then ask them to give you the name of their bestest bud cause usually people associate with people similar to themselves, who live in the same neighborhood , work at the same place, etc.
You can have the biggest N in the world garnered from your personal viewing and discussions with friends, family, and coworkers, but this isn't in any technical sense "statistically good" You have, in other words, a biased sample.
you can still make an argument about reliability and validity of your findings, but that your data is garnered from sound statistical procedures is not one of them.
CAVEAT: I'm not a big fan of statistical survey methods and believe that they're not useful to Marxist theory most of the time. They are good for tracking general trends and the like. But, statistical survey data deals in data about individuals, not groups. And Marxist theory, on my version, is fundamentally about groups--or in the lingo, classes. The one exception to that claim, is mathematical sociology which seeks to statistically model groups, rather than individuals.
So, your claims about your social location I recognize as an argument that is set forth to justify the objectivity of one's findings and arguments. I have misgivings about what you're saying, but that's for another thread at a later date. I'm such a tease, ya know Charles. Worth the wait though I think.