Date: Thu, 18 Aug 1994 08:44:22 -0400 (EDT) From: Doug Henwood <dhenwood-AT-panix.com> Subject: Re: Marxism and academia Of course you have some valid points here, but I've got answers for most of them. The official unemployment rate is 6.1%, but there are also plenty of numbers published with that headline figure that allow you to make better estimates of a "real" unemployment rate - the number of "discouraged" workers, the number not in the labor force who would like a job, etc. In other words, just the stuff you need to calculate a more accurate social picture of unemployment - in the neighborhood of 10-13%. And yes, the survey uses Census data to weight the sample to extrapolate to the general population. This probably results in understating unemployment as well, since the poor and nonwhite are undercounted, but this hardly invalidates the data. The employment survey uses adjusted figures from the 1990 census - adjusted, that is, to try to correct some for these undercounts. So too the telephone problem; about 7% of US households - disproportionally black and Hispanic - don't have phones, but this doesn't fundamentally invalidate the process. The first interview when a household enters the Current Population Survey program, the source of the (un)employment data, is done in person; follow-ups are done by phone. The Bureau of Labor Statistics' survey of employers has recently been *understating* the rate of job growth, so errors cut both ways. Women are no longer asked different questions from men - and surprisingly, it turns out there is little difference in how they report their employment status from the old days, when the questioning assumed that they were likely to be housewives. Doug Doug Henwood [dhenwood-AT-panix.com] Left Business Observer 212-874-4020 (voice) 212-874-3137 (fax) On Wed, 17 Aug 1994, wesley david cecil wrote: > Doug, you are quite correct that there are tons of useful stats published, > but the question is where do those stats come from. For instance, what > is the unemployment rate? 6.2% or something. Of course this depends on > how you count. For instance, one major aspect, as I understand it, is > phone interviews of randomly selected households. This requires reliance > on a statistical model to represent regional population densities etc. > used to figure out at what point we have a "random" and "representitive" > sample. Then, what are the questions you ask? Recently, women who > answered the phone were asked if they had been working in the home in the > last 6 months, if they said yes, they were not counted as unemeployed. > Of course we can see a problem or two here. Add to this that poor people who > don't have phones are never going to receive this call . . . . stats look > nice on the page but coming up with them is a whole different game. Also > recognize that many Left economists would put the unemployment rate at up to > 12%, once again, it all depends on who youaks, what unemployment is, etc. > etc. > Wes > ------------------
Display software: ArchTracker © Malgosia Askanas, 2000-2005