Health Benefits for Laidoff Workers

can be characterwed a3 &rat&d samplmg. The standard error 1s a measure of the extent to which an estmmte based on a sample 13 hkely to vary from the pop&&on value of the parameter bemg estimated For this type of sample, the probablhty 13 about 0 68 that an estnnate of earnings, number of workers, or the respective proportlons will differ from t,hose population parameters by not more than one standard error. The probablhty of a difference not more than two standard error3 IS 095 and that of a difference not more than 2?/, standard errors IS 099 One group of population parameters dwussed m the present study 13 the proportion‘of SE1 to taxable earnmgs m the various earnmgs intervals Another group of parametws being estimated 1s the extent and the dlrectlon (posltws or negative) of the change m those proportions between each interval and the lmmedlately followmg interval The dIrectIon of thus change determines whether the average tax rate IS progressmg or regressmg For two given Intervals, If the observed change between them 1s 196 times the standard error It can be said, with only a &percent probablhty of berni: nustaken, that there 13 3ome change 1x1 the SE1 proportlons between these two mtervals m the whole population and that the population change 13 m the same direction a3 the change observed m the sample-that 1s) the change 13 statlstlcally significant. The standard error3 of the ratlo of wages and SE1 to taxable earnmgs are about 0 001 except m the mtervals with the largest number of workers, where they are somewhat smaller The standard errors of the change3 are approximately 00015 to 0 0016