Beruflich Dokumente
Kultur Dokumente
March 30
2014
Subject : Mobile Propagation Assignment NO 1
Figure 2. The 5istribution of 4ample 6eans !rom the 5istribution 4hown in !igure +, in hich 4ample 4ize / +)
Figure 3. The 5istribution of 4ample 6eans !rom the 5istribution 4hown in !igure +, in
hich
Example
2s a result of this theorem, we can, for example, estimate the population mean of, say, the earnings of female part7time ,ritish employees. 2 random sample of +8- such women in the late +-9)s provided a sample mean of earnings of :0.*8 per hour with a sample standard deviation, s of +.0;, for the particular population under consideration. e can take the sample standard deviation to estimate the population standard deviation, which gives us a standard error of +.0; < +8- = ).)-( . =n the basis of the calculated probabilities of a normal distribution, we know that the population mean will fall within the range 0.*8 > $+ ? ).)-(%, that is, between :0.;9 and :0... with .9@ probability; that it falls within the range 0.*8 > $+.-. ? ).)-(%, that is, between :0.(- and :0.8* with -*@ probability; and that it falls within the range 0.*8 > $0.*9 ? ).)-(%, that is, between :0.(( and :0.9+ with --@ probability. e can be virtually certain that the range covered by the sample mean, plus or minus ( standard errors, will contain the value of the population mean. ,ut as is clear, there is a trade7off between precision and confidence. Thus, we might be content being -*@ confident that the true mean falls between :0.(- and :0.8* while acknowledging that there is a + in 0) chance that our sample mean does not bear this relationship to the true mean. That is, it may be one of the samples whose mean falls into one of the tails of the bell7shaped distribution. The central limit theorem remains critical to statistical inference and estimation in that it enables us, with large samples, to make inferences based on assumptions of normality. The theorem can also be proved for multivariate variables under similar conditions as those that apply to the univariate case.