SOCI 220 Lecture Notes - Lecture 10: Statistic, Statistical Parameter

• School: Texas A&M University
• Department: Sociology
• Course: SOCI 220
• Professor: Judith Linneman

Document Summary

If sample size it too small that not all parts of population are represented, then it is not represe(cid:374)tati(cid:448)e, (cid:449)hi(cid:272)h (cid:373)ea(cid:374)s it"s (cid:374)ot safe to generalize. Do you like my chili recipe: sample: 15% say yes statistic, population: 17% say yes parameter, parameters are typically unknown, do(cid:374)"t k(cid:374)o(cid:449) true (cid:448)alue of populatio(cid:374)s. Sampling for the very reason of trying to find out the estimates: sampling error= any difference between a statistic and its corresponding parameter. How can sampling error be minimized: use high quality sampling method, make sure there"s (cid:374)o sa(cid:373)pli(cid:374)g (cid:271)ias. Estimating sampling error: goal: estimate how far away a sample statistic id from unknown population parameter. That is ho(cid:449) (cid:272)lose is our sa(cid:373)ple to (cid:449)hat"s true i(cid:374) the populatio(cid:374): need to know a whole lot about how sampling works so do it a lot. Take some population that you have already counted. Then take samples, and see how far off that particular sample was from the population.