CH. 16 Babbie, Inferential Statistics Memo

- inferential statistics are different than descriptive statistics in that they don't merely describe a sample with regression analysis, but they use a sample to make estimates about a larger population.  in order for various inferential statistical calculations to be relevant, they must follow a few conditions...

1) the sample from which statistical measures are taken must be drawn from the population about which inferences are being made.

2) the statistics assume sampling occurred with replacement, assume simple random sampling occurred in the sample, and assume the sample had a 100% completion rate. if these assumptions don't hold true, it is not usually a serious problem, as long as the sample remains representative of the entire population.

3) the statistics address sampling error only, and not nonsampling error (the imperfections of data quality that are due to things such as misunderstandings of questions by respondents or coding errors).  the calculated level of statistical significance depends on how likely the relationships observed in the sample are due to sampling error alone. however, since nonsampling error is often larger than sampling error, generalizing sample findings to a population must be done with caution.

- tests of statistical significance are statistical computations that indicate the likelihood that the relationship observed between variables in a sample can be attributed to sampling error (unrepresentativeness) only. statistical significance is always expressed in probabilities. the probability of the measured associations being due only to sampling error is called the level of significance.  some researchers specify in advance the level of significance they will regard as sufficient for the purposes of their test, and if the measured association is statistically significant at that level, they'll regard it as representing a genuine association between the two variables.

-  statistical significance is not to be confused with SUBSTANTIVE significance, which means that an observed association is strong, important, meaningful, etc.