OUTLINE FOR CHAPTER 6

1. Introduction
   1. The two main kinds of estimation:  interval estimation & point estimation
   2. Estimation (interval or point) compared with the second main form of inferential statistics: hypothesis testing
2. Interval Estimation
   1. Sampling Error
      1. What a sampling error is . . . and what it isn't?
      2. Should we expect sampling error when a random sample is drawn?
   2. Sampling Distributions and Standard Errors
      1. Sampling distribution . . . a frequency distribution showing the results of multiple random samples of the same size drawn from the same population
      2. Standard error . . . the standard deviation of a sampling distribution
      3. Estimating the standard error from a single sample
      4. Estimate standard errors for different things: mean, proportion, r, etc.
      5. The abbreviation SEM
      6. Reporting estimated standard errors in passages of text, tables, and graphs
   3. Confidence Intervals
      1. What they look like (and how they appear in research reports)
      2. How they're constructed? . . . and 4 things that affect the interval's width?
      3. How they ought to be interpreted?
      4. Their advantage over estimated standard errors
3. Point Estimation
   1. How does one do point estimation?
   2. The big problem with point estimation: it's likely to fail due to sampling error
   3. Places where point estimation is used by researchers:
      1. Reliability and validity coefficients
      2. As a core ingredient when building confidence intervals
4. Warning Concerning Interval and Point Estimation
   1. If two numbers are separated by a "±" symbol, the number that comes after the "±" symbol might be a SD . . . or it might be an estimated standard error
   2. Standard errors are not determined, they're only estimated
   3. Sometimes (as when dealing with r), the sample statistic will not be located exactly halfway between the end points of the confidence interval
   4. Neither interval estimation nor point estimation can be expected to work if based upon nonrandom samples

Copyright © 2003 Schuyler W. Huck
All rights reserved.