This page was rebuilt in 2026. The original video and text of this chapter were not preserved. The section "About this chapter" holds only facts recovered from the series; the section "Background" is a new, independent explanation of the topic, based on the sources listed at the end.
About this chapter
The Sampling Demonstration Series was a four-chapter video series on sampling for social norms campaigns, presented by Dr. Larry Hembroff and published on this site. Each chapter paired a video with exercises in SPSS. The final chapter, Sampling Demonstrations Part 4, covers non-response ("Non-Response is a Random Process") and is still available.
According to the series' final chapter, this third chapter showed how to use SPSS to generate random samples, in order to see how representative they were of what was already known to be true of the population. Random samples drawn this way usually turn out to be much more representative of the population than convenience samples.
Background: drawing a random sample in SPSS
When the full population file is available (for example, a data set in which the true values are already known), a simple random sample can be drawn in SPSS with Data → Select Cases → Random sample of cases, either as an approximate percentage of cases or as an exact number of cases. Running the same frequencies on the sample and on the full file shows how close the sample estimates come to the true population values.
Repeating the draw several times shows that random samples vary around the true value, but without the systematic lean that a convenience sample can carry. This is the basis for the margin of error reported with survey results.
Sources
- Groves, R. M., Fowler, F. J., Couper, M. P., Lepkowski, J. M., Singer, E., & Tourangeau, R. (2009). Survey Methodology (2nd ed.). Hoboken, NJ: Wiley.
- IBM SPSS Statistics documentation, "Select Cases" (random sample of cases).