| Sumario: | Sociologists and economists are compelled to employ sampling methods because, like natural scientists, businessmen, government officials, and everyone else, they cannot afford to secure all the data possible before they formulate generalizations or make decisions. Their problem is how to secure a body of data that will serve their particular purposes almost as well as complete information would and at a reasonable cost. They need some estimate of how good these data may be as a substitute for complete information or how serious a risk they are taking that the discrepancy between the data and the results which complete information would provide may exceed a specified degree of misrepresentation. Both in securing the samples and in estimating the risks, they encounter practical problems some of which are of sufficiently frequent and general occurrence to warrant special discussion. One of the simplest applications of sampling is in a partial tabulation of a large file of records or set of schedules when complete tabulation would cost too much or take too long.
|
|---|