Abstract:

In recent years development has concentrated, almost exclusively, on the application of new technology for improving efficiency of processing and speed of results in survey research. With CATI and CAPI, overnight results of surveys are now possible. Whilst these technological developments have proceeded at break-neck speed, we haven’t even thought there is another avenue to explore, namely, the use of basic statistical procedures. This paper shows: a) For many years we have missed opportunities to make equivalent levels of efficiency gain had we concentrated our efforts equally on exploiting statistical procedures as well as technological ones. b) Now we have the technology, we can combine it with rudimentary statistics to make even further advances. Whilst examples are numerous this paper concentrates on four key areas: a) the application of basic probability theory to minimise the extent of back checking, b) improving estimation for sampling error on quota surveys in order to permit smaller sample sizes on many surveys, c) the application of sequential analysis procedures in line with CATI, CAPI and postal surveys to reduce sample size and improve speed of processing, d) the application of significance testing on tabular output to reduce executive labour time and to gain maximum use of computing resources.