Nonresponse, Sample Size, and the Allocation of Resources
In: The public opinion quarterly: POQ, Band 39, Heft 4, S. 557
ISSN: 1537-5331
5 Ergebnisse
Sortierung:
In: The public opinion quarterly: POQ, Band 39, Heft 4, S. 557
ISSN: 1537-5331
In: Public opinion quarterly: journal of the American Association for Public Opinion Research, Band 39, Heft 4, S. 557-562
ISSN: 0033-362X
The reliability, precision, or 'goodness of measurement', of any statistic to be generalized from a probability sample to a population is given by the confidence interval around the observed statistic. Survey researchers know that the size (& shape) of the confidence interval is related to nonresponse bias, sampling error, & value of the statistic, eg, the observed proportion of dichotomous answers. The exact interrelationships between these factors are shown in a set of tables. Nonresponse rate is found to be the most decisive determinant of 'goodness of measurement', leading to the conclusion that "once the planned sample is large, increasing its size even dramatically does not alleviate nonresponse bias." Survey researchers should reduce sizes of target samples using the savings in resources for slight increases in response rates by means of relatively more costly techniques of data collection. An approach for estimating optimal allocation of resources among techniques varying in cost is presented. 3 Tables. AA.
In: The public opinion quarterly: POQ, Band 20, Heft 1, Special Issue on Studies in Political Communication, S. 314
ISSN: 1537-5331
In: Public opinion quarterly: journal of the American Association for Public Opinion Research, Band 20, Heft 1, S. 314-320
ISSN: 0033-362X
An analysis of the subject matter of the pictures published in the US illustrated magazine, Life, & the analogous Soviet publication, Ogonek, showed a similar profile & at the same time signif diff's of detail. The pictures were classified into 9 categories: theoretical (pure science & other forms of search for knowledge), econ (including applied science), aesthetic, religious, social, political, avocational achievement, `tensavoidance' (a term coined to designate acts that serve to release tension) & educ. Both Life & Ogonek devoted a very small part of their space to theoretical, religious, & educ'al subjects. Both devoted somewhat over 20% of their pictures to politics. The signif diff's were to be found in the greater weight attached to econ subjects by the Soviet publication (27.6% as compared to 16% for Life) & in the much lower importance attached to tensavoidance' (18% for Ogonek as compared to 30.8% for Life). In the latter case, the difference in quality was even more signif than the difference in quantity: not only did Ogonek devote much less attention to recreational activities, but such pictures as were published in this domain showed only organized forms of recreation; there was not a single instance of pencil-biting or any form of nervous distraction. The conclusion is that the Russian periodical presents the posed ideal of the Soviet prototype whereas Life tries to show real behavior patterns. IPSA.
In: Human relations: towards the integration of the social sciences, Band 1, Heft 4, S. 429-455
ISSN: 1573-9716, 1741-282X