Suchergebnisse

Survey researchers usually assume that retrospective data become less accurate with the passage of time between event & interview. Available research focuses on retrospective behavioral data, although the findings may be assumed to generalize to retrospective attitudinal data. Tested here is the assumption that attitudes in the past are recalled less accurately over time. Data come from a national sample survey (N = 1,491) of adult Americans, collected at the Survey Research Center, U of Mich, in 1973. Analyses utilize the 792 Rs who had least one experience with one of seven government agencies & compare correlations across five different time periods. Results show that people who recall attitudinal information about a bureaucratic encounter which occurred ten or more years in the past generally showed the same pattern of relationships among aspects of the encounter as did people who had contacted a government agency only two or three months prior to the interview. These results have implications for survey research methodology. 2 Tables. AA.

Der Aufsatz gibt eine Einführung in die Konfigurationsfrequenzanalyse (KFA) für Nichtmathematiker (und Anfangssemester). Beschrieben werden: Zweck und Vorteil der KFA, Nutzanwendung der KFA und die Methode selbst. Da die KFA erst durch Computereinsatz als Methode praktisch verwendbar wird, verweist der Autor auf ein Programm, das er auf der IBM 370/158 des Universitätsrechenzentrums der Universität Dortmund erstellt hat. (Interessenten kann es zugeschickt werden). Das Programm wird an einem Beispiel aus der Erziehungsstilforschung demonstriert. Die Konfigurationsfrequenzanalyse (KFA) ist eine von G.A. Lienert (1969) initiierte und später durch J. Krauth und G.A. Lienert (1973) inferenzstatistisch verbesserte Methode zur Auswertung großer Datenmengen. Die KFA ist ein parameterfreies Verfahren, (Normalverteilung der Variablen und Linearität der untersuchten Zusammenhänge werden nicht vorausgesetzt), das vorwiegend zur Darstellung von Wechselwirkungen höherer Ordnung geeignet ist. Es sei zu vermuten, meint der Autor, daß dieses Verfahren wegen seiner Genialität und Einfachheit in der nächsten Zeit immer mehr an Bedeutung gewinnt, so daß die KFA demnächst denselben Stellenwert innerhalb der multivariaten statistischen Methoden haben wird wie z.B. die Varianzanalyse oder die Faktorenanalyse. (KA)

The differential response rates of a mail questionnaire using 3 different stamps on the outside envelope & the same 3 stamps on the inside, or return, envelope are investigated. The stamps used were: metered postage, a regular 1st-class stamp, & an oversized multicolored commemorative. In all, 9 combinations of stamps on the outer & return envelopes were examined. The sample of R's was drawn from the mailing list of the National Forensic League, an organization of high school speech teachers. Every other name on the mailing list (N=530) was sequentially assigned to 1 of the 9 combinations of stamps. The results indicated: (1) when the stamp on the outer envelope is different from the stamp on the return envelope, the response rate is higher (61%) than when the 2 stamps are the same (51)%). This is a significant difference (z=2.36; p less than .01); (2) considering only dissimilar combinations of stamps, a pairing of a commemorative & a meter--in either combination on the inner & outer envelope--is superior (69%) to all other dissimilar combinations (58%). This is a significant difference (z=2.01; p less than .05). 1 Table. AA.

A low-cost method for obtaining acceptably high response rates in survey research is described: namely, the hand delivery of self-administered questionnaires. The R is contacted in his home as in an interview survey. The researcher asks to leave a questionnaire to be filled out at the R's convenience & sets a date for picking up the completed questionnaire. Surveys in Madison, Wis, & Boulder, Colo, yielded response rates of 70 & 84% respectively. Costs were $1.39 per usuable response in the Madison survey & $1.18 per usable response in the Boulder survey. Personal delivery of questionnaires may result in more valid responses than low-cost interviewing since inexperienced personnel probably introduce less error in the course of delivering questionnaires than while conducting interviews. Advantages over the mailed questionnaire are that oral warnings against receiving outside help can be made & that the person picking up the questionnaire can check with the R to make sure he answered the questions himself. The method is most appropriate with regard to samples drawn from geographically compact areas since unit costs increase with geographical dispersion. AA.

Applying sets of techniques aimed at working to one's advantage all of the manipulatable aspects of the mail questionnaire technique of data collection is seldom done. However, by doing so high response rates can be obtained from random samples of the general public. 75% response rates were obtained from 2 statewide samples of 4,500 Wash state residents whose names had been drawn randomly from telephone books. The response to the 10-page questionnaire requiring from 109 to 150 individual answers, was attributed to manipulation of questionnaire format, use of intensive follow-up procedures, anonymity considerations, formulation of appeals following established principles, & use of personalization techniques on a mass scale. Use of postage stamps vs postage meters for mail outs made no diff in returns. While the individual importance of each technique could not be established, follow-ups were found to be indispensable. AA.

Access to powerful new computers has encouraged routine use of highly complex analytic techniques, often in the absence of any theory, hyp's, or model to guide the res'er's expectation of results. Several techniques are examined with respect to their ability to fit random data. The techniques include, AID, aspects of multiple regression, factor analysis, & multidimensional scaling. It is argued that in "exploratory" res, results be replicated (cross-validated) as a more stringent requirement than using statistical signif tests. AA.

The analysis of Ur service delivery is extended beyond equality of units of service input, to measurement of the equity of both inputs & outputs (impact), via development of an original technique for measuring & indexing service equity, which is applied to police services in Tuscaloosa, Ala. The findings are generally consistent with other recent US research, in that only modest inequities in service inputs & outputs, & little evidence of systematic discrimination toward the poor & racial minorities are found. 3 Tables. Modified HA.

"Mit der ... Arbeit soll der Anspruch begründet werden, eine qualitative Methodologie sei notwendig, sowohl als Gegenpart, als auch zur Vervollständigung der insgesamt sehr viel weiter und teilweise, wie in der Stichprobentheorie, der Skalierung und der elektronischen Datenverarbeitung, zu eindrucksvoller Geschlossenheit entwickelten quantitativen Methodologie. Vergleichbare Literatur über qualitative Methodik in den Sozialwissenschaften gibt es nicht. Das Übergewicht der methodischen Beiträge auf Basis der quantitativen Denkweise ist so groß, daß man den Eindruck gewinnen mag, die quantitative Methodik sei die einzige, oder doch wenigstens die "wissenschaftliche" in den Sozialwissenschaften. Daß dem nicht so ist, soll hier gezeigt und begründet werden. Es soll deutlich werden, welchen Platz qualitative Sozialforschung im Gesamtbereich der Sozialforschung einnimmt, welchen methodischen Regeln sie unterworfen ist und welche Besonderheiten sie aufweist. Die qualitativen Methoden im einzelnen können hier nicht behandelt werden; es sollte aber klar werden und wird jedenfalls behauptet, daß jedem bisher bekannten quantitativen Verfahren eine qualitative Methodik gegenübergestellt ist, wobei sich beide Verfahrensgruppen auf charakteristische Weise voneinander unterscheiden." (Autorenreferat)