Fowler ch. 12: providing info about survey methods

• 2 general functions of a good methodological description:

  • provide a good understanding of how well sample estimates are likely to describe the population from which the sample was drawn

  • provide procedural differences between surveys that would affect comparability

• info that should be provided about any survey

  • sample frame with an estimate of the percentage of the population studied that had a chance of selection, and anything known about how the excluded people differ from the population as a whole

  • sampling procedure and any deviations from simple random sampling

  • field results, disposition of the initially designated sample: number of respondents and nonrespondents and reasoning for nonresponse

  • exact wording of the questions analyzed → the entire survey instrument can be included in an appendix for big reports

  • a brief overview of possible kinds of error

  • numerical estimates of the amount of sampling error

  • info about interviews if they were used

  • the effects of nonresponse on sample estimates

  • info about reliability and validity of the major measures used

    • if questions were subjected to cognitive lab testing or systematic pretesting, this and the results can be reported

    • researchers can present analyses that assess the validity of the question answers

    • citing the results of record-check studies of the reporting of similar items can provide readers with a basis for estimating error

• desiderata → documenting how well the measurement was done and estimating the amount and type of error in the results is crucial

D’Ignazio & Klein ch. 6: swimming upstream

• variable and comparative analysis, two regression-based quantitative techniques, overlook the interactive processes that create and maintain racial inequality

• variable analysis

  • scheme of sociological analysis which seeks to reduce human group life to variables and their relations

  • variable analyses of race aim to identify the factors that reduce the coefficient on the race variable, or explain the “race effect”

  • in this analysis, we interpret the independent variables that explain the race effect as the source of racial disparities in an outcome

• comparative analysis

  • scholars compare two or more similar groups to identify the factor(s) responsible for group disparities in success/failure

  • these analyses seek to pinpoint the variable set that distinguishes more successful racial groups from others, and subsequently interpret one racial variable as the source of racial inequality

• quantitative research on race must unveil the interactive processes that create racial inequality to identify the sources of inequality and in turn advise policy designed to alleviate existing racial outcome disparities

• allegory of the swim meet for modeling group difference

  • this allegory suggests three steps for identifying the source of group disparities:

    • recognize the significance of the swim meet structure

    • analyze the impact of prior conditions and reactions on characteristics

    • analyze the covariance between characteristics and swim times while accounting for prior conditions and reactions