Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
APPEND LX A: TECHNICAL COMMENTS ON SURVEY PLANNING AND SURVEY DES ION LIMITATIONS The following procedure is proposed for continuing assessments of the design implications of specific uses of survey data ant for identifying limitations in survey design for proposed uses. PURPOSE -- IDENTIFICATION OF DESIGN CHANGES NEEDED BY EXAMINATION OF . . . . . ~ ISTING SURVEY DESIGN Decisions about what survey design changes are needed to permit a specified use for the survey data can be made by first asking: Can the specified use be accomplished within the existing design? If not, why not? This permits the specific identification of constraints on the specified data use under examination that arise from the existing survey design. Answers to '~by not" yield identification of changes in survey design that would eliminate existing constraints. The end product would be the identification of design changes required (if any) to eliminate the existing constraint on the data use and thus permit that use or a clear decis ion that no change in survey des ign would permit that use . PROCESS -- A SERIES OF STEPS Tat PROVIDE To INFORMATION NEEDED FOR To CONTINUING ASSESSMENT OF To DESIGN IMPLICATIONS OF SPECIF IC USES FOR SURVEY DATA AND FOR IDENTIFYING T~ LIMITATIONS IN SURVEY DESIGN FOR PROPOSED USES · What is the stated purpose of the survey? · What in the specific use to be examined? · For further definition of that use, what action arises from it? - Program action? Referral to another data base for additional information? Development of a hypothesis to be tested in another data set? - Development of a hypothesis to be tested by planned experiment?
104 · What precise information is required for the scientific analyses required before this action? · What are the present cons traints to making those scientific analys es ? -- Is the existing design competent to permit these required analyses? If not, why not? -- Is there suf ficient scientific understanding to permit interpretation of data relevant to this use ? Knowledge of requirements? Knowledge of biologically acceptable concentrations of contaminants or toxicants ? Knowledge of physiologic response to dietary deficiencies or excesses? In sum: Is the use fees ib le ? In sum: Is a design change required for this use? · What design changers) may be required for this use? SAMPLING DESIGN/SUBJECT DESCRIPTORS · Need for additional information/descriptore for class if ication purpos es . · Inadequate n in the cells that would be used in analysis. · Total survey population, or population included in an analytic cell, is not sufficiently representative of the U. S. population or populat ion group . In sum: What are the impl ications of these for changes on sub ject descriptors to be collected and on design of sampling? UNRELIAB IL ITY OF DATA ~ D IETARY OR OTHER ~ . · Imprecision of data due to misreporting. · Imprecision of data due to food composition tables. · Imprecision of data due to inadequate description of foods. In sum: What are the implications of these for changes in des ign of s urvey ins truments ?
105 ~ INTRAINDIVIDUAL VARIABILITY (IN DIETARY OR OTHER DATA) - In sum: Wha~c i8 the design approach (e.g., replication of tata) needed? Can it be us et within the framework of a survey? INTERINDIVIDUAL VARIABILITY (VARIATION FROM EFFECTS NOT CONTROLLED IN STATISTICS ISIS) · Variability associated with independent effects. · Var lab it ity assoc later with nonindependent e f fee ts . In sum: What is the des ign impl ication of need to control for other sources of variation -- for descriptors to be collected and for sample size? Is statistical control possible? T IME:L INKS S · Does the use depend on rapid turnaround? In sum: What des ign changes would provide required turnaround t ime ? GENERAL SUMMARY -- PROCESS (1) Is the use feasible to implement? If not, stop. 2) Is a design change required for this use? If not, ~ top. 3 ~ What des ign changes are required for this use ? Describe .
