All Times ET
Virtual
Estimating Speech and Language Disorder from a Nationally Representative Sample Using SAS Survey Procedures vs. SAS-Callable SUDAAN (304115)
*Sana Nasir Charania, United States Centers for Disease Control and PreventionMelissa Lauren Danielson, United States Centers for Disease Control and Prevention
Keywords: SAS, SUDAAN, SURVEY DATA
Most large, population-based surveys implement complex sampling methods, such as clustering and stratification, which must be accounted for in the analyses. Additionally, incorporating sampling weights adds to the complexity of analyses for survey data. Until the release of the SAS survey module, common SAS procedures assumed that the data were derived from a simple random sample. Thus, SAS programmers typically used SAS-Callable SUDAAN to conduct analyses on complex data. With the release of SAS survey procedures (9.1v) and SAS-callable SUDAAN, more analytic approaches are possible with survey data.
The objective of this session is to compare the survey module in SAS to SUDAAN using speech and language indicators from the 2016-2018 National Survey of Children’s Health (NSCH) by illustrating examples of syntax and output. The SAS Survey procedures covered are PROC SURVEYMEANS, PROC SURVEYFREQ, PROC SURVEYLOGISTIC, PROC SURVEYREG and their counterpart procedures in SUDAAN.
Attendees will receive the resources to conduct analyses using SAS survey procedures and SUDAAN. This presentation is considered introductory and is targeted to an audience with minimal exposure to the topic.