Example of Replicate Weights

Replicate weights are a series of variables that contain the information necessary for correctly computing the standard errors of point estimates in survey data. The replicate weights themselves are derived from the survey design and processing, and need to be calculated outside of SuperCROSS. Each record in the survey must be assigned a number of replicate weights (anywhere from 30 to up to several thousand times depending on the actual methodology, and the required accuracy of the variance estimate).

These replicate weights need to be stored as columns in the fact table in the source data and then included in the SXV4 when it is channelled using SuperCHANNEL.

Replicate Weights Example

weightings.xml

XML

<DATABASE_WEIGHTINGS dbid="HouseholdSurvey2016" formula_file="formula_household.xml">
	<WEIGHT id="Household Replicate Weights" type="Replicate" level="Household record">
		<CODE>Household Weight</CODE>
		<CODE>Replicate Weight 1</CODE>
		<CODE>Replicate Weight 2</CODE>
		<CODE>Replicate Weight 3</CODE>
		<CODE>Replicate Weight 4</CODE>
		<CODE>Replicate Weight 5</CODE>
		<CODE>Replicate Weight 6</CODE>
		<CODE>Replicate Weight 7</CODE>
		<CODE>Replicate Weight 8</CODE>
		<CODE>Replicate Weight 9</CODE>
		<CODE>Replicate Weight 10</CODE>

		...

		<CODE>Replicate Weight 27</CODE>
		<CODE>Replicate Weight 28</CODE>
		<CODE>Replicate Weight 29</CODE>
		<CODE>Replicate Weight 30</CODE>
	</WEIGHT>
	<COEFF measure="all" weight_id="Household Replicate Weights">
		<VAR id="G" value="30"/>
		<!-- G = 30 Replicate Weights -->
	</COEFF>
	<WEIGHT id="Person Weight" type="Single" level="Person record">
		<CODE>Person Weight</CODE>
	</WEIGHT>
</DATABASE_WEIGHTINGS>

In this example there are 30 replicate weight columns in the fact table. The display names of these columns are referenced in the <CODE> elements within the <WEIGHT> definition.