Validates the schema of datasets containing outcome variables.

The NlsyLinks handles a lot of the plumbing code needed to transform extracted NLSY datasets into a format that statistical routines can interpret. In some cases, a dataset of measured variables is needed, with one row per subject. This function validates the measured/outcome dataset, to ensure it posses an interpretable schema. For a specific list of the requirements, see Details below.

ValidateOutcomeDataset(dsOutcome, outcomeNames)

Arguments

dsOutcome	A base::data.frame with the measured variables
outcomeNames	The column names of the measure variables that eventually will be used by a statistical procedure.

Value

Returns TRUE if the validation passes. Returns an error (and associated descriptive message) if it false.

Details

The dsOutcome parameter must:

1. Have a non-missing value.

2. Contain at least one row.

3. Contain a column called 'SubjectTag' (case sensitive).

4. Have the SubjectTag column containing only positive numbers.

5. Have the SubjectTag column where all values are unique (ie, two rows/subjects cannot have the same value).

The outcomeNames parameter must:

1. Have a non-missing value

2. Contain only column names that are present in the dsOutcome data frame.

Author

Will Beasley

Examples

library(NlsyLinks) #Load the package into the current R session.
ds <- ExtraOutcomes79
outcomeNames <- c("MathStandardized", "WeightZGenderAge")
ValidateOutcomeDataset(dsOutcome=ds, outcomeNames=outcomeNames) #Returns TRUE.
#> [1] TRUE
outcomeNamesBad <- c("MathMisspelled", "WeightZGenderAge")
#ValidateOutcomeDataset(dsOutcome=ds, outcomeNames=outcomeNamesBad) #Throws error.