Links79Pair dataset.Links79Pair. That
dataset contains the same pairs/rows, but only a subset of the
variables/columns.NOTE: In Nov 2013, the variable naming scheme changed in order to be more consistent across variables. For variables
that are measured separately for both subjects (eg, Gender), the subjects' variable name will have an _S1 or _S2
appended to it. For instance, the variables LastSurvey_S1 and LastSurvey_S2 correspond to the last surveys completed
by the pair's first and second subject, respectively. Similarly, the functions CreatePairLinksDoubleEntered and
CreatePairLinksSingleEntered now by default append _S1 and _S2, instead of _1 and _2. However this can be
modified using the `subject1Qualifier` and `subject2Qualifier` parameters.
Links79PairLinks79PairLinks79PairLinks79PairR. This includes all the R values we estimated, while R (i.e., the variable above) excludes values like R=0 for Gen1Housemates, and the associated relationships based on this R value (i.e., Gen2Cousins and AuntNieces).Links79PairTRUE for Gen1Housemates, Gen2Siblings, and ParentChild. This is FALSE for AuntNiece and Gen2CousinsNo=0, Yes=1, DoNotKnow=255.R is identically constructed, but it did use interpolation.1 and 2, respectively.1 and 2, respectively.ExtraOutcomes79 for more information about its source.ExtraOutcomes79 for more information about its source.Links79Pair.RelationshipPath variable. Code written using this dataset should
NOT assume it contains only Gen2 sibiling pairs. See an example of
filtering the relationship category in the in Links79Pair
documentation.
Please first read the documentation for Links79Pair. That
dataset contains the same pairs/rows, but only a subset of the
variables/columns.
The specific steps to determine the R coefficient will be described in an upcoming publication. The following information may influence the decisions of an applied researcher.
A distinction is made between `Explicit' and `Implicit' information. Explicit information comes from survey items that directly address the subject's relationships. For instance in 2006, surveys asked if the sibling pair share the same biological father (eg, Y19940.00 and T00020.00). Implicit information comes from items where the subject typically isn't aware that their responses may be used to determine genetic relatedness. For instance, if two siblings have biological fathers with the same month of death (eg, R37722.00 and R37723.00), it may be reasonable to assume they share the same biological father.
`Interpolation' is our lingo when other siblings are used to leverage insight into the current pair. For example, assume Subject 101, 102, and 103 have the same mother. Further assume 101 and 102 report they share a biological father, and that 101 and 103 share one too. Finally, assume that we don't have information about the relationship between 102 and 103. If we are comfortable with our level of uncertainty of these determinations, then we can interpolate/infer that 102 and 103 are full-siblings as well.
The math and height scores are duplicated from
ExtraOutcomes79, but are included here to make some examples
more concise and accessible.
library(NlsyLinks) #Load the package into the current R session.
olderR <- Links79PairExpanded$RExplicitOlderSibVersion #Declare a concise variable name.
youngerR <- Links79PairExpanded$RExplicitYoungerSibVersion #Declare a concise variable name.
plot(jitter(olderR), jitter(youngerR)) #Scatterplot the siblings' responses.
table(youngerR, olderR) #Table of the relationship between the siblings' responses.
ftable(youngerR, olderR, dnn=c("Younger's Version", "Older's Version")) #A formatted table.
#write.csv(Links79PairExpanded, file='~/NlsyLinksStaging/Links79PairExpanded.csv',
# row.names=FALSE)
Run the code above in your browser using DataLab