Kinship linking file for pairs of relatives in the NLSY79 and NLSY79 Children and Young Adults

Source: R/links-79-pair-doc.R

This dataset specifies the relatedness coefficient (ie, 'R') between subjects in the same extended family. Each row represents a unique relationship pair.

NOTE: Two variable names changed in November 2013. Subject1Tag and Subject2Tag became SubjectTag_S1 and SubjectTag_S2.

Format

A data frame with 42,773 observations on the following 5 variables. There is one row per unique pair of subjects, irrespective of order.

  • ExtendedID Identity of the extended family of the pair; it corresponds to the HHID in the NLSY79. See References below.

  • SubjectTag_S1 Identity of the pair's first subject. See Details below.

  • SubjectTag_S2 Identity of the pair's second subject. See Details below.

  • R The pair's Relatedness coefficient. See Details below.

  • RelationshipPath Specifies the relationship category of the pair. This variable is a factor, with levels Gen1Housemates=1, Gen2Siblings=2, Gen2Cousins=3, ParentChild=4, AuntNiece=5.

Source

Gen1 information comes from the Summer 2013 release of the NLSY79 sample. Gen2 information comes from the Summer 2013 release of the NLSY79 Children and Young Adults. Data were extracted with the NLS Investigator (https://www.nlsinfo.org/investigator/).

The internal version for the links is Links2011V84.

Details

The dataset contains Gen1 and Gen2 subjects. "Gen1" refers to subjects in the original NLSY79 sample (https://www.nlsinfo.org/content/cohorts/nlsy79). "Gen2" subjects are the biological children of the Gen1 females -ie, those in the NLSY79 Children and Young Adults sample (https://www.nlsinfo.org/content/cohorts/nlsy79-children).

Subjects will be in the same extended family if either:

  1. they are Gen1 housemates,

  2. they are Gen2 siblings,

  3. they are Gen2 cousins (ie, they have mothers who are Gen1 sisters in the NLSY79,

  4. they are mother and child (in Gen1 and Gen2, respectively), or

  5. they are aunt|uncle and niece|nephew (in Gen1 and Gen2, respectively).

The variables SubjectTag_S1 and SubjectTag_S2 uniquely identify subjects. For Gen2 subjects, the SubjectTag is identical to their CID (ie, C00001.00 -the SubjectID assigned in the NLSY79-Children files). However for Gen1 subjects, the SubjectTag is their CaseID (ie, R00001.00), with "00" appended. This manipulation is necessary to identify subjects uniquely in inter-generational datasets. A Gen1 subject with an ID of 43 has a SubjectTag of 4300. The SubjectTags of her four children remain 4301, 4302, 4303, and 4304.

Level 5 of RelationshipPath (ie, AuntNiece) is gender neutral. The relationship could be either Aunt-Niece, Aunt-Nephew, Uncle-Niece, or Uncle-Nephew. If there's a widely-accepted gender-neutral term, please tell me.

An extended family with \(k\) subjects will have \(k\)(\(k\)-1)/2 rows. Typically, Subject1 is older while Subject2 is younger.

MZ twins have R=1. DZ twins and full-siblings have R=.5. Half-siblings have R=.25. Typical first cousins have R=.125. Unrelated subjects have R=0 (this occasionally happens for Gen1Housemates, but never for the other paths). Other R coefficients are possible.

There are several other uncommon possibilities, such as half-cousins (R=.0625) and ambiguous aunt-nieces (R=.125, which is an average of 1/4 and 0/4). The variable coding for genetic relatedness,R, in Links79Pair contains only the common values of R whose groups are likely to have stable estimates. However the variable RFull in Links79PairExpanded contains all R values. We strongly recommend using R in this base::data.frame. Move to RFull (or some combination) only if you have a good reason, and are willing to carefully monitor a variety of validity checks. Some of these excluded groups are too small to be estimated reliably.

Furthermore, some of these groups have members who are more strongly genetically related than their items would indicate. For instance, there are 41 Gen1 pairs who explicitly claim they are not biologically related (ie, RExplicit=0), yet their correlation for Adult Height is r=0.24. This is much higher than would be expected for two people sampled randomly; it is nearly identical to the r=0.26 we observed among the 268 Gen1 half-sibling pairs who claim they share exactly 1 biological parent.

The LinksPair79 dataset contains columns necessary for a basic BG analysis. The Links79PairExpanded dataset contains further information that might be useful in more complicated BG analyses.

A tutorial that produces a similar dataset is http://www.nlsinfo.org/childya/nlsdocs/tutorials/linking_mothers_and_children/linking_mothers_and_children_tutorial.html. It provides examples in SAS, SPSS, and STATA.

RelationshipPath variable. Code written using this dataset should NOT assume it contains only Gen2 sibling pairs. See below for an example of filtering the relationship category in the in Links79Pair documentation.

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.

References

The NLSY79 variable HHID (ie, R00001.49) is the source for the ExtendedID variable. This is discussed at http://www.nlsinfo.org/nlsy79/docs/79html/79text/hhcomp.htm.

For more information on R (ie, the Relatedness coefficient), please see Rodgers, Joseph Lee, & Kohler, Hans-Peter (2005). Reformulating and simplifying the DF analysis model. Behavior Genetics, 35 (2), 211-217.

See also

The LinksPair79 dataset contains columns necessary for a basic BG analysis. The Links79PairExpanded dataset contains further information that might be useful in more complicated BG analyses.

A tutorial that produces a similar dataset is http://www.nlsinfo.org/childya/nlsdocs/tutorials/linking_mothers_and_children/linking_mothers_and_children_tutorial.html. It provides examples in SAS, SPSS, and STATA.

The current dataset (ie, Links79Pair) can be saved as a CSV file (comma-separated file) and imported into in other programs and languages. In the R console, type the following two lines of code:

library(NlsyLinks)

write.csv(Links79Pair, "C:/BGDirectory/Links79Pair.csv")

where "C:/BGDirectory/" is replaced by your preferred directory. Remember to use forward slashes instead of backslashes; for instance, the path "C:\BGDirectory\Links79Pair.csv" can be misinterpreted.

Download CSV If you're using the NlsyLinks package in R, the dataset is automatically available. To use it in a different environment, download the csv, which is readable by all statistical software. links-metadata-2017-79.yml documents the dataset version information.

Author

Will Beasley

Examples

library(NlsyLinks)     # Load the package into the current R session.
summary(Links79Pair)   # Summarize the five variables.

#>    ExtendedID    SubjectTag_S1     SubjectTag_S2           R        
#>  Min.   :    2   Min.   :    200   Min.   :    201   Min.   :0.062  
#>  1st Qu.: 3042   1st Qu.: 304200   1st Qu.: 304301   1st Qu.:0.250  
#>  Median : 5973   Median : 597302   Median : 597701   Median :0.500  
#>  Mean   : 5826   Mean   : 582911   Mean   : 584166   Mean   :0.381  
#>  3rd Qu.: 8418   3rd Qu.: 841852   3rd Qu.: 843201   3rd Qu.:0.500  
#>  Max.   :12675   Max.   :1267500   Max.   :1267501   Max.   :1.000  
#>                                                      NA's   :3681   
#>        RelationshipPath
#>  Gen1Housemates: 5302  
#>  Gen2Siblings  :11114  
#>  Gen2Cousins   : 5000  
#>  ParentChild   :11521  
#>  AuntNiece     : 9899  
#>                        
#>                        
hist(Links79Pair$R)    # Display a histogram of the Relatedness coefficients.

table(Links79Pair$R)   # Create a table of the Relatedness coefficients for the whole sample.

#> 
#> 0.0625  0.125   0.25  0.375    0.5   0.75      1 
#>    309   4466  11201    622  22506     13     38 

#Create a dataset of only Gen2 sibs, and display the distribution of R.
gen2Siblings <- subset(Links79Pair, RelationshipPath=='Gen2Siblings')
table(gen2Siblings$R)  # Create a table of the Relatedness coefficients for the Gen2 sibs.

#> 
#>  0.25 0.375   0.5  0.75     1 
#>  3448   622  7004    13    27