Comparability — Austria [top]

In 1992, the original variable includes the categories reference person, partner of reference person, child of reference person, parent of reference person, others. Using PARTID, PARNTID1, and PARNTID2 we desegregated the category "others" into sibling, daughter/son-in law, and parent in-laws.

Comparability — Canada [top]

In 2010, the survey does contain information on other household members and relations between the reference person and the people who engage in certain household, paid work or leisure activities. This information does not apply to the time diary element of the survey, and all participants are coded as RELREFP=1.

Comparability — France [top]

In 1998, the original variable does not differentiate between parents and parents-in-law or siblings and other relatives.

Comparability — Netherlands [top]

In 1975 and 1980, relationship categories are limited to the respondent, spouse/partner, and child.

In 2000 and 2005, only the household reference person and the spouse of the household reference person are eligible to be part of a couple.

Comparability — South Korea [top]

In 2004 and 2009, own child and son/daughter in law, and parent and father/mother in law were not able to be distinguished in the source data. The married child and his/her spouse were both coded as child (RELREFP=3), and parents (including the spouse side) were coded as parent (RELREFP=4). Grandchildren and grandparents were coded as ‘other relative’ (RELREFP=9). Only unmarried siblings were identified, because married siblings and their partners were coded as ‘others’ in the original data.

Comparability — Spain [top]

In 2009, the original variables do not include the categories Father / Mother-in-law or Brother / Sister-in-law.

Comparability — United Kingdom [top]

The original 1995 data did not identify siblings or in-laws of the diarist.

Comparability — United States [top]

In 1985, there is no household roster; therefore RELREFP is constructed. RELREFP=1 (reference person), RELREFP=2 (spouse of reference person), and PARTID are quite reliable. RELREFP=3 or RELREFP=4 and the parent ids (PARNTID1 and PARNTID2) are more speculative. Other household relations are highly speculative.

In ATUS, this variable relates to the relationship of the diarist to the CPS reference person. All people in the household aged 15+, including the reference person, had a chance of selection as the diarist for the sub-sample of CPS respondents selected to complete the time diary. Users should note that RELREFP is derived from CPS, which is conducted approximately two months before the diary collection. As a result of this time lag and/or coding errors in the original files, there are some inconsistencies between the RELREFP variable and relevant background variables in respondent and roster files. We check all such cases in all three files, and make necessary corrections with respect to the most recent file unless the information in CPS file is more reasonable.

In each ATUS sample, there are a reduced number of inconsistencies between RELREFP and the household roster. See the syntax harmonization files for more detailed information on these cases.