Information on the codes for this variable is not currently available.
OCOMBWT reports the original weights that bring the sample in line with the population from which it was drawn.
The episode file does not contain weights because, first, the weights are calculated at the diary level; second, the meaning of an episode varies by the context of the research purpose.
If the survey does not include a weight, OCOMBWT should be set to 0. Some surveys inflate the sample size by a factor to mirror the size of the whole population of the country. If the original weight is inflated, OCOMBWT is left inflated.
Comparability — Index
Users should use caution as these are the original weights. Researchers have to find official statistics describing the population by age and sex. The University of Oxford Centre for Time Use Research recommends the United Nations publication "World Population Prospects," which contains time-series (since 1950) of the population by age and sex for each country. Alternative internationally recognized sources, such as the International Labor Organization Yearbook also may be used. If the survey has enough cases for researchers to split age and sex groups by employment status -- this means if users have at least 50 cases of working and not-working for each sex and age group -- then also include employment status. Note that users may not include employment status for the youngest and oldest diarists if few are working, but include employment status for the working age population.
Comparability — Austria [top]
In Austria 1992, the original weight inflates to the size of the total population.
Comparability — France [top]
In 1985, the original survey weights account for households, all members of households, and the diarists, including the spouse where relevant. We use this latter weight (pondkc) for the diaries, but analysis of the questionnaire elements not included in the MTUS would require use of the pondn (household weight) or the pondk (all household member weight).
Comparability — Spain [top]
For Spain 2002, there are original weights that are defined in the household and the individual level. The weights for the household (FACTOR) are taking into account the location of the household and the day when the household (as a unit) is answering the diary questionnaire. The personal weights (F) are taking into account the day when the individual is answering the dairy questionnaire. Original household and individual weights must be combined (FACTOR*F) to obtain the personal original weight (OCOMBWT).
Comparability — United Kingdom [top]
In 1987, this survey does not have an original weight and the sample is not a random sample, but a travel-to-work area sample from England and Scotland. The survey team analyzed this data using a weight that balances the age and sex group distribution in relation to the age and sex distribution across the UK. The original SCELI survey sampled people aged 20-60. By the time of the diary survey, these people were aged 20-61. The SCELI survey also interviewed spouses where the main participant was in a couple. The diary survey also interviewed all other people in the household aged 14 and older. Very few people aged less than 20 live in households where all members are aged less than 20 or aged less than 20 and/or aged older than 61. This sample is very close to the actual distribution of people aged 14-19. The people aged 62 and older in this sample are relatively representative of people aged 62 and older who live with people aged 20-61. Older diarists are not representative of people 62+. The OCOMBWT weights for the whole good diary sample, but users should use caution and keep in mind that the people aged 62 and older reflect only a subgroup of older people in the UK.
Comparability — United States [top]
In 1975, the original survey weights reflect wave on wave attrition and the national population census distribution. The original weights were inflated to the population size. The original survey weights also account for the fact that the sample is of individuals and spouses have a differential probability of selection. The original weights treat single main respondents as two people, and the spouse and main respondent as one person each to account for this differential probability of selection.
In ATUS samples for 2003 forward, the original weight inflates to the size of the USA population.
- All persons.
- Austria: 1992, 2008
- Belgium: 1966
- Bulgaria: 1965, 2001
- Canada: 2005, 2010, 2015
- Czech Republic: 1965
- Finland: 1979, 2009
- France: 1966, 1985
- Germany: 1965
- Hungary: 1999, 2009
- Israel: 1991
- Italy: 2002, 2008
- Netherlands: 1975, 1980, 1985, 1990, 1995, 2000
- Peru: 1966
- Serbia: 1965
- South Africa: 2000, 2010
- South Korea: 1999, 2004, 2009
- Spain: 2002
- United Kingdom: 1983, 1987, 1995, 2000, 2005
- United States: 1985, 1993, 1995, 1998, 2003-2022