4
The Current State of Data on Time Use
The number of surveys that collect data on time use around the globe is growing, and some countries have made sustained commitments to collecting data on time use on a regular basis. Table 1, at the end of this chapter, summarizes major time-use studies around the world. Canada has just fielded its third time-use survey and will field one every 6 years. Australia conducted surveys of time use in 1992 and 1997 and will continue to do so every 5 years. The European Union’s statistical agency, Eurostat, is conducting a large scale time-use survey that will collect data across 18 countries. In contrast, large surveys of time use in the United States have been fielded only four times in the past three decades.
To put these developments in context, this section describes time-use data collections that have been conducted in the United States and other countries. It also highlights some smaller, more targeted United States time-use studies that were mentioned during the workshop.
TIME-USE SURVEYS IN THE UNITED STATES
University of Michigan Surveys
In 1965-1966, the University of Michigan conducted a national time-use survey.1 It was part of the Multi-National Time Budget Study (see Szalai et
al., 1972), which included studies in 12 countries in North America and Western Europe. The Michigan Survey was limited to persons aged 18-64 and did not include those who lived in cities with fewer than 30,000 people or households without at least one member working at least 10 hours per week in the nonfarm sector of the economy. More than 1,200 people were in the sample. Interviews collected data on household demographic information; the status and characteristics of respondents’ employment situation; ownership of land, vehicles, residence, and other consumer goods; and media usage and social-psychological measures. In addition, diaries were left with the sample members, to be filled out by respondents during the day after the interview. Respondents recorded the activities they were engaged in, the time the activities began and ended, whom they were with, where they were during the activity, and other activities engaged in simultaneously. The sample was designed so that the number of diaries collected on each day of the week would be equal across the week.2
The Institute for Social Research at the University of Michigan conducted two subsequent surveys of time use. The first was conducted in 1975 and 1976 (see Juster and Stafford, 1985 for details); the second was a follow-up with a subsample of the 1975-1976 sample, conducted in 1981 and 1982. Both studies used approximately the same design. In the first survey, 1,500 adults who were 18 and over (including those over the age of 64) and nearly 900 spouses were interviewed. Multiple interviews were conducted with each respondent over the year, approximately every three months to roughly cover the four seasons.
Unlike the 1965-1966 study, data for the 1975-1976 and 1981-1982 studies were collected through the use of a 24-hour recall diary, in which respondents were asked to record what they did the day before they were interviewed. Respondents were prompted to describe what they had done at one minute past midnight on the diary day. Interviewers also asked the respondent where they were during the activity, who was with them, and whether they were doing anything else at the same time. Then the interviewer asked the respondent what he or she did next and at what time the next activity began, and so on until all the time to midnight the next day was accounted for.
Over the four interviews in the year, data were collected for two weekdays, one Saturday, and one Sunday for each respondent and his or her spouse. The initial interview was conducted in person, and the next three were conducted over the phone. Some background information on respondents was
|
local and smaller scale time use-surveys were conducted prior to the Michigan study; see Bryant and Zick (1996b) for a summary. |
collected for each of the four waves during the year, while other information was collected only for a single wave. The interviews collected basic demographic and household income data, as well as information on employment status and characteristics. Information was also collected on household organization (e.g., who does what tasks), technology and media usage, on the house or residence of the household, and household durables.
The 1981-1982 follow-up sampled almost half of the 1975-1976 sample members and their spouses. For each respondent, interviews were conducted on the same days of the week as in the previous survey. Again, four waves of interviews were given, approximately every three months. The format of the surveys in 1981-1982 were basically the same as in the 1975-1976 study: the only significant change was that proxy reports of time use for up to three of a respondent’s children who were between the ages of 3 and 17 were also collected. The surveys also collected much of the same background data on the respondents and their spouses as the 1975-1976 study did. The 1981-1982 survey also collected data on the social supports available to respondents.
University of Maryland Surveys
In 1985, researchers at the University of Maryland conducted a large-scale national survey of time use, called the Americans’ Use of Time Study (Robinson and Godbey, 1997; Robinson, 1999). Data from over 5,300 persons aged 12 and over were collected. This study also used an open-ended diary survey; however, the method for collecting the data was different from that of the Michigan studies. Random-digit dialing was used to screen households for a respondent at least 18 years or older. This person was then given a 2- to 5-minute orientation interview and was invited to participate in the mail-in diary study. If the person agreed, a diary was mailed to that person and to every member of his or her household who was aged 12 or over. These 1-day diaries specified the day to which they were supposed to refer and were to be filled out as the day proceeded. The survey also included a random-digit telephone survey of day-before activities (for those respondents initially contacted through the telephone sample) and a personal in-home interview for some respondents (as part of a separate sample).
The designated days for which the diaries were to be filled out were spread evenly across days of the week and throughout the calendar year. Data on primary activities and secondary activities were collected. Data on when the activity began and ended, who the respondent was with, and where the activity happened were also collected.
Since the 1985 study, U.S. cross-section time diary studies were conducted as part of this project in 1992-1994, 1995, and 1997-1998. All of these studies covered adults over the age of 18, and the 1992-1994 study included data for children.
TIME-USE SURVEYS IN OTHER COUNTRIES
Although many countries have collected and are collecting time-use data, Canada and Australia have made it a priority to regularly collect data on time use for their populations. Both studies were motivated primarily by a desire to measure unpaid work as input for satellite accounts to the national economic accounts.
Canadian Time-Use Surveys
Canada began collecting time-use data in 1986 and has fielded time-use surveys in 1992 and, most recently, in 1998 as part of a General Social Survey. Lorna Bailie from Statistics Canada reported on the 1998 survey during the workshop.
The 1998 survey collected data on 10,000 households for persons aged 15 and over. A computer-assisted telephone interview (CATI) survey was conducted, with the data collection period running from February 1998 to January 1999. Statistics Canada developed the CATI application of the time diary and introduced on-line coding of the activities during this round of the survey. Both techniques were successfully implemented with extensive monitoring to ensure high-quality results.
The 1998 survey also asked respondents, for the first time, “did this activity help a person outside your household or an organization?” for certain activities. If so, the respondent was further queried as to whether the person was 65 or older, her or his relationship to the respondent, and whether that person had a long-term health or physical limitation.
The survey included a retrospective diary for 24 hours and included questions about where the respondent was, whom the respondent was with, as well as the “for whom” component for certain activities. Data were collected only on primary activities (not secondary ones), but the information was supplemented with a child care diary that detailed when the child woke up and went to sleep and time spent looking after children. There was also a module on spouse’s activities. The survey also included a series of stylized questions about unpaid work, education and learning, employment and working conditions, quality of life, cultural and sports activities, socioeconomic characteristics (such as income, place of birth, religion, language, perceived health status, sleep problems, and type of dwelling).
The survey was approximately 30 minutes long and of that, 10 minutes was spent responding to the diary. The interview began by obtaining a household roster, asking about everyone in the household—age, gender, and relationship. Respondents were allocated to a specific day of the week, and interviewers had 48 hours in which to complete the interview. Statistics Canada reported that the CATI was very helpful in editing and data quality control.
Because the interviewer was able to see the list of respondent reported activities, the amount of time spent for each and the list of individuals in the household, inconsistencies in reporting could be flagged and clarified as the interview was in progress.
Only one person in the household was interviewed, and only one diary was collected for each respondent. Prior to the main survey, Statistics Canada conducted a test in which two persons in the household were asked to complete a diary and two diaries for each person were collected. Nonresponse rates for this test were extremely high, and as a result, the traditional method of one person and one diary per household was subsequently used.
Australian Time-Use Surveys
The Australian survey was described as the “Mercedes of time-use surveys” by workshop discussant Lorna Bailie. This survey interviews all household members 15 years and older and collects time diary information over two days for each respondent. Two of these time-use surveys have been conducted, in 1992 and in 1997.3 Each survey had a sample of approximately 7,000 people who completed diaries for two days of the week. Therefore, in total, a sample of 2,000 diaries were kept for each day of the week. The purpose of the surveys is to provide data to make estimates of the time that individuals spend in different activities for use in policy development and planning.
The 1997 Australian Time Use Survey information was obtained partly by interview and partly by self-completed diaries, which were left with the respondents to record their activities over the two days. One randomly selected member of each household was first interviewed to collect information on household composition, characteristics of individuals in the household aged 15 and older, and the use of technology and outsourcing of domestic tasks (for example, maid services or lawn mowing services). Diaries for each household member over the age of 15 were then left behind to be filled out and collected later. A paper-and-pencil diary was used to collect data for two consecutive days. The diary collected information on what the activity was, when it took place, where it took place, and with whom the activity took place. Data on simultaneous activities and care-giving activities were also collected. Both the 1992 and 1997 surveys included modules for child care activities and information on any disabilities of any people who were receiving care from the respondent.
Eurostat’s Harmonized Time-Use Survey and Pilot Survey
Eurostat, the statistical agency for the European Community, has conducted a pilot time-use survey in 18 countries and is planning to conduct a full study that is harmonized across all countries (see Table 1 for the countries involved in the study). The objectives of the study are to improve national account estimates, better understand time use at work, better understand gender and family policy, understand common transport policy regarding passenger transport and tourism, and measure time use in leisure and cultural activities.
The pilot survey included household questionnaires, individual questionnaires, and time-use diaries. There were a core set of questions to be asked in all countries involved in the study. Individual countries added their own questions to this core.
All members of sampled households over the age of 10 were instructed to complete two diaries for two nonconsecutive days for each individual (one weekday and one weekend day). Time on the diaries was broken down into 10-minute slots, for which the individual described what he or she did. Questions covered secondary activities, who else was present, and who the respondent was helping if applicable. The 1996-1997 pilot study interviewed people in 3,400 households, with a total of 13,600 diaries. In conjunction with the larger Eurostat study, some countries are also conducting their own time-use studies. Italy will be collecting time-use diaries from 30,000 people, including children aged 3 and over. The Bulgarian study will collect diaries from 20,000 people, including infants and children.
OTHER TIME-USE SURVEYS
Each of the surveys described above are general purpose surveys aimed at obtaining nationally representative data for the population. Other time-use surveys have been targeted towards a particular population or group. A number of surveys have focused on the time use of children, to better understand child development, socialization, and well-being. Another study, conducted for the California Air Resources Board, was designed to determine the effect of second-hand smoke. Such studies sometimes ask only stylized questions about time use, such as how many minutes do you use the telephone each day, or how many times do you read to your child. Some have used both time diaries and stylized questions. Others have used the experiential sampling method of collecting data, in which respondents are given a beeper and are paged at random times during the day and asked to report what they are doing. This section describes some of these studies that were discussed at the workshop.
One of the larger time-use studies on children was recently conducted as
a supplement to the Panel Study of Income Dynamics (PSID) at the University of Michigan, and was sponsored by the National Institute of Child Health and Human Development. The study targeted children aged 0-12 in 1995 PSID households. Time diaries over a 24-hour period for up to two children in each household were collected, usually from the child’s primary care giver. Data for 3,600 children were collected. The interviews included some stylized questions and time diaries, which gave a chronological report about the child’s activities over a specified 24-hour period. Diaries were collected for one weekday and one weekend day. Information on simultaneous activities was also collected. Sandra Hofferth reported some results of the study, specifically, on how the time diary data compare with data collected from stylized questions. These results are discussed in the following section. Several other targeted time-use studies for children and youth have been conducted (Almeida, 1997; Larson, 1989; Huston et al., 1997). In addition to children’s time use, the PSID has also collected, on an on-going basis through stylized questions beginning in 1968, data on time spent doing housework.
An example of a more targeted time-use study is the 1987-1988 California Activity Pattern (CAP) Survey. This time-use diary study was funded by the California Air Resources Board to better understand time spent in daily activities “that had implications for air pollution exposure (presence of smokers, use of cooking equipment, use of solvents, etc.)” (Robinson et al., 1994:3).
The CAP Survey is a probability random sample of 1,579 Californians aged 18 years and older in 1987-1988 who have telephones. A random-digit dialing survey was conducted. One eligible household member was randomly selected to be interviewed from each household. One 24-hour diary was collected from each sample member. The days of the week for which diaries were collected were spread throughout the week, with Sundays overrepresented. Diaries were also collected throughout the year except for the months of May and June. In addition to recording times spent in activities, data were collected on the locations of the activities and whether smokers were present during the activity. Interviews also collected general background information on respondents.
From data collected through a 24-hour time-use diary and through direct questions on smoking behavior, estimates of exposure to environmental tobacco smoke have been computed. Among other things, the data were used to estimate exposure to environmental tobacco smoke through the reports of activities, locations, and the presence of smokers.
TABLE 1 Major Time-Use Surveys
Country |
Years |
Sample Size |
Response Rate (%) |
Australia |
1992 |
3,300 households yielding 12,000 diaries |
69 (households); 83 (individuals) |
1997 |
* |
72 (households); 84 of (individuals) |
|
Bulgaria |
1970-1971 |
* |
* |
1976-1977 |
* |
* |
|
1988 |
27,506 individuals in 9,150 households |
98.4 (households) |
|
Canada |
|
||
Halifax-Dartmouth Area |
1971-72 |
2,141 individuals |
* |
National Pilot Study |
1981 |
2,685 households subsampled from 1971 study |
52 |
National Study |
1986 |
12,500 households |
80 |
1992 |
12,765 households |
77 |
|
1998 |
* * |
|
|
Cuba |
1967 |
* * |
|
Denmark |
1975 |
* * |
|
EUROSTAT Harmonized |
1996-1997 |
3,400 households yielding 13,600 diaries |
60-65 over all countries |
Survey |
|
||
Finland, Greece, Ireland, Italy, Luxembourg, Portugal, Spain, Sweden, UK, Albania, Bulgaria, Hungary, Poland, Slovenia, Estonia, Latvia, Lithuania, FYROM |
|
||
EUROSTAT Harmonized |
Proposed 1998 |
* |
* |
Survey France |
|
||
EUROSTAT Harmonized |
Proposed 1999 |
* |
* |
Survey Finland, Italy, UK |
|
Country |
Years |
Sample Size |
Response Rate (%) |
Finland |
1930, 1940s |
Subsamples of school children and farm wives |
* |
1967 |
Subsamples |
* |
|
1969 |
Subsamples |
* |
|
1971 |
Subsamples |
* |
|
1975 |
Subsamples |
* |
|
1979 |
Subsamples |
* |
|
1987-1988 |
10,574 individuals |
74.4 |
|
1990 |
* |
* |
|
France |
1986 |
* |
* |
Germany |
1992 |
* |
* |
Great Britain |
1974 |
* |
* |
1980 |
Subsample of elderly |
* |
|
Hungary |
1963 |
12,000 individuals from 12,000 households |
* |
1976-1977 |
27,607 diaries |
* |
|
1986-1987 |
8,297 diaries |
* |
|
Ivory Coast |
1979 |
3,352 individuals |
56 |
Japan |
1960-1961 |
170,000 diaries |
* |
1965 |
24, 000 diaries |
* |
|
1976 |
* |
* |
|
1980 |
68,000 diaries |
* |
|
1986 |
* |
* |
|
1991 |
* |
* |
|
1996 |
99,000 households with 270,000 individuals |
* |
|
Latvia |
1971 |
* |
* |
1973 |
* |
* |
|
1987 |
891 households |
* |
|
Lithuania |
1974 |
* |
* |
1988 |
984 employed individuals |
* |
|
The Netherlands |
1986 pilot |
* |
* |
1987 |
6,668 individuals in 3,817 households |
47 |
|
1988 |
* |
* |