Download presentation
Presentation is loading. Please wait.
Published byErnest Mathews Modified over 8 years ago
1
Comparing a 2 and 7 day registration diary in time-use research Evaluating the EUROSTAT guidelines Contact: Joeri.Minnen@vub.ac.be Ignace.Glorieux@vub.ac.be Vrije Universiteit Brussel (Belgium) Paper prepared for the European Conference on Quality in Official Statistics, 8-11 July, Rome
2
European Conference on Quality in Official Statistics, 8-11 July 2008, Rome Time-use registration Goals HETUS-project Since Szalai a number of conventions with regard to the methodology of time budget research have been established Nevertheless: a great variety of recording methods at the moment of recording the information still exists –Fixed / open time interval –Precoded activity list / after coding –Number of diary days recorded –… HETUS: one general guideline for all member countries => Guidelines for recording the time-use information
3
European Conference on Quality in Official Statistics, 8-11 July 2008, Rome Time-use registration Goals HETUS-project Collecting time-use information in a comparable way => Harmonisation To make comparisons between countries possible Major concerns –Design of the diary –Field work Our interest centres on the recommendation regarding the number of diary days
4
European Conference on Quality in Official Statistics, 8-11 July 2008, Rome Time-use registration Recommendations to the number of diary days It is recommended to use two diary days, i.e. one weekday (Monday-Friday) and one weekend-day (Saturday and Sunday) It is strongly recommended that diary days/dates be allocated to households/individuals by a controlled random procedure It is recommended that the survey days/dates are representative of, and cover a full 12 month period
5
European Conference on Quality in Official Statistics, 8-11 July 2008, Rome Time-use registration Presenting the time-use results Calculating estimates for: –Weekdays (Monday-Friday) –Saturday –Sunday –All days of the week (Monday-Sunday) Synthetic workweek: weekday*5 Synthetic weekend: weekend day*2 Synthetic week: (weekday*5)+(weekend day*2)
6
European Conference on Quality in Official Statistics, 8-11 July 2008, Rome Time-use registration Goal this presentation Rationale for 2 vs. 7 days: –Costs of fieldwork –Response rate –Learning/boredom –Number of respondents –Zero observations –Intra-personal variability –More statistical options –… Testing the consequences of the recommendation of recording only 2 days instead of 7 consecutive days
7
European Conference on Quality in Official Statistics, 8-11 July 2008, Rome Time-use registration Daily vs weekly cycles: an example The week cycle is an important cycle in our social life –F.e. working time, work during weekdays, week/weekend contrast, the specific character of Sunday and Saturday in the weekcycle Certain activities follow a weekcycle and occur once or several times a week –F.e. sports, social life, taking part in associations, cleaning, …
8
European Conference on Quality in Official Statistics, 8-11 July 2008, Rome Time-use registration Data used Flemish time budget dataset Respondents filled in 7 consecutive days, starting day was randomly determined Pooled dataset –1999 in total 1394 respondents –2004 in total 1860 respondents Open time interval / precoded activity list
9
European Conference on Quality in Official Statistics, 8-11 July 2008, Rome Time-use registration Creating 2 day dataset Selection of 2 days with the guidelines of EUROSTAT in mind Selection of the weekday (total, n=3.254) was not random => We selected the first registration day if it was a weekday => If not, randomly selected one weekday Selection of Saturday or Sunday (total, n=3.254) was random Weighting –Socio-demographic (sexe, age & education) –Diary days (Mon=Tue=Wed=Thu=Fri); (Sat=Sun)
10
European Conference on Quality in Official Statistics, 8-11 July 2008, Rome Time-use registration Mean time/week, hours and minutes: all respondents 2-day registration 7-day registration Diff. Paid labour20:5120:46-0:05 Housework19:2019:10-0:10 Childcare2:30 0:00 Personal care15:3115:35+0:04 Sleep & rest60:5361:11+0:18 Education2:372:35-0:02 Social participation9:489:46-0:02 Leisure27:0727:10+0:03 Waiting0:090:11+0:02 Travel7:147:12-0:02 Other1:541:48-0:06
11
European Conference on Quality in Official Statistics, 8-11 July 2008, Rome Time-use registration Frequency of activities Workweek (Mon-Fri)SaturdaySunday 1-day5-daysDiff. Paid labour8,0 0,0 Housework18,819,0-0,20,0 Childcare2,82,7+0,10,0 Personal care22,322,4-0,10,0 Sleep & rest11,5 0,0 Education1,21,1+0,10,0 Social part.5,04,8+0,20,0 Leisure12,2 0,0 Waiting0,30,4-0,10,0 Travel11,811,7+0,1 0,0 Other3,43,0+0,4+0,1
12
European Conference on Quality in Official Statistics, 8-11 July 2008, Rome Time-use registration Results in words The estimates on population level between 2-day and 7-day registration for a synthetic week are comparable Positive and negative factors (methodology & quality) keep each other in balance The largest difference is 18 minutes for the activity ‘Sleep & rest’, however as a percentage of the total time spent sleeping and resting it is nearly 0 Essential conditions: –Large(r) sample –With respect to the length of the registration period –Equal dispersion of the diary days
13
European Conference on Quality in Official Statistics, 8-11 July 2008, Rome Time-use registration Mean time/week, hours and minutes: all respondents Random selection (50% of the dataset) 2-day7-day Diff 2-day vs 7-day tot Diff 7- day vs 7- day tot Paid labour21:0520:49-0:19-0:03 Housework19:0219:09+0:08+0:01 Childcare2:282:30+0:020:00 Personal care15:2515:31+0:10+0:04 Sleep & rest61:0661:13+0:05-0:02 Education2:382:34-0:03+0:01 Social participation9:479:42-0:01-0:04 Leisure27:1427:16-0:04-0:06 Waiting0:100:11+0:010:00 Travel7:057:10+0:07+0:02 Other1:531:46-0:05+0:02
14
European Conference on Quality in Official Statistics, 8-11 July 2008, Rome Time-use registration Mean time, hours and minutes: all respondents 1999 (n=1394)2004 (n=1860) Diff. 2-day Diff. 7-day 2-day7-day2-day7-day Paid labour21:2220:4920:2820:43-0:54-0:06 Housework19:3619:3219:0718:52-0:29-0:40 Childcare2:312:282:302:32-0:01+0:04 Personal care15:2415:3915:3615:33+0:12-0:06 Sleep & rest60:5761:1560:5061:08-0:07 Education2:352:39 2:33+0:04-0:06 Social part.9:029:0510:2210:16+1:20+1:09 Leisure26:4126:5227:2627:24+0:45+0:32 Waiting0:090:110:100:11+0:010:00 Travel6:587:037:277:19+0:29+0:16 Other2:392:211:201:23-1:19-0:58
15
European Conference on Quality in Official Statistics, 8-11 July 2008, Rome Time-use registration Results in words Due to a lower sample size, the results are less stable Individual diary days are represented by a smaller number of respondents Some activity groups are less biased, there is an indication that ‘the more a (sub-) population takes part in an activity as a group, the less a large sample size is needed for a good estimation.’ Socio-demographical variables / breakdowns to: –Sexe / age / education / … –Sexe x age / … Closer look with other (international) data -> lager sample
16
European Conference on Quality in Official Statistics, 8-11 July 2008, Rome Time-use registration Variation in different days of the week Kalton (1985) variability across the weekdays, Monday till Friday is relatively small Stinson (2000) finds evidence that the activity pattern of Friday is different from the other weekdays, and also from Saturday and Sunday And, could change even more in the future Szalai T-Test Results pooled dataset 2-day (agg)MonTueWedThuFriSatSun 7-day (agg)MonTueWedThuFriSatSun 7-day (ind)MonTueWedThuFriSatSun
17
European Conference on Quality in Official Statistics, 8-11 July 2008, Rome Time-use registration Participation rate Time-use accounting identity => Ti=PiT’i Ti = mean time population; Pi = population participation rate; T’i = mean time participants Participation rate per day, or average along longer period HETUS Pi(weekday)=(Pi(mon)+Pi(thu)+Pi(wed)+Pi(thu)+Pi(Fri))/5 Participation rate across f.e. weekday 7-day registration Pi(weekday)=Pi(Mon thru Fri) F.e. trafic gestion
18
European Conference on Quality in Official Statistics, 8-11 July 2008, Rome Time-use registration Rhythm of a weekday (Mon-Fri)
19
European Conference on Quality in Official Statistics, 8-11 July 2008, Rome Time-use registration Participation rates Workweek (Mon-Fri)SaturdaySunday 1 day5 daysDiff. Paid labour50,462,7-12,3-0,6+0,3 Housework83,395,3-12,0+0,5-1,3 Childcare21,133,1-12,0+0,4-0,2 Personal care99,3100-0,7-0,2+0,5 Sleep & rest99,9100-0,10,0 Education11,422,4-11,0+0,2-0,3 Social part.53,786,7-33,0-1,4-0,9 Leisure93,699,7-6,1-0,7+0,3 Waiting4,716,4-11,70,0-0,3 Travel73,189,2-16,1+0,4+0,6 Other33,464,8-31,4-0,5+2,0
20
European Conference on Quality in Official Statistics, 8-11 July 2008, Rome Time-use registration Cumulative participation rate Number of days1234567 Paid labour50,457,359,661,162,7 62,9 Housework83,389,592,294,195,397,197,6 Childcare21,126,028,931,233,135,036,5 Personal care99,399,9100 Sleep & rest99,9100 Education11,415,018,120,822,423,524,1 Social part.53,769,377,882,786,791,794,7 Leisure93,698,799,499,699,799,899,9 Waiting4,79,612,414,116,418,419,4 Travel73,180,884,987,389,291,392,2 Other33,443,853,559,564,869,672,7 Sports6,511,615,418,018,820,521,7
21
European Conference on Quality in Official Statistics, 8-11 July 2008, Rome Time-use registration Conclusions Weekcycle is an important rhythm is our society Estimations on population level for a synthetic week on basis of a 2-day registration is comparable with a 7-day registration, the sample size is nevertheless an important factor Participation rate and mean time per participants can not be calculated for a weekcycle To understand individual behaviour it is appropriate to collect information over 7 consecutive days The actual length of the diary period relies on the purpose of the research
22
European Conference on Quality in Official Statistics, 8-11 July 2008, Rome Time-use registration Future tasks Try to find a method / circumstances to extrapolate participation rates Make comparisons with international datasets with a larger sample size to test our findings and broaden our recommendations
Similar presentations
© 2024 SlidePlayer.com. Inc.
All rights reserved.