GN 33246 Quarterly Labour Force Survey Notes for Users of Quarterly Labour Force Survey Series ------------------------------------------------------- Labour Force Survey (LFS) Reweighting Project 2007 -------------------------------------------------- In 2007, the Office for National Statistics (ONS) undertook a reweighting project, whereby LFS data were reweighted using population estimates for 2007-2008. As a result, reweighted editions of LFS back to 1992 were deposited at UKDA during 2008. All applicable QLFS datasets have been updated accordingly. Users should note that in the End User Licence (EUL) datasets: * the single year of age variable (AGE), and 4-digit unit SOC for main occupation, has been reinstated in the reweighted datasets. * only calendar datasets have been reweighted to 2007-2008 figures. The previous seasonal datasets are no longer updated or supported by ONS and have been withdrawn from the UKDA catalogue accordingly. For differences between the EUL and Special Licence (SL) QLFS datasets, see below. Weights ------- After the 2007 reweighting exercise, new weighting variables were introduced: PIWT07 and PWT07 . Please note that the documentation may still refer to weighting variables by their previous names (PWT03 and PIWT03) until the next documentation update. Zero weights ------------ Users should be aware then when running tables at person level, the weighting variable is sometimes missing or has a value of zero. These cases can be ignored, because they are non-responders about whom there is no information other than knowing that they exist. This situation arises if a member of a household was not present at the time of interview, and the respondent was unwilling, or unable, to give any information about that person. LFS Disclosure Risk Changes --------------------------- During 2008, at the request of the depositor, the variables age, dobm, doby, nuts2, soc2kap, soc2kl, soc2kr and soc2ks were dropped from LFS datasets from 2003 onwards, and a banded age variable (ageeul) created. After the 2007 reweighting exercise, year of age (AGE) was reinstated, though some datasets may still include AGEEUL as well. These changes have been made to reduce the risk of disclosure, and will be reflected in all new LFS datasets deposited at the UKDA. Further details may be found in the document 'briefing_note_for_ukda_lfs.pdf' contained in the documentation for affected LFS studies. Special Licence Access datasets are unaffected. Special Licence Access QLFS data -------------------------------- From the January-March 2003 quarter, a Special Licence (SL) version of the QLFS data is also available in addition to the version made available under the standard End User Licence (EUL). The SL version contains extra variables, and therefore is subject to more restrictive access conditions. Prospective users of the SL edition will need to complete an extra application form and demonstrate to the data owners exactly why they need access to the extra variables, in order to get permission to use that version (see 'Access' section of the catalogue record). Therefore, most users should order the standard (EUL) version of the data. In order to help users choose the correct dataset, 'Special Licence Access' has been added to the dataset titles for the SL versions of the data. Typically, variables that can be found in the SL data, and that are not available in the standard EUL version, are: month and year of birth; Nomenclature of Units for Territorial Statistics Level 2 (NUTS2) (county-level); 4-digit Standard Occupational Classification (SOC) for occupation in apprenticeship, last job, second job and job made redundant from; unitary authority/local authority for place of residence and place of work; urban/rural indicator. Data for households of size 10 or above, which are excluded from the standard EUL data, can also be found in the SL data. LFS move from seasonal to calendar quarters ------------------------------------------- In accordance with EU regulations, the LFS moved from seasonal (spring, summer, autumn, winter) quarters to calendar quarters (January-March, April-June, July-September, October-December) in 2006. Following the 2007-2008 reweighting project (see above), calendar versions of all datasets in the main QLFS series were deposited at the UKDA, and the previous seasonal datasets were removed from the UKDA catalogue at the request of ONS, as seasonal data are no longer supported. However, some seasonal datasets may still exist for other LFS series, and ONS advise that, because of the method of construction and the weighting factors used in the datasets, comparison cannot be made between datasets of a calendar and seasonal nature. Time series and longitudinal analysis should only be conducted on datasets of the same type. Further information on the seasonal to calendar quarter change and its impact on LFS data may be found in the following online article: Madouros, V. (2006) Impact of the switch from seasonal to calendar quarters in the Labour Force Survey, London: ONS. Available from: http://www.statistics.gov.uk/articles/labour_market_trends/CQ_article.pdf Identifying Households in the LFS Data -------------------------------------- Households in the survey can be uniquely identified by using REMSERNO (which is a concatenation of WEEK, W1YR, QRTR, ADD WAVFND and HHLD) plus either QUOTA or CASEID, as these are identical. These numbers will remain the same for a household while it is in the survey. If REMSERNO does not appear in a dataset it can be calculated using the following equation: week x 10000000 + w1yr x 1000000 + qrtr x 100000 + add x 1000 + wavfnd x 100 + hhld Persons within households can be identified by using RECNO. Individual Identifier - CASEID ------------------------------ A new variable CASENO has been included in the dataset from the April - June 2002 quarter (SN 5420) onwards. This variable gives every case in the dataset a unique identifier. CASEID is calculated by the following equation: quota x 100000000000 + week x 1000000000 + w1yr x 100000000 + qrtr x 10000000 + add x 100000 + wavfnd x 10000 + hhld x 100 + persno Publication Thresholds ---------------------- For LFS analysis not involving earnings data, ONS recommend that estimates of less than 10,000 people are not used as they are likely to be unreliable. Estimates which do involve earnings data are subject to different thresholds because the sample size is smaller, and the grossing procedure is different. Much of the output from analysis of earnings data is in the form of 'means' or 'proportions', so users need to be aware of the size of the population on which such estimates are based. The threshold when using earnings data is set at 60,000 people for estimates based on wave 5 only (for quarters prior to Spring 1997), and at 30,000 for estimates based on combined waves 1 to 5 data (for Spring 1997 onwards). Users are advised to consult Volume 1 of the documentation, 'Background and Methodology', for further information on this subject.