Series details: CRDA/52

The Learning and Training at Work surveys were initiated in 1999 as an annual sample survey of employers in England, conducted by a contractor on behalf of the Department for Education and Employment (DfEE) and its successor, the Department for Education and Skills (DfES). NDAD holds datasets for the surveys conducted between 1999 and 2002. These surveys gathered information on on-the-job and off-the-job training, learning opportunities, employers' awareness of and participation in training initiatives, the training of young employees, and the costs of training. The data was collected to inform policy decision making in these areas by the DfEE and DfES.1

The Learning and Training at Work surveys were introduced to replace the DfEE's Skill Needs in Britain (SNIB) surveys, which were conducted annually between 1990 and 1998. SNIB gathered data on learning and training at work, recruitment difficulties, skills shortages and skills gaps. The increasing emphasis on skills issues on the one hand and employer provided training on the other, and the growing length and complexity of the SNIB questionnaire, led the DfEE to decide to replace it with two separate surveys on training and skills. The Learning and Training at Work surveys were the successor to SNIB on the training side. In the area of skills, SNIB was succeeded by the Employers Skill Survey, which was first conducted in 1999 to gather information on the extent, causes and implications of skill deficiencies.2

The Learning and Training at Work survey has since been replaced with the National Employers Skills Survey (NESS). In 2005, there was a set of questions covering the same questions as those asked in the expenditure section of Learning and Training at Work (LTW) 2000, and some of the questions on incidence and extent of training by employers covered in LTW 2000. The NESS survey is conducted by the Learning and Skills Council (LSC), not DfES. There has been no Learning and Training at Work survey since 2002.

The 1999-2002 Learning and Training at Work surveys were conducted on behalf of the DfEE and DfES by IFF Research Ltd, London. IFF gathered the data using telephone interviewing techniques, and analysed the data to produce a report on each survey (see Publications produced by the originating department).

Responsibility for the 1999-2002 surveys within the Department for Education and Employment and the Department for Education and Skills rested with the Employability and Adult Learning Division of the department's Analytical Services Directorate. In 2002 the Division was renamed Adult Learning and Skills Division; in 2003 it was renamed again, as Adults Division. When the datasets for the 1999-2001 datasets were transferred to NDAD in 2002, direct access to the data files was limited to the Division's Statistics and Research on Skills team. NDAD was told that other parts of the DfES primarily relied on the analysis in the survey reports.3

For further information on Employability and Adult Learning Division, and the DfEE and DfES, see the Administrative History of the Education Departments.

The datasets for the 1999-2002 surveys primarily consist of employers' answers to the main telephone questionnaire, plus a small number of variables from a screening questionnaire which was completed before the main questionnaire. In addition, the 2000 dataset includes data from a supplementary questionnaire (referred to as a "datasheet") on training costs. This was posted to all respondents who indicated, in the main questionnaire, that they had provided some form of training to employees during the previous 12 months.4 While most of the fields in the datasets relate directly to questions asked in the survey (and usually have descriptions which link the field to the original question), the datasets also include a certain number of derived fields, and fields with administrative functions. For example, the 2000 dataset includes a number of fields which indicate whether values in fields relating to training costs have been "simulated" (see How the data was originally captured and validated for further details).

With a few exceptions, the questions in the 1999-2002 surveys asked respondents to supply information relating to the sampled location: i.e. to the individual establishment which had been contacted by the survey contractor. The datasets do not cover employing organisations as a whole, unless the organisation was based on a single site. Each dataset consists of one record per respondent. The records are anonymised: identifying details of employers are not included, and were not passed to the DfEE/DfES by the survey contractor. All types of public sector and private sector employers with one or more employees were eligible for inclusion in the sample, with the exception of schools and local education authorities. For further information on the methods used to collect the data, see How the data was originally captured and validated.5

In order to support comparative analysis of the data, the questions asked in the main questionnaire in the 2000-2002 surveys were based, as far as possible, on those used in previous sweeps of the Learning and Training at Work surveys.6 Consequently, the datasets cover the same general subject areas, though there are differences in the extent of coverage of certain areas and in the individual questions which were asked. The main subject areas are summarised below, with examples of the types of topics covered. More specific information on the contents of individual datasets is provided in the dataset catalogues (see Links to dataset catalogues):

(1) General characteristics of the sampled establishment: e.g. the establishment's industry sector, region, number of employees, types of staff employed (e.g. "managers and senior administrative occupations", "professional occupations"), and whether the establishment was part of a larger organisation.

(2) Staff turnover and recruitment difficulties: the 1999 survey included questions on the number of new recruits in the past 12 months, the numbers of staff who had left in the same period, and the size of the establishment 12 months previously. Both the 1999 and 2000 surveys asked whether the respondent currently had any hard-to-fill vacancies. Questions in these areas were not included in the later surveys.

(3) Young employees (aged 16-24): questions such as whether the establishment had employees aged 16-24; the number of young employees recruited in the past 12 months; whether recent recruits were participating in national training initiatives; what factors were taken into account when recruiting young employees; the methods used to recruit 16-24 year olds; and the proportions of young employees who had obtained or were working towards different levels of qualifications. The 2000 survey included fewer questions in this area than the other surveys.

(4) Skills needs: questions such as whether the need for skills was increasing, decreasing or static; the proportion of existing staff regarded as fully proficient in their jobs; and whether the establishment had built links with outside organisations (e.g. schools, National Training Organisations) to help meet long term skills needs.

(5) Management and delivery of training: questions such as whether the establishment had a business plan, a human resources plan, a training plan and a budget for training; whether the organisation as a whole (not just the individual establishment) had a senior manager responsible for training, and a separate training facility; whether the organisation employed training staff; the numbers of training staff employed in the overall organisation.

(6) Learning opportunities: questions such as the types of learning opportunities offered to staff (e.g. basic literacy, basic numeracy, information technology); whether the employer helped staff to learn things not connected with their work; whether the establishment had a trade union or staff association, and whether it was involved with learning and training; whether the establishment had an equal opportunities policy; whether the establishment was recognised as an Investor in People.

(7) Provision of off-the-job training, defined in the surveys as "all training away from the immediate work position," including training elsewhere on the employer's premises: questions such as whether off-the-job training had been funded or arranged over the past year; the numbers and types of staff who had been provided with off-the-job training; the average numbers of days of training which had been funded for each person receiving training; where training took place; the types of training provided (e.g. induction, health and safety training); the methods used to provide training; whether the employer was satisfied with the training providers; and the types of qualifications, if training was intended to lead to formal qualifications.7

(8) Provision of on-the-job training, defined as "training given at the desk or place where the person usually works": whether the employer had carried out on-the-job training in the past 12 months, and if so, the methods used. The 2000-2002 datasets also record the reasons for not providing training, if an employer had not provided any on-the-job or off-the-job training during the past 12 months.8

(9) Awareness of and involvement with training initiatives: questions such as the employer's awareness of national learning and training initiatives (e.g. National Vocational Qualifications (NVQs), Modern Apprenticeships, New Deal for Young People); whether the employer was involved with training initiatives; whether NVQs or Scottish Vocational Qualifications (SVQs) were offered to employees: the extent of takeup of NVQs/SVQs among staff (numbers who had attained NVQs/SVQs and numbers working towards NVQs/SVQs); whether the employer was satisfied or dissatisfied with NVQs/SVQs, and the reasons for this. The 1999 survey asked significantly more questions in this area than the later surveys.

(10) Training costs: the datasheet on training costs used in the 2000 survey asked for information about the costs of off-the-job training over the past 12 months, and costs of on-the-job training in "a typical month, preferably the last calendar month". Questions on off-the-job training were broken down into the costs of training courses, training centres, training staff and equipment, training organisations, and non-course based training.9

The Learning and Training at Work datasets and related dataset documentation were selected for preservation by the National Archives under theme 2.1.5 (formulation and delivery of social policies) of the National Archives' Acquisition Policy Statement.10

The final Learning and Training at Work annual survey was carried out in 2002. Therefore, no further transfers of datasets to NDAD are expected.

The Learning and Training at Work datasets and related dataset documentation are public records under the Public Records Acts 1958 and 1967. The National Archives has assigned the datasets and documents to class NV 3.

The Learning and Training at Work datasets and related dataset documentation are open without restriction. Data is available for browsing on demand by users of NDAD and does not have to be booked in advance.

The Learning and Training at Work datasets and related dataset documentation are Crown Copyright. Copies may be made for private study and research purposes only.

The Learning and Training at Work datasets are not subject to registration under the Data Protection Act.

The language of the materials is English.

Copies of research briefs and reports relating to the 1999-2002 Learning and Training at Work surveys have been transferred to NDAD, and can be consulted via the Dataset Documentation Catalogue.

Copies of the datasets for the 1999-2002 Learning and Training at Work surveys are also held at the UK Data Archive, study numbers 4181, 4393, 4513 and 4706.

As part of its work on the 1999-2002 surveys, IFF Research Ltd produced reports analysing the results of each survey and summarising the methodology used. These reports were published by the DfEE and DfES in the department's Research Reports series. In addition shorter Research Briefs summarising the main findings of the surveys were also produced by IFF Research Ltd, these were published in the DfES Research Briefs series. Copies of both sets of publications have been transferred to NDAD: see the Dataset Documentation Catalogue.

When the Learning and Training at Work datasets were transferred to NDAD, the DfES's Statistics and Research on Skills team accessed them using networked PCs. It is not known what hardware was used by the survey contractor, IFF Research Ltd.

Microsoft Windows was the operating system for the version of SPSS used by the DfES's Statistics and Research on Skills team in 2002. It is not known what operating system was used by IFF in conjunction with Quancept CATI (see Application software). In 2003 the manufacturer of Quancept CATI specified the following operating system requirements: "SUN Solaris 2.5, 2.6, 2.7 or 2.8 for SUN Ultras, SCO Open Server 5 for Intel Pentiums, [or] HPUX 10.20 for HP 9000s".11

The data for the 1999-2002 surveys was gathered by IFF Research Ltd using Quancept CATI, a computer assisted telephone interviewing (CATI) package produced by SPSS MR, a division of SPSS Inc. The data was exported to SPSS (SPSS Inc's basic statistical analysis software) for analysis by IFF and for transfer to the DfEE/DfES. SPSS 10.1 was used by the DfES's Statistics and Research on Skills team to access the datasets in 2002. By 2003 the team had upgraded to SPSS 11.5.12

Each of the Learning and Training at Work datasets consists of a single table, comprising all the data from one sweep of the survey. The datasets were transferred with original field descriptions which indicate the questions to which the fields relate. Further information, and links to the table catalogues, are provided in the catalogues of the individual datasets (see Links to dataset catalogues).

The Learning and Training at Work datasets are closed, in the sense that once the data from the survey was entered and validated, no further changes were made to the data.

This section summarises how data in the 1999-2002 Learning and Training at Work surveys was gathered, checked and processed. More detailed information is provided in the survey reports (see Publications produced by the originating department).

Survey design

Each of the surveys was preceded by pilot survey, to test the questionnaires and to ensure that the approach conformed with the objectives of the study. Pilots were carried out in October 1999 (for the 1999 survey), May 2000 (2000 survey), September 2001 (2001 survey) and August 2002 (2002 survey). 100 pilot interviews were conducted in 1999, 2001 and 2002, and 200 in 2000, with employers from a range of industry sectors, establishment sizes and regions. Only minor changes were made to the questionnaires as a result of the pilot interviews.

The main stage of interviewing was carried out during the periods specified below:

• 1999 survey: 3 November - 21 December 1999.
• 2000 survey: 17 July - 20 October 2000.
• 2001 survey: 20 October - 4 December 2001.
• 2002 survey: 30 September - 8 November 2002.

The significantly longer time period required in 2000 can be accounted for by the two-stage method adopted to gather data on training costs (see 'Survey questionnaires', below). In each survey, the average length of the interview (including the time taken to collect training costs data in 2000) was 20 minutes.13

Sampling technique

About 4000 interviews in the main questionnaire part of the survey were achieved in the 1999-2002 sweeps, representing a response rate of 62% in 2002, 63% in 1999 and 2001, and 66% in 2000. 924 employers provided responses to the 2000 datasheet on training costs (a response rate of 24%), of which 883 were of sufficient quality to be usable.

The same sampling methodology was applied in each of the surveys. BT's Business Database was used as the sampling frame. This provided a listing of all establishments with a business telephone line, apart from very recent start-ups, businesses with whom BT was in sensitive commercial negotiations, and establishments in the Kingston-upon-Hull area (which was served by a separate telephone company). The sample design involved "quota sampling with stratification by 9 Government Office Regions, 5 industry sectors (defined by the 1992 SIC [Standard Industrial Classification]), and 6 sizes of establishment defined by the number of employees at the location, using variable sampling fractions". Sampling targets were set to ensure that there was a sufficient number of interviews in smaller regions; sufficient interviews in smaller industry sectors; and sufficient interviews with smaller employers, while also providing the oversampling of larger employers which was necessary because of the disproportionate share of employment which they represented. Further information on the sampling methodology is provided in the survey reports (see Publications produced by the originating department).

The interviews were conducted by business to business interviewers in IFF's telephone interviewing centre in London. Interviewers were instructed to seek contact with the owner or managing director, if the establishment had 1-24 employees. In the case of larger establishments, contact was sought with "the personnel or training director or manager / the director or senior manager here who is responsible for training at this establishment". In a small number of cases, there was no person matching this description, and the interview was conducted with the most senior person at the site (whose job title is recorded in the datasets).14

Survey questionnaires

In each survey, all respondents were asked the questions in a screening questionnaire (also referred to as a "screening sheet"), and a main telephone questionnaire. The screening questionnaire recorded the attempts by interviewers to contact an establishment, the outcome of attempted contacts, and basic information about the firm (e.g. number of employees, industry sector) if this was provided. As previously noted (see Scope and content), a small number of variables from the screening sheets are included in the Learning and Training at Work datasets, though only for contacts which led to an interview in the main questionnaire. In the 1999 survey, the wording of questions in the main questionnaire was kept as similar as possible to the Skill Needs in Britain surveys, to ensure comparability with SNIB. Questions in the 2000-2002 surveys were based, as far as possible, on earlier sweeps of Learning and Training at Work. A number of questions on NVQs and the New Deal for Young People, which had featured in the 1999 survey, had to be deleted from the 2000 survey to keep the overall length of the questionnaire and the datasheet within the target of 20 minutes.15

In the 2000 survey, a two stage process was adopted to gather data for the training costs component. Respondents who indicated in the main questionnaire that they had provided some on-the-job or off-the-job training in the previous 12 months were asked if they were willing to answer questions on training costs. The training costs datasheet was posted to those who were willing to co-operate, to allow them time to gather the information by consulting colleagues or written records if necessary. After a few days, respondents were contacted by IFF to collect the data by telephone.16

Computer assisted telephone interviewing (CATI) software was used in each of the 1999-2002 surveys (see Application software). This meant that the questionnaires were held solely in electronic form, with data being entered directly by IFF's interviewers.

Data validation and data weighting

The CATI software used in the 1999-2002 surveys included automatic question routing and some in-built validation checks. The data was also validated by examining "outliers" (responses at the extreme end of a range, which might indicate a coding error); by comparing data with external sources such as SNIB, the Employers Skill Survey and the Office for National Statistics' Labour Force Survey; and by examining year on year changes.17

A particular issue which arose with the 2000 survey was the fact that an employer's training costs could only be calculated if all of the questions in the datasheet were answered. However, some respondents did not answer every question. A process of data simulation was therefore adopted in order to supply data values in cases where a response was missing. Further information on how this was done is provided in the 2000 survey report (see Publications produced by the originating department). The 2000 dataset includes fields which act as flags indicating data items where the value was simulated.18

When the data was analysed to produce the survey reports, the results of the 1999 survey and the non-training cost results from the 2000 survey were grossed up to population estimates derived from the Annual Employment Survey (AES). The 1997 AES was used in 1999, and the 1998 AES in 2000. Different procedures were used to weight the training cost results in the 2001 and 2002 surveys. By the time of the 2001 survey, the AES had ceased to be produced, so results from the 2001 and 2002 surveys were weighted using population data derived from the Inter Departmental Business Register (IDBR). Results from the 1999 and 2000 surveys were re-weighted using the IDBR to allow for comparison between the surveys. Further information on how the survey results were weighted is provided in the survey reports (see Publications produced by the originating department). The datasets contain fields which give the weight factors applied to the data.19

Details of the content and transformation validation checks performed by NDAD on the Learning and Training at Work datasets are contained in the catalogues of individual datasets (see Links to dataset catalogues).

Dataset catalogues provide more detailed information about individual datasets, and are currently available for the following dataset(s):

1. Dataset Documentation Catalogue, reference CRDA/52/DD/1/1, p. 3.

2. Dataset Documentation Catalogue, reference CRDA/52/DD/1/1, p. 3; reference CRDA/52/DD/1/1, p. 3. On the Employers Skill Survey, see information in National Statistics' StatBase database (http://www.statistics.gov.uk/STATBASE/Product.asp?vlnk=9115&More=Y) and the DfES's Skillsbase database (http://www.skillsbase.dfes.gov.uk/Database/Database.asp?sect=11&page=3).

3. Note of telephone conversation between NDAD and the DfES on 14 April 2003.

4. Dataset Documentation Catalogue, reference CRDA/52/DD/1/2, p. 154.

5. Note of telephone conversation between NDAD and the DfES on 14 April 2003; Dataset Documentation Catalogue, reference CRDA/52/DD/1/2, p. 126; reference CRDA/52/DD/1/3, pp. 5, 106.

6. Dataset Documentation Catalogue, reference CRDA/52/DD/1/2, p. 125; reference CRDA/52/DD/1/3, p. 105.

7. The definition of off-the-job training is taken from the 1999-2002 main questionnaires (see the Dataset Documentation Catalogue, references CRDA/52/DD/1/1-4).

8. The definition of on-the-job training is taken from the 1999 and 2000 main questionnaires; an almost identical definition was used in 2001 and 2002 (see Dataset Documentation Catalogue, references CRDA/52/DD/1/1-4).

9. Dataset Documentation Catalogue, reference CRDA/52/DD/1/2, pp. 157-158.

10. Note of telephone conversation between NDAD and the National Archives on 18 June 2002.

11. Website of SPSS MR, "Quancept CATI System Requirements" (http://www.spss.com/spssmr/Quancept_CATI/sysreqs.htm) consulted on 28 May 2003

12. Data owner transfer form for 1999-2001 datasets; notes of telephone conversations between NDAD and the DfES on 14 April 2003 and 23 April 2003.

13. Dataset Documentation Catalogue, reference CRDA/52/DD/1/1, pp. 3, 123; reference CRDA/52/DD/1/2, pp. 3-4, 125; reference CRDA/52/DD/1/3, pp. 5-6, 106; reference CRDA/52/DD/1/4, p. 6.

14. Dataset Documentation Catalogue, reference CRDA/52/DD/1/1, pp. 3-4, 123, 125; reference CRDA/52/DD/1/2, pp. 3-4, 125-126; reference CRDA/52/DD/1/3, pp. 5-6, 105-106, reference CRDA/52/DD/1/4, pp. 6-7, 121-122.

15. Dataset Documentation Catalogue, reference CRDA/52/DD/1/1, p. 123; reference CRDA/52/DD/1/2, p. 125; reference CRDA/52/DD/1/3, p. 105.

16. Dataset Documentation Catalogue, reference CRDA/52/DD/1/2, pp. 126, 154.

17. Data owner transfer form for 1999-2001 datasets; note of telephone conversation between NDAD and the DfES on 23 April 2003.

18. Dataset Documentation Catalogue, reference CRDA/52/DD/1/2, pp. 128-129.

19. Dataset Documentation Catalogue, reference CRDA/52/DD/1/1, pp. 3, 123; reference CRDA/52/DD/1/3, pp. 4, 132; reference CRDA/52/DD/1/3, pp. 6, 109; reference CRDA/52/DD/1/4, pp. 6, 125.