How to derive the NS-SEC

To create the NS-SEC on the Census and social surveys, data on occupation and employment status are required.

An NS-SEC category is allocated by using a combination of:

• information about occupation coded to occupational unit group (OUG) level of the Standard Occupational Classification 2000 (SOC2000)

• information about employment status and size of organisation in the form of an employment status variable

Although it is expected that the vast majority of users will use some form of software to combine OUG and employment status to derive NS-SEC, the derivation tables provided can be thought of as a 'look up' table similar to that used for SC and SEG in Volume 3 of the 1990 Standard Occupational Classification (OPCS 1991) which cross-classifies OUGs with employment status categories.

The derived employment status variable is created by combining data on whether an individual is an employer, self-employed or an employee, size of organisation (where collected) and supervisory status.

Employer, self-employed or employee

NS-SEC needs to distinguish employers, those who employ others, the self-employed who work on their own account with no employees, and employees who are employed by an individual or organisation.

Size of organisation

For size of organisation, NS-SEC uses information on the number of employees in the workplace in order to distinguish between employers in large and small establishments and, for some occupations, between higher and lower managers. To date, the distinction between large and small employers has consisted of applying a size rule cut-off of 25 employees. Individual employers in organisations with 25 or more employees are deemed to own 'large' organisations; those owning enterprises below this threshold are classified as 'small' employers. In government social surveys, size of organisation has been related to the workplace, for example the local unit of the establishment at which the respondent works (see GSS 1996:45). The 2001 Census also used this rule. However, it is preferable that size of organisation should refer to an 'enterprise' as defined in the Inter-Departmental Business Register and not to a local unit (IDBR - see ONS 1998:3). Local unit or workplace should only be used if it is impossible or impractical to obtain information at the level of the enterprise.

Supervisory status

Supervisors are employees who are not managers but who are responsible for supervising the work of other employees. Previous surveys have used a variable where the respondent was asked if he or she had managerial duties, supervisory duties, or was an employee. This was used to create the employment status variable regardless of the occupation and with some exceptions. However, under SOC2000 managers can only be allocated to occupations in SOC2000 major group 1. This removes the need to ask for self-reported managerial status and on the Census and future surveys respondents will only be asked if they supervise other employees or not.

Managers

In SOC2000, managers are coded to major group 1 only (see Elias et al. 2000 for a fuller explanation). In the original derivation tables the managerial cells are only valid for SOC2000 codes 1111 to 1239. For these codes the cells for other employees (including supervisors) are left blank. For SOC2000 major groups 2 to 9 the managerial codes are not valid, as managers in these occupations should be coded to major group 1 and, therefore, the managerial cells are left blank in the original derivation tables.

Blank cells

There are other blank cells in the original derivation tables which correspond to situations deemed not to arise, such as a self-employed police officer. In practice some responses to surveys and censuses may correspond to blank cells (for example, as a result of coding error). In previous census and survey practice these were edited according to rules by either changing the occupation or the employment status code so as to achieve an allowable combination. While such editing can correct coding or keying errors, it could also be the case that the original derivation tables do not allow for combinations that do appear in the real world. This can be the case for occupations where self-employment is deemed not to occur but where labour market changes have nevertheless created a new combination.

Three derivation methods

There are three methods to derive the functional categories L1 to L13 of the NS-SEC. The choice of method depends on the information gathered about employment status.

The reduced method was developed for sources unable to collect information on size of organisation. The simplified method provides a last resort solution. If used on its own no records will be allocated to the first category of the classification.

At the eight-class level the reduced method correctly allocates 98 per cent of cases compared to the Full method. The simplified method correctly allocates 83 per cent of cases compared to the full method. For more information about the performance of the reduced and simplified methods compared to the full method.

The following tables show the performance of the reduced and the simplified derivation methods compared to the full method. The data are from the 2000 Summer Quarter of the Labour Force Survey (June-August 2000).