Methodology

This document was compiled by the ISDC - from the CBS original

CHAPTER B. - METHODS

The methods of sampling, estimation and investigation employed in the Labour Force Survey since 1954 when this survey first began, have changed over the years. The methods described here represent the situation following changes, which were incorporated in the new estimate for 1985, based on the results of the 1983 Census of Population and Housing.
A detailed explanation of the previous methods, and the changes that were introduced as a result of the 1983 Census of Population and Housing, can be found in "Labour Force Surveys" 1985 no.801, in the series of Special Publications (page 253).
In the present format, four quarterly surveys are conducted each year, the interviewing for each quarterly survey being spread over the entire quarter. In each survey approximately 12,000 households are sampled. Since some households are interviewed twice a year, the result is that about 22,000 different households are interviewed every year.

1. Population

The survey population includes all the "permanent population" aged 15 and over, as detailed in Chapter D.

2. Frames

A "frame" is a list or file from which a sample is taken. In the Labour Force Survey the sample is drawn from a number of different frames. In the following description of these frames, the distinction is made between the main frames from which most of the sample is taken and secondary frames.

2.1 Main frames

2.1.1 General

As explained below in Section 4.1 the main sampling is done in two stages. In the first stage localities are sampled and, in the second stage, dwellings (or households) in those localities included in the localities sample.
The sample of the localities included in the survey is taken from a frame based on the Bureau's file of localities. From this file, the type of locality, administrative sub-district and number of persons aged 15 and over, are obtained for each locality. For moshavim and qibbuzim, further information is added before the sampling, as detailed below.
For those localities included in the sample, (see par. 4.1 below) a localities frame is obtained.
The coverage of the frame is almost complete: the non-coverage will be described in paragraph 2.3.

2.1.2 The main frame for most localities

is obtained from the Arnona registers of local authorities. Such registers exist for all urban localities and for some of the villages.
Every local council, by law, possesses a property register of the dwellings, other types of buildings, lots, etc., for the purpose of Arnona collection.
For each property there are usually the following details:
a. type of property, i.e. residential dwelling, non-residential dwelling, "other", with certain details.
b. means of identification, i.e. the address of the property and the name of the occupant. There are also other details.
In almost all of the urban localities and therefore, for a very high percentage of residential dwellings, the register is on a computer file containing an entry for each property.
The degree of updating of the registers or files varies from one local authority to another. In some local authorities, the builder registers the dwelling upon completion of construction, whereas in others the dwelling is registered only when the residents take possession of the dwelling.
Generally, files of all the dwellings in the localities of the sample are received once a year, at some date after the end of the fiscal year on March 31. The files received are updated to March 31 or, to the date on which they are copied for the Bureau.

2.1.3 Other locality frames

a. Villages
If there is no property tax register, a list of households of the village is obtained.

b. Moshavim
A list of farms, dwellings or households is obtained.

c. Qibbuzim
A list is obtained of all persons aged 15 and over who reside permanently in the qibbuz.

d. East Jerusalem
In East Jerusalem, because addresses on the Arnona register are difficult to locate, a different kind of frame is used, namely the list of enumeration districts of the 1983 Census. For each of these districts there is a corresponding number of households, based on the Census. This number is updated in those districts, which are sampled according to the count of households in the area.

e. New buildings in urban localities
Once every half a year a list is prepared in the Bureau of new building additions for the localities in the sample. These additions are meant to represent dwellings (households) not included in the frame from which the main sample was taken.

2.2 Special Frames

a. The list of student dormitories and the list of absorption centers. For each student dormitory, the number of persons aged 15 and over living there is estimated and for each absorption center an estimate is made of the number of persons living there. From each of these two lists, a separate sample is drawn each year.

b. Two frames were obtained from the 1983 Census of Population and Housing:

- a list of residents and households within institutions (not including student dormitories).

- a list of Bedouin households in the South living outside of any specific locality.

A sample was drawn from each of these frames, and data from the Census were used in each sample. These data serve as the basis for estimating the appropriate populations in each quarter.

2.3 Non-coverage in locality frames

a. Of the non-institutional population living outside the given localities, only the Bedouins living in the South are represented. (The population not represented in the frame is represented in the estimate, as explained in Paragraph 6.2.2).

b. In the registers used for localities frames (see paragraph 2.1), there is a degree of delay in the updating of new dwellings and occasionally empty dwellings are not registered.

3. Panels and their Mode of Investigation

The sample, which is drawn once a year is divided into four groups or panels as described in Section 4. The four panels are introduced into the investigation during four consecutive quarters, a panel in each quarter beginning in the 4th quarter of the calendar year when the sample is drawn. Each pane1 is investigated four times as follow:

a. Two investigations during two consecutive quarters - investigation number 1 and investigation number 2.

b. After a break of two quarters there are two additioal investigations during two consecutive quarters - investigation number 3 and investigation number 4.

The sample in each quarter is composed of 4 panels from two or three sampling years. In order to characterize the sample upon which the estimate is based - a characterization which is important for evaluating the properties of the estimate, a diagram such as the one below is used from which the composition of the sample investigated in 1990 may be seen.

Sample Diagram of Panel Investigated Quaterly, According to Year Sampled and Investigation Number
(The investigation number appears in the diagram)

In urban localities the sample of dwellings is, in general, divided into four panels. By contrast, in each of the rural localities sampled, the entire sample belongs to only one panel.
The system of investigation by panel was devised for the purpose of providing good estimates for the difference between:
- two consecutive years,
- consecutive quarters,
- parallel quarters from consecutive years.

Note: In urban localities considerable effort is invested in allocating the yearly sample between the four panels in such a way as to increase the sample efficiency (in terms of sampling errors).

4. Sampling

The main sampling is done in two stages. In the first stage localities are sampled from the localities file of the Bureau. In the second stage, dwellings, (or households), are sampled from the frames of those localities included in the sample. The sampling is done in such a way that the final probability of being included in the sample is the same for every household in the population, (about 0.8%).

4.1 Sampling of localities

Distinction should be made between three types of localities:

a. Jewish localities with 4,000 households or more, (about 14,000 persons) as well as Nazareth and East Jerusalem, are automatically included in every yearly sample.

b. Jewish localities with 2,000-4,000 households (about 7,000-14,000 persons) as well as non-Jewish localities of more than 1,000 households (over 6,000 persons) are included once every two years in the yearly sample.

c. Other localities, only some of which are included in the sample. These localities are divided into strata, each stratum comprising an entire locality group or part (the locality groups are listed in paragraph 6). In each stratum, a sample of localities is drawn, the probability of selection of a given locality being proportional to the number of persons aged 15 and over in that locality.

The order of sampling in a stratum of urban localities is according to the number of persons aged 15 and over in the locality. In villages the order is different. Localities of this type are henceforth referred to as "probability localities".

4.2 Sampling in the localities using the main sampling method

4.2.1 Size of sample and its division into panels

In localities from which a sample is drawn every year, the sample constitutes about 0.8% of the dwellings in the locality. In localities where a sample is drawn once every 2 years, the sample constitute 1.6% of the dwellings.
In other urban localities, the sample is about 40 dwellings.
In urban localities, the sample is usually divided into 4 panels, but in some of the non-Jewish urban localities it is divided into two panels.
For non-urban localities sampled, the sample is about 10 households, in qibbuzim about 12 families, and the entire sample belongs to one panel.

4.2.2 Sampling in the localities

Urban localities (not including East Jerusalem)
In localities for which an Arnona file is obtained, non-residential dwellings (business, lots) are eliminated from the file. The file is classified geographically and then a systematic random sample (*) of dwellings is drawn.
In the few localities where the sampling frame is a printed list the sampling is done manually. Afterwards, non-residential places are eliminated from the sample.

_____________________
(*) For example, if 1/100 of the dwellings are sampled, the serial number of the dwellings included in the sample from the file r,r+100, r+200, where "r" is a random number between 1 and 100.
_____________________

Villages and Moshavim
A systematic random sample of households is drawn from the list of local households. Before the third quarter of 1989, two samples of households were drawn: the first for investigations 1 and 2, and the second, a year later, for investigations 3 and 4. After this date, however, the sampling method was changed, only one sample being drawn for all the investigations, as in the urban localities (**).

Qibbuzim
A systematic random sample of persons is taken from the list of persons aged 15 and over in the qibbuz. Each sampled person who is not a head of household is removed from the sample (a zero case), and if the person is head of household, then the entire family is included in the sample. (Every single person aged above 18, is counted as a household).

The sample in the qibbuz is drawn anew for each of the 4 investigations.

4.3 Sampling in East Jerusalem

The sampling includes the following stages:

a. Enumeration districts are sampled from the frame of enumeration districts from the 1983 Census of Population and Housing, after they have been up-dated.

b. The enumeration districts in the sample are mapped and divided into sections, and one section is sampled in each district.

c. All households in the sampled section are interviewed (about 12 households). For the difficulties of investigation since the end of 1987 see Section 5 below.

4.4 Sampling new building

The investigation of the panels in the Labour Force Survey in urban localities continues up to two and a half years after the sample was drawn. Therefore, the original sample of dwellings in the panel does not represent the population which moved into dwellings whose construction was completed after the sample was drawn and before the investigation began. In contrast, in those rural localities where the same sample is used for all four investigations, the survey is completed after a year and a quarter from the time that the sample was drawn. Hence the problem of non-coverage of new building is much less severe.
To remedy this situation, in Jewish urban localities, a sample of additions is drawn. Each half year a sample is drawn from a frame of new building additions for panels sampled in previous years which have still to be investigated during the quarters following the sampling of the additions. Sampling of the additions is taken from new building in all urban localities.

_____________________
(**) In villages and moshavim whose third investigation was conducted before the third quarter of 1989, a new sample was drawn before the third investigation.
_____________________

4.5 Sampling in institutions

The institutions are divided as follows:

a. Student dormitories and absorption centers

Once a year a sample of each type of these institutions is drawn, the sampling probability being proportional to its size. The size of absorption centers is determined by the number of persons there and that of student dormitories by the number of persons aged 15 and over. Within the institutions selected either dwellings or rooms are sampled.

b. Other institutions

A permanent sample i.e. one that does not change is selected. A sample with each unit having a sampling probability of 1:50 is drawn from the "other institutions" in the 1983 Census. The estimates that are presented in the tables of the present publication, were calculated on the basis of data obtained for this sample from the Census (see also paragraph 6.2.2).

4.6 Sampling of Bedouins in the South

Due to the difficulties in locating the Bedouins in the South who live outside localities (***) they are not included in the sample. A sample of 1:25 was drawn from this population group in the 1983 Census. The estimates that are presented in the tables of the present publication were calculated on the basis of data obtained for this sample from the Census.

4.7 Size of the Sample

On average, in 1990, approximately 22,000 persons aged 15 and over were interviewed quarterly, (excluding Bedouins in the South and residents of institutions).
The sample of institutions includes about 1,500 persons and the sample of Bedouins about 400 persons.

_____________________
(***) These localities are Rahat, Tel-Sheva, Kseifa and 'Aro'er.
_____________________

The breakdown of persons aged 15 and over interviewed quarterly, is as follows:

5. Enumeration and Questionnaire

The enumeration is conducted by trained enumerators of the Central Bureau of Statistics who visit every dwelling in the sample. Dwellings not used for residential purposes, empty apartments, businesses etc., are considered zero cases. In residential dwellings, the enumerators interview one of the household members from whom they receive information about all members of the household. For households where nobody is at home on the first visit, the enumerator usually makes two further visits. If also on these visits no one is at home, special questionnaires are usually left to be filled in and returned by post. For each visit that no household member is at home, the enumerator indicates the reason in a "Non-investigation" questionnaire. A certain percentage of households included in the sample are not enumerated due to absence of household members, refusal to answer, inability to locate etc.. Households that do not fill in the questionnaire are considered non-respondents.
On the first visit to the family, while filling in the questionnaire, the enumerators have to find out whether the households sampled have telephones and if so, to note their numbers.
In urban localities the enumerators were asked to receive the permission of the households with telephones to conduct stage B and C by telephone. In 1990, about 42% of the interviews were conducted by telephone.
During the years 1954-1956 the enumeration was carried out during one week each year, in 1957 during two weeks and in 1958 and 1959 one week during each quarter. As from January 1960, the enumeration is carried out continuously each and every week for the entire three months of the survey. During each week, about 1/13 of the households included in the survey are interviewed. The "determinant week" in this period always refers to the "previous week", namely the week ending on the Saturday prior to the visit of the enumerator. The data obtained for any period (quarter - year, year, etc.) reflect the situation during an "average" week in this period.
For each household one questionnaire is completed containing information pertaining to the entire household and to each member aged 15 and over.
In 1978 the investigation of some of the topics in the regular questionnaire was extended (for details of changes, see the publication, Labour Force Surveys 1978. Special Series No. 653, Appendix 1). The questionnaire for an individual includes questions on work in the determinant week, number of work hours in general and in the determinant week, number of work hours less than usual in the determinant week, number of overtime hours, reasons for part-time work, reasons for absence from work during the entire determinant week or a part of it, place of work, geographical mobility of all employed persons, type of work and employment status. In addition, information is obtained on search for work, search for full-time or part-time work, reasons for unemployment, whether the individual ever worked in Israel and when, and previous employment of unemployed persons. The amount of annual work of employed persons and reasons for working only part of the year are also investigated. Those persons not in the civilian labour force but who had worked during the year preceding the interview, were asked about the type of work they had done. Those not in the civilian labour force and who had not worked during the year preceding their interview, were asked why they did not work.
Apart from details about work, demographic information is also collected: age, sex, marital status, country of birth, period of immigration, level of education (number of years of schooling and type of last school attended).
The household questionnaire includes information on the number of persons in the household, the number of rooms in the house, the number of children in the household and the number of hours of work of any paid domestic help.
Besides the regular questions about work, household characteristics, and the demographic questions mentioned above, the questionnaire may occasionally relate to other matters such as housing conditions, domestic appliances in the possession of the household and various other subjects. The inclusion of these subjects in the Survey is intended to provide statistical material for a detailed investigation of the connection between work patterns of the household and other areas of behaviour. Also included occasionally in the Survey are questions related to specific aspects of work such as seniority at work, labour mobility and so on. They are intended to provide further information on household and individual behaviour in the work field.
Because of the events in East Jerusalem since the end of 1987, the enumerators of the Labour Force Survey could not enumerate the families in the sample as planned. To make possible the first stage of enumeration, those preparing the maps of the sampled enumeration districts were requested to obtain the telephone number of the families in the sample. The data processed were collected by telephone. Therefore caution is recommended in using the data of the non-Jewish population in East Jerusalem from 1988 and onwards.

6. Processing of material and estimates

6.1 General

The collected material is checked, completed, edited, coded, punched and undergoes logic editing. Following these procedures, estimates are calculated for each quarterly sample. The annual estimates presented in the tables of this publication are averages of the four quarterly estimates.

6.2 Inflating coefficients and estimates (excluding Jewish immigrants arriving in 1990)

In order to translate the sample data into estimates for the population as a whole, the sample results must be "inflated". The population is divided for this purpose, into "inflating groups" defined by age x sex x type of locality and geographical area (district and sub-district).

6.2.1. Inflating groups

Starting with the data of 1986, localities of 2,000 or more persons are classified as urban, the others being rural. From 1986-89 the distinction between these two types of locality was determined by population estimates obtained in the 1983 Census, and from 1990 by population estimates at the end of 1988 (see Chapter E for a fuller discussion). The locality groups are as follows:

Groups of Jewish urban localities:
1. Western Jerusalem
2. Tel Aviv - Yafo
3. Haifa
4. Northern District
5. Haifa District (excluding the city Haifa)
6. Tel Aviv District (excluding the city Tel Aviv - Yafo)
7. Sharon and Petah Tiqwa Sub-District (part of the Central District)
8. Ramla and Rehovot Sub-District (part of the Central District)
9. Development localities in Southern and Jerusalem Districts
10. Localities in the Southern and Jerusalem Districts which are not development localities, as well as Jewish urban localities in Judea, Samaria and the Gaza Area.

Groups of Jewish non-urban localities
11. Qibbuzim
12. Moshavim and collective moshavim
13. Other rural localities.
This group includes those places outside the localities (from which a sample was not drawn).

Groups of non-Jewish urban localities
14. East Jerusalem
15. Localities with 10,000 persons and more
16. Localities with 2,000-9,999 persons

Groups of non-Jewish, non-urban localities
17. Non-urban localities
18. Bedouins in the South who live outside any specific locality.

Age groups for the locality groups are:

a. Jewish urban localities:
1. 15 - 17        6. 55 - 59
2. 18 - 24        7. 60 - 64
3. 25 - 34        8. 65 - 69
4. 35 - 44        9. 70+
5. 45 - 54

b. For the other locality groups, persons aged 55+ are divided into two groups only, 55-64 and 65+.

6.2.2. Inflating coefficients

There are two sets of inflating coefficients:

a. One set for the regular population (including absorption centers (****) and student dormitories).

b. A second set for institutions (excluding absorption centers and student dormitories).

The inflating coefficients are obtained as follows:

a. Population estimates are calculated for the inflating groups: from the Demography Section the size of each population group (age x sex, both for Jews and non-Jews) is obtained every quarter (updated from reports of births, deaths, immigration and emigration). These numbers are then multiplied by the percentage of each type of locality within each group, also obtained from the Demography Section at the end of every year and subsequently corrected each quarter, starting from the second quarter of the year.

b. Population estimates are calculated for institutions (*****) and their derived inflating factors.

c. The estimates of the size of the regular population in the inflating groups are obtained by subtracting the institution population size from the regular population size.

d. The inflating coefficients are obtained by dividing the population estimate in the inflating group, which was obtained for the middle of the quarter to which the Survey relates, by the number of cases in the (quarterly) sample for that inflating group.

_____________________

(****) Since 1985, when population estimates have been based on the 1983 Census of Population and Housing. Before 1985, when population estimates were based on the Census of 1972, absorption centers were not included in the population estimates.

(*****) Based on the assumption that the percentage of the institutionalized population in the inflating groups did not change since the time of the Census.
_____________________

The quarterly estimates are obtained by multiplying each datum by its inflating coefficient. The annual estimates are the average of the four quarterly estimates.
The method described above ensures that the weight of each inflating group in the estimate is correct and thus reduces biases resulting from differential non-response between inflating groups. In addition, it also reduces the sampling error of the estimate (For explanation of sampling error, see chapter C).

Note
For the purpose of calculating inflating coefficients, estimates of the Jewish and non-Jewish population of villages include persons living outside the locality, with the exception of Bedouins in the South. This means that although persons outside of localities are not included in the sample, they are nevertheless represented in the estimates by means of persons living in villages.

6.3 Inflating coefficients and estimates for Jewish immigrants of 1990

To ensure a correct representation of the immigrants who arrived in 1990, in the population, it was decided to allocate them to separate inflating groups, defined by age (as detailed in par. 6.2.1.). This was done starting in the third quarter of 1990.
Demographic estimates were obtained for sex x age groups for these immigrants, who were then divided between the regular and institutionalised populations, on the assumption that the percentage in institutions was the same as in the rest of the population. Inflating coefficients for the regular population were then derived in the same way as described in par. 6.2.2.
For the institutionalised population, a copy of the file containing the sample in institutions (as described in par 4.5) was made. The cases in this new file served as a sample of the immigrants in institutions.

CHAPTER C. - RELIABILITY OF THE ESTIMATES

The data presented in this publication are estimates for the population as a whole and were calculated on the basis of a survey conducted amongst only a sample of the population. As such, they are prone to errors, which may be of two types:
1. Sampling errors
2. Non-sampling errors

1. Sampling Errors

Sampling errors occur because only a sample and not the entire population were investigated. The particular sample used in this survey is one of a large number of possible samples of the same size that could have been selected using the same sample design. It is clear that the estimates derived from the different samples will differ from each other, and almost all of them will differ from the value obtained had a complete census been taken- "the census value".

In view of the variability of sample estimates, it may be preferable to consider a range of values where the census value is likely to be found, with a given probability (or level of confidence) rather than to rely on the specific value of the estimate obtained from the sample. This can be done using the sampling error, which is a measure of the variability of the estimate due to the random manner in which the sample is selected.

Since we can say that the changes are about 67 out of 100 that an estimate from the sample would differ from the census value by less than the sampling error, it therefore follows that, with the same degree of certainty, i.e. 67% of the census value will be found within one sampling error of the sample estimate. In other words, we can say with 67% confidence that the census value lies within the interval bounded below by the sample estimate minus its.sampling error and above by the.sample estimate plus its sampling.error. Similarly, we can say with 95% confidence that the census value lays in the interval between the sample estimate minus two sampling errors (below) and the sample estimate plus two sampling errors (above).

Intervals defined in this way, that with a prescribed level of confidence, for example 67% (or 95%), they contain the census value, are known as "67% (95%) confidence intervals".

The Methodological Appendix contains tables of approximate sampling errors together with instructions on how the tables may be used to obtain sampling errors for the various estimates presented in this publication and illustrations are given how to use the sampling errors to construct confidence intervals.

It should be mentioned that it is sometimes more convenient to measure sampling error in terms of the relative sampling error, defined as the sampling error divided by the estimate.

Estimates whose relative sampling error exceeds 20% are enclosed in parentheses (or denoted..) in order to caution the reader of their low reliability. In some cases, parentheses are used for estimates with even smaller relative sampling errors (******).

2. Non-Sampling Errors

Non-sampling errors may stem from many "sources in the various stages of collecting and processing the survey data and may occur equally in a full census. Of the many possible non-sampling errors, with relevance to the present publication, the following should be mentioned:

1. Errors, which stem from non-response: errors, which are caused by the fact that households are not investigated due to absence from home or refusal. This may cause some bias in the estimates since the characteristics of persons belonging to these households may differ from those of persons who were investigated.
In 1990 89% of the households supposed to be investigated were in fact enumerated (compared to 88% in 1989). The total from which this percentage is calculated does not include cases not meant to be investigated such as unoccupied dwellings, non-residential dwellings and persons not belonging to the survey population.

2. Response errors: errors due to misunderstanding of the questions unwillingness or inability to answer correctly, or mispresentation of questions.

3. Errors in processing: errors that occur at the stage of processing the material, such as errors in the coding and punching of the questionnaires. Some of these errors are corrected by means of checks that the material undergoes.

4. Some of the households were interviewed in a week, which was not the "determinant week" (as defined in Chapter B, paragraph 5). This too causes a bias in the estimates.

In contrast with sampling errors, which can be estimated on the basis of the survey data, non-sampling errors are difficult or even impossible to estimate. Thus emphasis is laid on controlling such errors, rather than on indicating their magnitude in the data.

For estimates relating to years other than 1990, it is possible that estimates enclosed in parentheses have relative sampling errors lower than those given above.

______________________
(******) To simplify the process of denoting unreliable estimates, different levels of relative sampling errors were used for different types of estimates:
(1) For some frequency estimates (for absentees and unemployed) and for most percentage estimates
______________________

CHAPTER D. - DEFINITIONS AND CLASSIFICATIONS

A. Population and Labour Force

1. Population - The survey population includes the permanent (de jure) population of Israel aged 15 and over, including the Jewish residents living in Judea, Samaria and the Gaza Area, potential immigrants and permanent residents living abroad for a period of less than one year. It does not include tourists and temporary residents unless they have been living in Israel continuously for more than a year. As of 1968, the survey population includes the residents of East Jerusalem. As of 1982, the survey population includes the non-Jewish population of the Golan Sub-District.

Some of the tables do not include all the population (see the notes to the tables themselves).

2. The determinant week is the week ending on the Saturday preceding the visit of the enumerator to the household.

3. Characteristics of the civilian labour force

3.1 Civilian labour force - persons aged 15 and over who were "employed" or "unemployed" during the determinant week, according to the definitions given below.

3.2 Employed persons - persons who had worked for at least one hour during the determinant week, at any type of work, for pay, profit or other consideration; all those working in qibbuzim (whether in services or any other branch); family members who had worked without pay for more than 15 hours during the investigation week; persons in institutions who had worked more than 15 hours during the week. Persons who were temporarily absent from their work and did not seek other employment are also included in the employed population.

Employed persons are divided into tree sub-groups:

a. full-time workers - all those who had worked 35 hours or more during the determinant week, (including preparation hours).

b. part-time workers - those persons who had worked 1-34 hours during the investigation week (including preparation hours).

c. temporarily absent from work - this group includes all those who had been temporarily absent from their regular work during the determinant week, due to illness, vacation, reserve army service, strike, unsuitable weather, temporary disruption of work for up to 30 days, etc. Those absentees seeking work during their absence are not included.

3.3 Unemployed persons - all those who had not worked for even one hour during the determinant week and who had actively sought work during that week, by registering with the Labour Exchanges of the Employment Service or any other employment office, either in person or by mail, by attempting to establish an independent business etc. Workers who had been temporarily absent from their work and who sought other employment are considered unemployed.

Unemployed persons are divided into two sub-groups:

a. unemployed persons who had worked in Israel during the twelve months preceding the determinant week.

b. unemployed persons who had not worked in Israel during the twelve months preceding the determinant week.

3.31 Seeking work

The possible ways of actively seeking work are:

- Labour Exchanges of the Employment Service
- Other employment bureau
- Adverts in newspapers
- Application to employer either in person, or in writing
- Friends or relatives
- Attempt to establish a business
- Sought work in some other ways.

Each interviewee is requested to answer on every one of these ways. For publication purposes, we create a variable - "the principal way of seeking work", for each interviewee, which is determined by the rank order of ways in the questionnaire.

3.4 Not belonging to the civilian labour force - all persons aged 15 and over who were neither "employed" nor "unemployed" during the determinant week. Included in this group are students, non-paid volunteers, housewives who had not worked for even one hour outside their homes, persons not fit to work, persons living off their pensions or other income, who did not work for even one hour during the determinant week. Also included in this group are soldiers doing their regular army service, (conscription or in the permanent army), family members who had worked without pay less than 15 hours per week, persons in institutions who had worked less than 15 hours during the determinant week.

4. Years of study - the number of years spent in regular studies in school and not studying on one's own, or attending irregular courses. The number of years consists only of completed years. If a person is studying at the time of the enumeration, that year is counted.

5. Type of last school attended - by the investigated person, even if he did not complete his studies there.

6. Head of household - the oldest male earner (employed, unemployed or serving in the regular army) in the family is considered the head of the household. If there is no male provider, the head of the household is the oldest female earner. If there are no providers, the oldest male family member is considered the head of the household, and if there is no such male, then the oldest female is the head. This is all conditional upon the head of the household being 18 years of age or more, except for qibbuzim and institutions where youths under the age of 18, who have no family in the qibbuz or the institution are considered the head of the household. In addition, every person aged 18 and over in a qibbuz or institution is considered a head of household, unless he is married and the spouse is considered the head of the household.

7. Number of children, by age group - information is obtained on the number of children born to each non-single female in the household. In addition, all the children in the household are counted, even if they were not born to the females belonging to the household. Based on this, two series of data on female employment are processed: in the first series, females are classified according to the number of their children and in the second series they are classified according to the number of children living in the household. The data are presented in two sections in the chapter on female employment.

8. Work hours per week - the number of work hours per week of employed persons or employees, during the investigation week, includes: overtime in the same place of work or in other places of work, preparation hours of teachers and artists, waiting hours (e.g. a driver or porter waiting for work), as well as work hours of a non-paid family member in the family business or farm, if the person worked 15 hours or more, on the average, per week. The total number of work hours is the sum of work hours of all employed persons, or employees. The average work hours per week are obtained by dividing the total number of weekly work hours of all employed persons by the number of employed persons, including those persons temporarily absent from their work.

9. Type of locality - Until the end of 1989 localities were classified by type according to their estimated population in the 1983 Census of Population and Housing. In order to reflect the changes, which occur in population size in various localities, "the type of locality" is updated between censuses according to the size of the population.

Beginning with the data of January-March 1990, the localities were classified by "type of locality" according to the estimate of their population at the end of 1988. The localities are divided into two main groups, the distinction between them being according to size of locality (number of residents):

a. Urban localities, in which there are 2,000 or more residents. These localities are classified into sub-groups by size;

b. Rural localities, in which there are less than 2,000 residents.

For further details, see Technical Publication No.59 -"List of Localities, Their Populations and Codes, 31/12/1990".

10. District and Sub-District - The Sub-Districts are defined according to the "List of Localities. Their Population and Codes, 31 XII 1990. No. 59

Work District and Sub-District - Employed persons are classified according to the name of the locality of work. Employed persons working in Judea, Samaria and the Gaza Area are classified as a separate group.

11. The Tel Aviv and Haifa conurbations

Conurbations are large urban concentrations, which include more than one locality, usually having a continuous build-up area, and common economic, social and cultural characteristics. Since the conurbations are continuous areas, their population includes not only persons living in localities, but also those living outside localities within this area.

The conurbations are divided according to their internal structure:

Core: The area of the urban locality, which serves as the focus of the activity of the population of the conurbation.

Inner Ring: The area containing the localities surrounding the Core.

Middle Ring: The area containing the localities surrounding the Inner Ring (only in the Tel Aviv conurbation).

Outer Ring: The area containing the localities surrounding the Middle Ring.

Sections (Northern, Eastern, Southern): Additional Sub-divisions of the rings of the Tel Aviv conurbation.

Two conurbations were defined: Tel Aviv and Haifa. (Below: the internal structure of the two conurbations with the list of localities).

a. The Tel Aviv Conurbation

Core: Tel Aviv-Yafo
Inner Ring:
     Eastern Section:  Bene Beraq, Giv'atayim, Ramat Gan
     Southern Section: Azor, Bat Yam, Holon, Miqwe Yisra'el
Middle Ring:
     Northern Section: Gelil Yam, Herleliyya, Kefar Shemaryahu, Ramat HaSharon, Ra'anana
     Eastern Section:  Or Yehuda, Ef'al-Bet Avot, Ef'al-Merkal Seminariyoni, Giv'at Shemu'el,
                                  Ganne Yehuda, Ganne Tiqwa, Yehud, Kefar Alar, Newe Efrayim,
                                  Ben Gurion Airport, Savyon, Petah Tiqwa, Qiryat Ono, Ramat Ef'al, Ramat Pinkas
     Southern Section: Bet Dagan, Nahalat Yehuda, Rishon LeZiyyon
Outer Ring:
     Northern Section: Giv'at Hen, Hod HaSharon, Kefar Malal, Kefar Sava, Ramot HaShavim
     Eastern Section:  Lod, Rosh Ha'ayin, Ramle
     Southern Section: Be'er Ya'aqov, Malkeret Batya, Nes Ziyyona, Qiryat Eqron, Rehovot

b. The Haifa Conurbation

Core: Haifa and Oil Refineries' area
Inner Ring:
     Tirat HaKarmel, Kefar Bialiq, Nesher, Qiryat Atta, Qiryat Bialiq, Qiryat Yam, Qiryat Motzkin
Outer Ring:
     Oranim, Kefar HaNo'ar HaDati, Kfar Hasidim Alef, Kefar Hasidim Bet, Qiryat Tiv'on, Rekhasim.

12. Development Localities

Development localities are as defined by the "Development Towns Advisory Center", under the joint auspices of the Ministry of Labour and Social Welfare, The Ministry of Industry and Trade and the Ministry of Construction and Housing. In recent years, the Center has increased the number of localities in this category and besides the localities detailed below, many other localities are included, such as: Moshavot in the Galilee, Mitzpim in the Galilee and many localities in Judea and Samaria.

The following 25 localities comprise the sub-group "development localities" in the processing of the Labour Force Surveys.

Development localities in the North:

Bet She'an, Hazor Hagelilit, Tiberias, Yoqne'am Illit, Karmiel, Migdal Hafmeq, Ma'alot-Tarshiha, Nazerat Illit, Akko, Afula, Zefat, Qiryat Shemona, Shelomi.

Development localities in the South:

Elat, Ofaqim, Bet Shemesh, Dimona, Yavne, Yeroham, Mizpe Ramon, Netivot, Arad, Qiryat Gat, Qiryat Mal'akhi, Sederot.

B. Classification of Employed Persons by Status at Work. Economic Branch and Occupation

Data on employed persons relate to their work during the determinant week. For persons temporarily absent from their work, the data refer to the work from which they were absent. For unemployed persons who had worked in Israel during the 12 preceding months, the data relate to their last work. For those not in the civilian labour force, including soldiers in compulsory military service, who had worked in Israel during the last 12 months, the data refer to their last work.

1. Status at Work

Employee - a person working for another person in exchange for daily or monthly wages, piece work or work for any other kind of compensation.

Employer - a person who employs other persons for wages or any other compensation, or a partner in a business employing others, including a farm owner employing others for wages.

Self-employed - a person working in his own business or farm, who does not employ others for wages or any other compensation.

Member of cooperative - a person who, in addition to his salary, shares in the profits of a cooperative. This is also the case for a member of a collective moshav.

Member of gibbuz - includes members, candidates for membership, relatives residing permanently in the qibbuz and training groups. It does not include Nachal (Pionner Fighting Youth), classified as not belonging to the civilian labour force, or volunteers.

Non-paid family member - any person working in the family business 15 hours or more during the determinant week without any kind of compensation. A family member receiving a salary is considered an employee.

2. Economic branches - the classification according to economic branch is determined by the type of establishment, in which the interviewed person worked. Each person is classified according to the economic branch to which the establishment (or institution) in which he works, belongs. The branch is determined by the main product or service of the establishment. In an establishment whose activities cover various branches, (e.g. a complex textile plant, or an institution, such as a municipality), the branch is determined by the activities of the department, or division in which the interviewed person works: a subsidiary unit serving the needs of the establishment only, is not considered a separate department. If the establishment is engaged in several economic activities, which cannot be separated according to departments, the classification of the establishment is determined by its final product.

3. Occupation - Classification is determined by the type of work performed by the interviewed person at his place of work, without regard to the subject he studied, if his work is not in that field.
The classification of economic branches is based on the international classification of the United Nations (International Standard Industrial Classification) (I.S.I.C.) and the classification of occupations is based on the international classifications of the International Labour Office (I.L.O.), International Standard Classification of Occupations (I.S.C.O.).
Both classifications were adapted to the requirements of the Israeli economy.

CHAPTER E. - EFFECT OF UP-DATING THE “TYPE OF LOCALITY” ON SURVEY DATA (JANUARY-MARCH 1990)

Until the end of 1988, the "type of locality" of the localities in the "permanent file" was based on the results of the 1983 Census. Following the numerous changes, which have occurred since, the "permanent file" was updated according to the population estimates for the end of 1988, and the Labour Force Surveys started to use the up-dated file in January-March 1990.

If the "type of locality" is unchanged over a number of years, it makes possible a better comparison of the characteristics of the labour force over the years. Nevertheless, the type of locality should be updated in the "permanent file", in order to reflect the changes in size of population in the localities.

The data of the Labour Force Survey - January-March 1990 were processed according to type of locality in the 1983 Census and also on type of locality at the end of 1988, in order to estimate the effect of these changes on the results of the survey.

The effect of the change on the percentage of those belonging to the civilian labour force, on the percentage of unemployed and on the distribution of employed persons by economic branch at the national level is tiny and insignificant. But in the comparisons among the groups of "size of locality" significant differences are to be found in the percentage of those belonging to the labour force, the percentage of unemployed and the distribution of employed persons by economic branch.

For example:

(1) The percentage of males belonging to the civilian labour force in localities with a population of 50,000 to 99,999 residents, according to type of locality at the time of the 1983 Census was 54.3% and according to the type of locality at the end of 1988 it was 60.5%; the reason appearing to be the transfer of Nazareth to this group. Since the participation of non-Jewish males in the labour force is higher than that of Jewish males, there is an increase in the percentage of those belonging to this group.

(2) The percentage of unemployed females in localities with 10,000 to 19,999 residents according to the 1983 Census classification is 14.3% and according to the classification applicable to the end of 1988, it is 11.8%.

The inevitable conclusion is that in regard to type of locality, comparison cannot be made between data of the two classifications, without taking into account the change in method, although at a national level the change has no influence.