1991 Occupational Segregation variables supplied by Catherine Hakim, LSE

There are three versions of the JobSeg variable, all created on the same basis and from the same 1991 Census source, differing only in the sex ratio used to define the mid-point in the classification, and in the width of the middle Integrated occupations band.

The source for all three versions of the JobSeg variable is the 1991 Census 10% published census statistics on the occupational structure: Table 4 of the 1991 Economic Activity report. The statistics for all in employment were used to compile sex ratios for all occupations, and occupations were then grouped into three bands: male-dominated, integrated (or mixed) and female-dominated. The three versions are:

JobSeg1

• Male occupations <25% female
• Mixed occupations 25%-55% female (40% — 15%)
• Female occupations >55% female

JobSeg2

• Male occupations <30% female
• Mixed occupations 30%-70% female (50% ? 20%)
• Female occupations >70% female

JobSeg3

• Male occupations <15% female
• Mixed occupations 15%-45% (30% ? 15%)
• Female occupations >45% female

JobSeg1 In 1991, the total workforce was 44% female, so 40% was used as the mid-point for JobSeg1. This is the variable most people would choose to use for analyses of the whole workforce, either nationally or for particular regions, including comparisons of the full-time and part-time workforces.

JobSeg2 Some people prefer to use a single unvarying midpoint of 50% for all analyses of occupational segregation, on the grounds that the midpoint then does not vary from one year to another, or from one census to the next. The argument against this choice is that women have never constituted 50% of the workforce, so the distribution of the workforce will invariably be unbalanced around this artificial midpoint.

JobSeg3 In 1991, the full-time workforce was 32% female, so 30% was used as the midpoint for JobSeg3. This is the appropriate variable for analyses focused on the full-time workforce only, or for analyses of long-term trends covering the nineteenth century as well as the twentieth century.

Key references

Hakim C (1992) `Explaining trends in occupational segregation: the measurement, causes and consequences of the sexual division of labour', European Sociological Review, vol 8: pp 127-152.

Hakim C (1993) `Segregated and integrated occupations: a new framework for analysing social change', European Sociological Review, vol 9: pp 289-314.

Hakim C (1994) `A century of change in occupational segregation 1891-1991', Journal of Historical Sociology, vol 7: pp 435-454.

Hakim C (1996) Key Issues in Women's Work, London: Athlone.

Specification for creation of JobSeg variables in 2% SAR individual file

Missing values OCCPATN ()

RECODE OCCPATN
      (-9=8)
       (31,49,32,50,48,22,54,29,56,72,24,26,27,28,33=7)
       (39,46,17,13=6)
       (63,51,62,04=5)
       (57,09,08,01,21,58,47,25,03,12,40,23,-8=4)
       (10,52,14=3)
       (67,18,53,15,42,55,20,30,61,59,71,07=2)
       (05,45,43,66,68,02,16,73,44,19,11,60,06,64,37,65,38,
      70,36,35,41,34,69=1) INTO JOBCODE

RECODE JOBCODE
      (8=4)
      (7=3)
      (6,5,4=2)
      (3,2,1=1) INTO JobSeg2

RECODE JOBCODE
      (8=4)
      (7,6=3)
      (5,4,3=2)
      (2,1=1) INTO JobSeg1

RECODE JOBCODE
      (8=4)
      (7,6,5=3)
      (4,3,2=2)
      (ELSE=1) INTO JobSeg3

VALUE LABELS JobSeg1 JobSeg2 JobSeg3
      1 'male' 2 'mixed' 3 'female'

MISSING VALUES JobSeg1 JobSeg2 JobSeg3 (4)

Missing values Occpatn (-9)


 Specification for creating JobSeg variables in the 1% SAR household file



Missing values Occpatn ()

Recode OCCPATN
      (-9=8)
      (094,134,135,233,238,136,139,235,240,137,138,237,242,236,
      234,230,187,254,096,102,239,093,130,241,350,143,140,095,
      100,261,355,063,114,099,131,253,260,125,109,349,262,231,
      127,029,015,270,107,079,226,064,110,190,126,244,142,075=7)
      (065,228,227,129,117,124,296,016,097,295,298,291,351,192,
      038,077,207,132,225,224,195,034,327,018,263,293,062=6)
      (256,185,222,030,060,019,037,141,265,248,002,189,012,076,
      103,128,031,098,255,211,055,191,123,011=5)
      (297,080,111,041,118,247,266,197,032,299,205,078,115,269,
      356,113,061,112,039,271,258,036,056,106,287,252,026,
      033,249,054,184,043,-8=4)
      (003,058,186,009,067,292,203,324,246,070,221,057,215=3)
      (027,105,161,059,193,017,243,066,277,101,305,196,346,325,
      069,021,073,268,068,089,008,294,085,206,202,232,108,334,
      014,267,071,116,251,250,316,163,276,040,245,010,090,208,
      286,264,083=2)
      (273,122,133,289,188,121,001,119,217,025,120,345,013,048,
      272,194,078,042,332,020,218,358,004,209,028,314,072,084,
      274,035,086,104,007,024,229,023,322,171,223,052,160,279=1)
      (ELSE=1) INTO JOBCODE

RECODE JOBCODE
      (8=4)
      (7=3)
      (6,5,4=2)
      (3,2,1=1) INTO JobSeg2

RECODE JOBCODE
      (8=4)
      (7,6=3)
      (5,4,3=2)
      (2,1=1) INTO JobSeg1

RECODE JOBCODE
      (8=4)
      (7,6,5=3)
      (4,3,2=2)
      (ELSE=1) INTO JobSeg3

VALUE LABELS JobSeg1 JobSeg2 JobSeg3
      1 'male' 2 'mixed' 3 'female'

MISSING VALUES JobSeg1 JobSeg2 JobSeg3 (4)

MISSING VALUES OCCPATN (-9)