2 .

C o n cep ts, D a ta , a n d S trategies o f E n q u iry

29

T H E C H O IC E O F C O N C E P T U A L C O N T E X T

A s will h ave becom e a p p a re n . enough from th e

^
of social m obility has been p r o m p te d by a range of
,
conflicting, in terests, socio-political as well as academ ic (see, fu rth e r, G oldthorpe 1980^/1987: ch. 1). It is thus scarcely surprising to find a sim ilar y
S e
r a n g e o f variation in the conceptual a p p ro a ch e s th a t have b e en
followed by those actively engaged in investigating and analysing 5m obi 11 y.
T he m ost basic divergence w hich, as reco u n ted elsew h ere (G o ld th o p ,
1985a) can be traced back to n in e te en th -ce n tu ry origins re la te s to th
conceptual context w ithin w hich m obility is to b e d e f i n e d and in tu rn
observed and m easured. Tw o m ain tra ditions can h e re b e d i s t i n g u i s h e d ^
in which m obility is envisaged as occurring w ith in 'a class structure, and t e
other in which it is envisaged as occurring w ithin som e fo rm of social
hierarchy. F o r those w ho follow th e first tra d itio n , m obility re fe rs to th e
m ovem ent of individuals as betw een social p osition s th a t a re iden tified in
term s of relationships w ithin lab o u r m arkets and p ro d u c tio n units; fo r th o se
who follow the second, m obility refers to th e m o v e m e n t o f individuals as
betw een social groupings o r aggregates th a t a re ra n k e d acco rd in g to such
criteria as their m em b ers p restige, status, e conom ic re so u rc es, etc.
T o be su re, particular m obility studies m ay n o t alw ays n e atly fall in to o n e
ra th e r than th e o th e r of th ese tw o traditions. Som e investig ato rs a p p e a r to
shift betw een a class-structural and a h ierarchical c o n te x t m o re o r less
unknow ingly, w hile o thers have sought delib erately to w o rk w ithin b o th .
H ow ever, w hat m ay be said is th a t in any study o n e c o n ce p tu a l a p p ro a c h or

Concepts, D ata, and Strategies

of Enquiry
In this c h ap ter we consider a variety o f m ethodological issues which range
from ones o f a quasi-philosophical ch aracter through to o th ers of statistical
ec n ique. T o som e re a d e rs, o u r discussion of these issues may seem to
constitute an excessively lengthy pream ble to the treatm ent of the substantive
m a tte rs th a t a re o f u ltim ate im p o rtan ce. If so, we m ust rem ain u n repentant.
W e do n o t believe that the way in which in a sociological study methodological
p ro em s a re defined and resolved can be seen as simply a backstage
o p e ra tio n , to be re p o rte d o n , if at all, only in th e decent obscurity of an
app en d ix . T h e tre a tm e n t o f m ethodological and o f substantive problem s is
in fact often difficult to separate, and this should not be concealed. M oreover,
as w e h a v e arg u ed a t the e n d o f the previous ch ap ter, the unsatisfactory state
o f c o m p arativ e m obility research , as regards the persistence of large areas of
d isag ree m e n t on essentially em pirical questions, has to be a ttrib u ted largely
to m eth o d o lo g ical difficulties, and th e re is then a particular need for these to
b e explicitly identified and addressed.
W e b eg in th e c h ap ter w ith a discussion o f th e different conceptual
c o n te x ts w ithin which the study o f social m obility may be u n d ertaken, and
w e explain o u r decision to ad o p t a class-structural perspective. W e then give
an acco u n t o f th e class sch em a th ro u g h which w e have attem p ted to m ake
this decision o p e ra tio n a l, and which provides the basis for the num erous
m obility tab les th a t we shall su b seq u en tly construct. From such problem s'of
concep ts and th e ir re p re se n ta tio n , we m ove on to those of data and, most
im p o rtan tly , to th o se associated with achieving data o f an ad eq u ate standard
o f cro ss-n atio n al com parability'. W e describe the approach th a t we have
ta k e n to th ese pro b lem s and th e pro ced u res th at we have followed in
constitu tin g th e d a ta -se t o n which o u r em pirical analyses rest. O u r next
concern is w ith analytical tech n iq u es. W e seek to show w hy, following the
concep tu al a p p ro a ch th a t w e have a d o p te d , we m ay m ost appropriately
w ork th ro u g h th e analysis o f m obility tables, using in addition to sim ple
p ercen tag in g tech n iq u es o f log-linear m odelling, and how in this way we
can im p lem en t w h at is fo r us a crucial distinction betw een absolute and
relative m obility rates. Finally, we tak e u p ilarge, yet highly consequential,
is su e s o f th e m e th o d o lo g y o f co m p arativ e studies. W e note the sharp
divergencies o f o p in io n an d practice th a t hfere exist, and explain the com
bin ed o r m ix e d strateg y th a t we will ourselves a tte m p t to follow.

r*

'

1 A s regards the origins o f th ese different con ceptual tradition s, a com parison o f p assages
from M arx and John Stuart M ill is o f interest. M arx, in o n e o f th e rather few instances in w hich
he discusses social m obility directly, alludes to the ease w ith w hich in th e U n ited S tates o f the
m id-nineteenth century w age-labou rers w ere able to turn th em selv es in to in d ep en d en t selfcurtaining p easants so that th e position o f wage-labourer w as for the: m ajority o f th o se w ho
occupied it but a probationary state w hich they are su re to lea v e w ithin a lon ger or a shorter
term (1865/1958:444). A n d w hat is then for M arx o f ch ief significance a b out this m ob ility is not
that it represents any form o f social ascent he m akes no attem p t to characterize it in v ertical
term s but rather that it serves to prevent what he elsew h e re refers to as a d ev elo p ed
form ation of classes. B eca u se o f th e prevailing high rate o f m ob ility, M arx argu ed , classes in
A m erican society have, not yet b ecom e fixed but con tin ually ch an ge and in terchan ge their
elem en ts in con stan t flux "(1852/1958: 2 5 5 ).
'

M ill, on the oth er hand, has a q uite d ifferent p ersp ective o n m ob ility. A n ticip atin g liberal
theorists, he em phasized th e way in w hich advancing industrialism in m id -n in eteen th -cen tu ry
E ngland w as beginn ing to-brfeak dow n the barriers that had traditionally ex isted to m ovem en t
b etw een differen t lev e ls or grades o f em ploym ent. H ith erto, th e d em arcation b etw een th ese
grades had b een alm ost eq u ivalent to an hereditary d istinction o f ca stes, each em p loym en t
b eing chiefly recruited from th e children o f th ose already em p lo y ed in it, o r in em p loym en ts o f
the sam e rank with it in social estim ation . . .. B ut currently, according to M ill, th e habits and
disabilities w hich chained p eo p le to their hereditary con d ition s are fast w earing aw ay, and
every class is exposed to increased and increasing com petition from at least the class im m ediately
b elow it (1 8 4 8 : 4 6 2 -3 ). T h u s, although M ill h ere u ses the langu age o f class, h e clearly s e e s th e
con text o f m obility as b ein g in fact a social hierarchy: that is, o n e d eterm in ed by th e typ e o f
work perform ed and its 'social estim ation .

2 . Concepts, Data, and Strategies o f Enquiry

2 . Concepts, Data, and Strategies o f Enquiry

the other tends to be dominant; and this is, in our view, to be expected since
he two approaches are directed towards the treatm ent of different even
though overlapping, sets of problems.

30

In consequence, where mobility is analysed in a hierarchical context as

represented by occupational prestige or status scales, it becomes difficult for
the structural influences that bear on mobility rates and patterns to be
adequately isolated and displayed. Occupational groupings that are treated
as equivalent will in fact often be ones that are affected in quite different
ways by, for example, shifts in demand, technological innovation, or the
policies of national governments, and that may thus be following, within the
overall course of economic development, quite divergent trajectories of
expansion or decline (cf. Westergaard and Resler 1975: 287-8).
If, in contrast, mobility is studied in the conceptual context of a class
structure, it is not vertical movement on some social scale that will be at the
centre of attention but, rather, mobility understood in terms of relational
changes: specifically, changes in the nature of individuals involvement in
relations within labour markets and production units. Such changes may
not, in fact, be readily interpretable as implying social ascent or descent.
For, while the relationships that constitute a class structure can be seen as
expressing differential social advantage and power, they do so in quite
varying ways; and class positions need not therefore be ordered in any
consistent unidimensional fashion (cf. Carlsson 1958: ch. 3; D ahrendorf
1959: 74-7; Giddens 1973:106).3 Further, though, within a class-structural
approach individual chances of mobility, whether vertical or otherwise, are
unlikely to be the only issue raised. In contrast with the social groupings
found at similar levels of prestige or status, classes in the minimal sense of
collectivities defined by their common class locations can be expected to
show some degree of homogeneity not only in the kinds and levels of
resources that their members command but further in their exposure to
structural changes and, in turn, in the range of at least potential interests that
they may seek to uphold. Thus, within a class perspective, it becomes
possible for the investigation of mobility chances to be included within a
larger concern. While the significance of such chances in themselves and in
'actually revealing the inequalities in advantage and power associated with
different class locations can be fully recognized, other issues can also

For our own part, we have chosen here to take up the class-structural

Lnd7 r;TeVen?rd
SthatltW
0UldaPoverall
Peart0base^regards the
in any event, the more
promising

range of

alTenllm

We haV' iC1,e<1' we inte" d > concentrate our

wh ' H n L i l l d b y l

S'

S k P ' S d ,0 i

- i! mh! ta, lltyiS St di^ d Within the COncePtual context of a social hierarchy, it
may be supposed that a prime focus of interest is on mobility in some
vertical sense that is, on social ascent or descent in terms of prestige
status, or whatever is taken as the ordering principle of the hierarchy; for it is
such upward or downward movement that will, of course, be implied in each
individual instance of mobility. Such an approach has an obvious relevance
for the examination of a num ber of m ajor issues within the field of mobility
research. It is, for example, well suited to treating questions concerning the
determ inants of individual success or failure or, in other words, ques
tions of who gets ahead and why? Thus, scales of occupational prestige or
of socio-economic status have formed the basis for most studies carried out
within what has become known as the status-attainment paradigm (see
e.g., Blau and D uncan 1967; Jencks 1972, 1979). And in turn, then, a
hierarchical (perspective can also be taken up in macrosociological and
com parative 'analyses where the aim is to characterize societies as being
m ore or less o p en according to the relative importance of ascription and
achievem ent as represented, say, by family and education in promot
ing success or guarding against failure.
H owever, what must at the same time be noted is that occupations that are
found in close proximity to each other on scales of prestige or status need
not, and often do not, have much else in common with each other, and may
indeed hold-quite disparate locations within the social division of labour.
Thus, within the sam e narrow .band of-sale values one may find, say,
groupings of skilled industrial workers brought together with certain types
of small proprietor and minor official; or, again, farmers and smallholders
placed alongside artisans, industrial labourers, and personal service workers.2

F or instan ce, in th e case o f th e D uncan Index o f O ccupational Status (D uncan 1961; BJau
and Duncan 1967:122^-3), o n e finds within the score interval 5 5 -9 electrotypers and stereotyp es
and lo co m o tiv e en gineers grouped togeth er with funeral directors and w h olesale proprietors,
railroad con d uctors, and certain types o f forem en; or again, in the interval 1 0 -1 4 , farmers and
fishermen bracketed with shoem akers and repairers (nonCfactory), longshorem en and stevedores,
Jab .ou rers in m achinery m anufacturing, hospital attendants, and taxi-drivers. Inspection o f the
m ore d etailed H o p e -G o ld th o r p e scale o f the prestige pr general d esirability of occupations
(G old th orp e and H o p e 1974: 9 6 -1 0 9 ) w ould p rovide still more num erous exam ples o f such
h eterogen eity. It should be em ph asized that the aim h ere is not to q uestion the validity o f these
scales. T h e point is, rather, that, w h ere occupations are brought together according to their

BSf

X ..

31

s statu l, prestige, e tc ., there is no reason why they should also be h om ogen eou s in terms o f their
S;"j structural locations and that, in all probability, they will n ot be.
3
Our position here d o es, h ow ever, differ significantly from that o f Runcim an. Although
Runcim an initially defines classes as sets o f roles w h ose com m on location in social space is a
function o f th e nature and d egree of econom ic power (or lack o f it) attaching to them through
their relation to th e institutidnal p rocesses o f production, distribution and exchange (1990:
. 377; em phasis ad ded), h e then w ishes to treat as equivalent, on a single d im ension, roles w hose
shared level o f econom ic pow er derives from different institutional relations . . . (1990: 380).
W hether or n ot it is the case that the idea o f such equ ivalence is in principle familiar enough to
sociologists o f all theoretical sch ools, w e find it difficult to understand why Runcim an should go
on to concern h im self with th e q uestion o f H ow many classes are there? For it w ould seem far
m ore con sisten t with his general position for him to conceptualize the differing d egree of
econ om ic p ow er attaching to class roles its differing nature being effectively discounted as
form ing a continuum . /
%

32

2 . Concepts, Data, and Strategies of Enquiry

Should no Jediffi h ClaSS' StrUCtUralas aSainst

hierarchical approach
would ?nr th
t
0 aPPrec,ate- To put the m atter more concretely we
would for the task m hand, see a decided advantage in being able to t o
rin d S tn a T w a f
160118 f ^
cate80ries- s ch as, say, those
P
1 *7 * PCuSantS rfarm erS=S3laried emPloyees, pro
prietors and self-employed workers, e tc .- r a th e r than in terms of categories
which represent simply levels distinguished within a prestige or status
of th e f11 ' h ll 'S WC would m a^ntain>far more as members of categories
the form er than of the latter kind that individuals have in fact experienced
have'd e v e l o p '
J
economic- through which their societies
m rtl
7 h
u ,ndustnal on4es and have in turn responded to, and
participated in, these processes.
To argue thus, we would stress, does not require us to claim that classes
are m ore real than collectivities distinguished by reference to prestige or
s a us nor to regard a class-structural approach as being superior to a
hierarchical Qne in any absolute sense. As we have sought to make clear our
preference for the former approach is entirely a matter of choosing one
conceptualization over another because we believe that, on balance it is the
m ore suited to our purposes. Accordingly, we would not seek to deny that
along with the advantages offered, certain disadvantages are also entailed
Most obviously, within a class-structural perspective the possibility is lost of
treating mobility systematically in terms of social ascent or descent, and
further, then, the possibility of applying various elegant analytical techniques
which depend upon the adoption of-categories ordered within a social
hierarchy.
i
;
A t the same time, though, we would reject the implication of arguments
such as those of Kelley (1990) to the effect that a class-structural context for
the study of mobility is in some way less naturalor conceptually more
arbitrary than a hierarchical one in being less securely grounded in the
ways in which mobility.is typically thought of and lived through by the Iky
m em bers of a society. While we would not, of course, question that popular
conceptions of mobility do often imply a hierarchical context, the idea jof
4
W e would regard it as significant that it is in fact in, terms o f classes, rather than o f prestige
or status levels, that historians o f (he m odern period have usually wished to discuss social
cleavages, conflicts, and m ovements; and working within essentially the sam e conceptual
fram ew ork as they have done is itself an advantage that w e will seek to exploit.

2 . Concepts, Data, and Strategies o f Enquiry

33

mobility between class positions which cannot be readily characterized in

vertical terms is also, we believe, one that finds a wide correspondence at the
level of actual social experience.
For example, in the course of industrialization the most important of all
mobility flows, numerically at least, are those of peasants and small farmers
and of their offspring into semi- and unskilled wage-earning jobs in industry;
and there can be little doubt that radical changes thus ensue not only in the
market and work relationships in which these individuals are involved but
further in their entire way of life and its social setting. Yet perhaps one of the
least revealing questions to be asked about such mobility, whether from the
standpoint of the sociologist or from that of the social actors in question, is
whether it is upward or downward in direction. Typically, such mobility
would appear to bring, and to be experienced as bringing, varying kinds of
both gains and losses for which no common denominator is readily available,
for example, increased income and shorter hours of work but devaluation of
previously acquired skills and reduced autonomy; the chance of entering
modern affluent society but also the risk of cultural and social impoverish
ment (cf. Smelser and Lipset 1966; Johnson 1981).3 M oreover, other
instances of mobility in which similar trade-offs seem often present and
likewise ambivalence'on the part of those contemplating, or who have made,
the transition are not difficult to cite: for example, that from employee to
independent, self-employed status or from a rank-and-file to a first-line
supervisory position (see, e.g., Chinoy 1955: chs. 5 and 7; Goldthorpe etal.
1968: ch. 6; Mayer 1977; Bland, Elliott, and Bechhofer 1978; Crossick and
Haupt 1984). And here again what would appear from the monographic

5 Johnson has written as follow s on the im plications o f mobility from the peasantry into
industrial work in Eastern E uropean societies in the post-war period.
It should be stressed that the m ove from the village to the town or city represented an
: enorm ous change in nearly every aspect o f the lives o f th ose involved. It was not sim ply a case
o f changing on e 'job for another. An enorm ous gulf separated the relatively m odern
lifestyle o f the industrial centers from the age-old patterns and traditions o f the relatively
isolated peasant com m unities. The urban migrant left behind the em otional support and the
econom ic security (how ever tenuous) provided by m ultibonded ties o f fam ilial relationship
and lifelong acquaintance for the anonym ity and impersonality o f life among strangers.
Instead o f the relatively self-sufficient patterns o f consum ption characteristic o f subsistence
agriculture and handicraft production, he now found it necessary to depend . . . upon his
. capacity to acquire the cash to purchase [com m odities] by selling his labor. Instead o f the
autonom ous, self-scheduled and periodically varied patterns o f work characteristic, o f smallscale agriculture, he now had to accustom him self to the unfamiliar discipline o f the foreman,
the punch-clock and a minutely specialized and routinized task on the assem bly line. With his
skills as a farmer and husbandryman essentially useless in his new setting, and lacking in
formal education, he almost necessarily had to settle for the low w ages and often unpleasant
working conditions o f an unskilled or sem iskilled laborer; yet at the same tim e the exam ple of
his acquaintances and the proximity and visibility o f amenities and 'luxury' commodities
almost unknown in the village stimulated him towards acquisitiveness and a consumptionoriented mentality that linked the sen se o f self-worth with the outward trappings o f material
prosperity. (1981: 33.-4)

WMOTECA SAN JOAOGf*

" O s '*

SfSTEMA DE BIBLIOTECA*
PONTIFIC1A u a

o e C H IU

34

35

2- Concepts, Data, and Strategies o f Enquiry

2. Concepts, Data, and Strategies o f Enquiry

literature to be of chief significance is the relational changes that are involved

rather than movement along some vertical, attributional dimension
Since our mam aim thus far has been to bring out as clearly as possible the
Jfferences between the class-structural and hierarchical approaches to the
s udy of m o h h ty and to show how these have been developed in regard to
!etS of pr blem s>we have in effect represented the two approaches
as alternative paradigms between which a choice must, implicitly or explicitly
be made In conclusion, however, we should qualify this account by recog-

between the social inequalities that are conceptualized in terms euhe of a

class-structure or of a prestige or status hierarchy, it is, we beheve sn l
case that no perfect mapping from one conceptualization to the othe can b e
achieved and t h a t - t o return to the point from which we began there is
good reason why, in any particular study, one or the other shoul
privileged.7
TH E CLASS SC H EM A

a J T ?

that the tW aPProaches are not entirely incompatible

sought
m PraCtlC6 S me d6gree f COmPromise between them may be
Thus, in the case of investigators taking up a hierarchical perspective on
n
r StatUS SCaleS k is notable that>whlle>for example,
Glass (1954) and Svalastoga (1959) made little or no attempt to reduce the
heterogeneity of occupations that were placed at the same hierarchical level
ater exponents of this approach have often been concerned to do so At
least for the purposes of mobility-table analysis, the practice has become
quite regular of starting with a set of occupational categoriesderived say
rom official sta tis tic s-th a t are relatively homogeneous in their structurai
locations and of then ordering these categories in the light of some average of
i
Vn 'UeS f th d r constituent occupations (see, e.g ., Blau and Duncan
1967: 26-9; Featherm an and H auser 1978: 25-30). Indeed, it may also be
the case that the ranking thus produced is itself subject to further ad hoc
modification ,>with the aim, it would seem, of bringing it into yet closer
accord with what might be described as class-like criteria.6This latter step
is, however, one which, in our view, risks achieving compromise at too great
a cost m terms of conceptual consistency: it becomes unclear just what the
context of mobility is taken to be.
In turn, where a class-structural perspective is adopted, the most obvious
way of attempting to capture some of the hierarchical aspects of mobility is
then to order the classes distinguished by reference to prestige, status, or
some other appropriate criterion of an external kind. This is, in fact, the
strategy that we shall ourselves take up, is described in the sectionthat
follows. None the less, we shall pursue it to only a rather limited extent
While at an empirical level a fairly high correlation may be expected
a c c l d i 1 ^ 0/ eXample B !au and D u ncan, having ranked seventeen occupational categories
mod f v l i . h
maintain fhp

KCOn miC ' f ? 5 n the basis o f median incom e and Vea o f schooling? then
by m o''1P g',Salf sm e n ' retai1 abov three categories o f craftsm en so as to

m akin, . h t aH T anU, ",man,ual dlstlnction ( 1967: 2<?)- (T hey appear, how ever, to forgkt
scalinB of mnh i ^ a"
later seekln8 t0 ar8ue ( T . 7 1 -5 ) that the multidim ensional
and H a u ^ S f
^
Centraty f so d o 'e ^ no c status.) Likewise, Featherman
S
onaTdfsl L , r SOC1 - t1om lc ki"g
sam e seventeen categories, invoking

p r o v id e d ^ ofber^ank criteria^(1978^
above t w o c V Z J ^ f
f
(

' ^
Salesmen>!retai1

W o rm a J
are again prom oted, this time

five strata into w hkh

mClUded in the sec n d - 'l o w e r nonm anual- o f
rata into wh1Ch the seventeen categories are for som e purposes aggregated

Having opted to conduct our analyses of mobility within a class-structural

context, we need then to translate this decision into practice by establishing
some set of class categories which will provide the basis for our empirical
work. However, there is no obvious and uncontroversial way in which this
may be done: to the contrary, the concept of class is a notoriously contested
one. Our position in the face of this difficulty is the following. W e take the
view that concepts like all other ideas should be judged by their con
;
sequences, not by their antecedents. Thus, we have little interest in argu
ments about class that are of a merely doctrinal nature. The class schem a
that we have developed in connection with the present and previous
(Erikson, Goldthorpe, and Portocarero 1979) com parative investigations
possesses a theoretical rationale which, we trust, endows it with a m easure of
internal consistency. But, as will become apparent, it is in its inspiration
rather eclectic. We have drawn on ideas, w hatever their source, th at
appeared to us helpful in forming class categories capable of displaying the
salient features of mobility among the populations of m odern industrial
societies and within the limits set by the data available to us. C orrespond
ingly, it is by reference to this objective that, we would hope, the value of the
r schema will be assessed.
:
Table 2.1 provides a description of the categories of the schema. The most
extended version, shown in the first column of the table, is presented,
together with the diagrammatic account of Figure 2.1, primarily to help
readers follow the rationale of the schema that we set out thematically
rather than class by class in the remainder of this section. It is actually the
seven-class version of the schema, found in the second column of Table 2.1,
that will for the most part serve as the basis of our empirical analyses, since,
as we report in the next section, some collapsing of categories was necessary
in order to maintain standards of cross-national comparability in our data.
The explanation belowjof features of the full version of the schema that we
cannot in facj utilize should enable readers to m ake their own judgem ent of
^ In p rinciple, a com p lete recon ciliation o f the two. approach es m ight b e th ought possib le
sim ply through th e elab oration o f a sufficiently large num ber 'of categories: for exam p le,
through d istinguishing p restige or status strata w ithin each o f a set o f classes. In practice,
h ow ever, very large sam ples w ould then be required in order to h ave reliab le cell cou n ts in
m obility tables based on such a categorization, and q uestion s o f the theoretical focus o f
particular an alyses w ould in any ev e n t rem ain.

Basic class positions

2. Concepts, Data, and Strategies o f Enquiry

37

how far in its seven-class form its underlying p rinciples are ^ P rmiS^ e
T he five- and three-class collap ses that are also sh ow n in T ab le 2 .1 are
clearly less satisfactory but w ere u n avoid ab le in a num b er o f m ore_c o m p e x
analyses w here cell counts w ou ld o th erw ise h ave b e c o m e u n reliab ly lo .
T h e aim o f the class schem a is to d ifferen tiate p o sitio n s w ithin la b o u r
m a rkets and p ro d u c tio n units or, m ore sp ecifica lly , o n e co u ld sa y , o
differentiate such position s in term s o f the e m p l o y m e n t re la tio n s that th ey
en ta il.9 T h e principles o f differentiation that w e a d o p t h a v e b e e n m ainly
derived from classic sou rces, in particular, from M arx and M ax W eb er. B ut
these principles h ave been ad ap ted , under th e in flu en ce o f v ariou s later
authors, to try to m eet the specific req uirem ents o f a n alysin g class m o b ility
within the total pop ulation s o f m id -tw en tieth -cen tu ry industrial n a tio n s,
both capitalist and state socialist.
For M arx and W eber alike, it could be said th at e m p lo y m e n t re la tio n s are
crucial to th e delineation o f the structure o f class p o sitio n s w ith in m o d ern
society even though th ese authors w ould accord a so m e w h a t d ifferin g
significance to these rela tio n s.10 From both so u rces, w e can d er iv e a b a sic,
threefold division o f class position s, as follow s:
1. em ployers: i.e . th o se w h o buy th e labour o f o th ers and th u s a ssu m e
som e degree o f authority and control o v er them ;
2. self-em p loyed w orkers w ithout em p loyees: th o se w h o n eith er buy th e
labour o f others nor sell their ow n;

8 The origins of the present schem a lie in one devised by G o ld th o rp e an d L lew ellyn (1977)
for use on British data. T his was then modified for purposes of th e co m parative m obility
analyses reported in Erikson, G o ldthorpe, and P ortocarero (1979) at th e sam e tim e as E rikson
was involved ifi revising a Swedish classification (A ndersson, E rik so n , and W arneryd 1981)
which already in its original form (Carlsson et a i 1974) had close sim ilarities w ith th at of
G oldthorpe and Llewellyn. F u rth er refinem ents w ere proposed in th e course of la te r B ritish
work by G oldthorpe and Payne (1986) in collaboration with A nthony H eath. T h e .ra th e r
cum bersom e labelling of classes by com binations of rom an num erals and letters is here
retained, since this enables earlier and later versions of the schem a to be system atically related.
9 In earlier presentations we have referred to the class schem a as aim ing to bring to g eth er
.individuals holding similar m arket and w ork situations. T he revised form ulation h ere used
seetcs to bring o ut m ore clearly that the schem a is intended ultim ately to apply to p o sitio n s, as
defined by social relationships, rath er than to persons ^ although, for purposes o f describing
the classes distinguished, it is difficult to avoid referring to actual incum bents e.g. m anagers,
p roprietors, w orkers, etc. W e might also add here th at, while w e then allocate individuals to
class positions within our schem a, we do not thereby com m it ourselves logically or sociolog
ically to the view th at th e'class of all individuals alike will be m ost validly d eterm ined by
reference to their ow n em ploym ent. Q uestions o f derived class, turning on w hether the
individual or the family is the m ore appropriate unit of class com position, will be tak en u n at
length in discussing th e class mobility of women in C hapter 7.
We would, in fact, believe that the opposition between M arxian and W eberian conceptions of
class that is by now enshrihed in sociology textbooks is in m any respects exaggerated and
especially in view o f the fact that the work of neither author can be regarded ^ p r o v id in g a
canonical statem ent o f his position. O ur own approach has been often refe rred to and
discusssed as W eberian!, b ut we would not regard this as particularly inform ative o r otherw ise
helpful: to rep eat, it is consequences, not antecedents, that m atter.

I
1

T a b l e 2 .1 .

The class schem a

Full version

Collapsed versions
Seven-class*

II

Higher-grade professionals,
administrators, and officials;
managers in large industrial
establishments; large
proprietors

R outine non-manual
em ployees, higher grade
(administration and com m erce)

III6

R outine non-manual
em ployees, low er grade (sales
and services)
.

IVo

Small proprietors, artisans^etc., with em ployees

IV>

Small proprietors, artisans,

etc. .w ithout em ployees

IV c

/
I+ II

Low er-grade'professionals,' /
. Administrators, and officials;
higher-grade technicians;
managers in small industrial
establishm ents; supervisors of
non-manual em p loyees
.

Ilia

primary production

III

IVc

VI

V +VI

Skilled manual workers

etc.)
V IIb Agricultural and other workers
in primary production

Noteci:

.
^

Non-manual
workers

Skilled workers: lower-grade

technicians; supervisors of
manual workers; skilled manual
workers
N on-skilled w orkers: sem i-an d
u n s k i l l e d m a n u a l workers (not
in agriculture, etc.)

V IIb

A gricultural labourers.
agricultural and other w orkers
in primary production

IV a + b

Petty bourgeoisie

IV c+V IIb Farmworkers

V +VI

Farm workers

Skilled workers
Manual
workers

V ila

ion o f the schem a is the on e we shall m ost frequently use in o u r analyses. W e give h ere th e nam e

rh,

i.

Farmers: farmers and

smallholders and other self
employed workers in primary
production

V ila
V ila S em i-and unskilled manual
workers (not in agriculture,

White-collar workers

Routine non-manual workers:

routine non-manual employees
in administration and
commerce; sales personnel;
other rank-and-file service
J
workers

IV a + b P etty bourgeoisie: small

proprietors and artisans, e tc .,
with and without employees

Lower-grade technicians,
supervisors of manual workers

Service class: professionals,

administrators and managers;
higher-grade technicians;
supervisors of non-manual
workers
I-III

F a r m e r s and s m a l l h o l d e r s ;
, other self-em ployed workers in

"

Three-class

Five-class

the seven-class version onnc

l

usually re fe r to each Cass.

service

40

2 . Concepts, Data, and Strategies o f Enquiry

2 - C o n c e p t , D a ta , a n d S trategies o f E n q u iry

3. employees: those who sell their labour to employers and thus place
themselves to some degree under their authority and control.
In the construction of the class schema, this threefold division may be
regarded as the starting-point, as Figure 2.1 illustrates. However, consider
able modification and elaboration then follow, in respect chiefly of two
closely related developments within the twentieth-century industrial world:
first, the transform ation of property into corporate forms, whether private
or public, which has resulted in most m ajor employers being organizations
rather than individuals; and, secondly, the growth of employees as a
proportion of the total active population, accompanied by a greater differ
entiation of the forms of em ployer-em ployee relations as employing organ
izations have becom e increasingly bureaucratized.
As regards, first, the now predom inantly corporate or, in socialist
societies, state-owned nature of productive property, the main implication is
that the class of em ployer becomes one made up overwhelmingly of sm all
employers with, say, work-forces that num ber at most in tens rather than in
hundreds. Such small proprietors are comprised by Class IVa of our schema,
while Class IV b covers self-employed workers without employees apart in
both cases from those engaged in primary production, who go together in
Class IVc.
As can, th en , be seen from Table 2.1, no separate class of large employers
is provided for. Large proprietors are in fact included in Class I, where they
might appear to constitute a rather anomalous element. However, more
detailed exam ination of the cases in question, which we have been able to
m ake m several of our national samples, suggests that, as well as being few in
num ber, they are also less straightforward than they might seem.
To begm with, it should be emphasized that we are n o t dealing here, other
than quite exceptionally, with members of a capitalist elite or leaders of
industry. R ather, large proprietors turn out to be most typically the owners
of stores, hotels, restaurants, garages, small factories, or building or trans
portation firms. Indeed, it might be argued that, since their operations tend
to be on only a slightly m ore ambitious <;cale than those of the small
proprietors of Class IV a, little is gained in seeking to separate them from the
latter. The main reason for so doing, and for including them in Class I, is
that, in so far as such large proprietors tend to be quite extensively involved
in managerial as well as entrepreneurial activities, they may be regarded as
aving a yet greater affinity with those salaried managers to be found in
Class I who have a substantial share in the ownership of the enterprises in
which they work. M ost often, the difference here will be simply one of
11 W e w ould e s tim a te th a t, ac ro ss th e W e s te rn in d u s tria l so c ie tie s th a t w e m n .i H .r %

41

whether enterprises of very much the sam e kind as was '" d'^ at^ ^
u
that is, relatively m odest o n e s - h a v e becom e incorporated o r not. a m atter
which will be of som ew hat varying consequence cross-national y, ep e
g
on prevailing com pany law and business practice, b u t in any event m ore
from a legal, financial, and fiscal than from a sociological stan d p o in t.
It m ight further be noted at this point th at considerations sim ilar to th e
foregoing underlie the inclusion in Classes I and II of th e schem a of all
professionals, w hether reportedly self-em ployed or em ployees. A m ong
professionals, a variety of legal and conventional arran g em en ts is to be
found through which independent practice and salaried em p lo y m en t are
com bined or the distinction betw een them is e f f e c t i v e l y blu rred ; an d , even
w here it might be possible for tru e independents to be identified, it w ould
then seem likely that in m ost of the national sam ples at o u r disposal th eir
num bers would be too small to allow sep arate analysis.
W hen we turn, secondly, to the growth in num bers of em ployees and th e
increased differentiation in their relations with em ployers (usually, em p lo y
ing organizations), w hat we are clearly forced to recognize is th e inadequacy
of treating all em ployees alike as holding sim ilar class p ositions. It is
apparent th at, in consequence of em p lo y er-em p lo y ee relatio n s being based
on quite heterogeneous principles, em ployees in fact occupy a ran g e of
different labour-m arket and w ork situations, am ong w hich m eaningful
distinctions can and should be m ade in class term s.
The line of division to which we would h ere give g reatest em p h asis
following W eber and in turn R enner (1953) and D a h re n d o rf (1 9 5 9 ,1 9 6 4 )
is that which stem s from differences betw een, on th e one h an d , th e lab o u r
contract and, on the o th er, the conditions of em ploym en t w hich typically
obtain within organizational bureaucracies, both public and private. E m ploy
m ent relationships regulated by a labour contract en tail a relatively sh o rt
term and specific exchange of m oney for effort. E m ployees supply m ore-orless discrete am ounts of labour, under the supervision of th e em ployer o r of
^ e m p lo y e r 's agents, in retu rn for w ages which are calculated on a p iece
or tim e basis. In contrast, em ploym ent relationships w ithin a b u reau cratic
context involve a longer-term and generally m ore diffuse exchange. Em ployees
render service to their em ploying organization in re tu rn for co m p en satio n ,
I t s h o u ld b e n o te d t h a t , w h e re in tru ly la rg e -s c a le e n t e r p r i s e s m a n a g e r s h a v e a s h a r e in
o w n e rs h ip , th is w ill t y p ic a lly a m o u n t t o o n ly a v e r y s m a ll fr a c tio n o f t h e t o ta l e q u ity a n d e v e n
if su ffic ie n t to g iv e th e s e in d iv id u a ls c o n s id e r a b le p e r s o n a l w e a l t h w ill s till n o t c o n s t i t u te th e
b asis o f th e ir a u t h o r i t y , w h ic h w ill r e m a in b u r e a u c r a tic . I n o t h e r w o r d s , th e w e a lth o f s u c h
m a n a g e rs d e r iv e s fr o m t h e ir p o s itio n s o f a u th o r ity a n d n o t v ic e v e r s a , a s w ith t h e e n t r e p r e n e u r
in th e h e r o ic p h a s e o f c a p ita lis t in d u s tr y .
T h u s , f o r e x a m p le , in E n g la n d , w h e r e s e lf- e m p lo y e d p r o f e s s io n a ls i n c lu d e , fo r la rg e ly
te c h n ic a l r e a s o n s , all g e n e ra l m e d ic a l p ra c titio n e r s w o r k in g f o r t h e N a tio n a l H e a lth S e rv ic e a n d
m a n y C h u rc h o f E n g la n d c le rg y m e n , th e c a te g o ry still a c c o u n ts f o r o n ly a little o v e r 1 p e r c e n t
o f all m e n w ith in th e w o r k -fo rc e a n d fo r a b o u t 5 p e r c e n t o f all a llo c a te d to C la s s e s I a n d II
t o g e th e r , w h ile in S w e d e n , w h e r e a s tr ic te r d e f in itio n o f s e lf- e m p lo y m e n t a p p lie s , t h e c o r r e s
p o n d in g p e r c e n ta g e s a r e ro u g h ly h a lv e d .

Scanned by CamScanner

42

2. Concepts, Data, and Strategies o f Enquiry

which takes the form not only of reward for work done, through a salary and
various perquisites, but also comprises important prospective elementsfor
example, salary increments on an established scale, assurances of security
both in employment and, through pensions rights, after retirement, and,
above all, well-defined career opportunities.
As argued at greater length elsewhere (Goldthorpe 1982), a service
relationship, rather than one formulated in terms of a labour contract, is
likely to be found where it is required of employees that they exercise
delegated authority or specialized knowledge and expertise in the interests of
their, employing organization. In the nature of the case, such employees
must then be accorded a legitimate area of autonomy and discretion and, to
this extent, their performance will depend on the degree of moral commit
ment that they feel towards the organization rather than on the efficacy of
external sanctions. A service relationship can thus be understood as the
means through which an employing organization seeks to create and sustain
such commitment; or, that is, as a functional alternative to direct control in
regard to those employees whom the organization must to some significant
extent trust to make decisions and to carry them through m ways that are
consistent with organizational values and goals.14
It is, therefore, the distinction between employees involved in a service
relationship with their employer and those whose employment relationships
are essentially regulated by a labour contract that underlies the way inwhich, within our class schema, different employee classes have been
delineated. The most obvious division to be made in this respect is that
between the predominantly salaried professional, higher technical, admin
istrative, and managerial positions of Classes I and II and the predominantly'
wage-earning manual occupations of Classes VI and VII. The former may be
taken as those positions with which a service relationship is most character
istically associated, and thus as constituting the basis of the service class jar
salariat of modern industrial societies; the latter, as those where the labour
contract usually prevails, and which thus constitute the basis of the working
class. We find it of interest and significance that something close to this
division receives rather widespread linguistic recognition: for example, |in
the distinction m ade in English between staff and workers; in French,
between cadres or employes and ouvriers] in German, between Beamte or
Angestellte and Arbeiter; or in Swedish between tjansteman (literally, service
m en) and arbetare. We would, furtherm ore, see a reflection of a similar
division in the distinction drawn in state socialist societies between intellig

14
W hile on e w ould, then, from this point o f view , expect to find a close association b etw een
- t y p e o f em ploym ent relationship and the con tent o f work tasks, and roles, it should be
em phasized that it is th e/orm errath er than the latter that, for us, is decisive in determ ining class
position. Thus, in the case suggested by som e exp on en ts o f the proletarianization thesis (e.g.
Crompton 1980) o f em ployees w ho are perform ing routine administrative work'altho'ugh with
the cosm etic title o f manager, the crucial question is w hether this title does or d oes not carry
wi it the advantages, o f a service relationship with the em ploying organization.

Scanned by CamScanner

2 . Concepts, Data, and Strategies o f Enquiry

43

entsia and workers (cf. Szczepansk. 1970; Hardin 1976), even though the
historical context and wider connotations are here of course m some respe
" ^
0^ , at the same time to recognize that the contrast that we
have set upbetween a service relationship and employment regulated via a
labour contract is one of an ideal-typical kind, and that ^ a c t u a l empioymen
relationships will often only approximate one type or the^oth
Y
indeed fall rather ambiguously between them. Conseque Y2.1 again brings out, we elaborate on our basic division of employees in two
ways.
First, the distinctions between Classes I and II and Classes VI
made within the service class and the working class respectively. P o rtio n s
covered by.Class I may be taken as those to which the largest responsibi
in decision-making attach and which will in turn offer the fullest rang
beneficial conditions associated with the service relationship; while in the
case of the lower-level positions of Class II, certain of these features may be
attenuated. Conversely, it is with non-skilled labour as represented by Class
VII, where there is least need for employees to be allowed autonom y and
discretion and where external controls can be most fully relied on, th at the
labour contract will tend to operate in its purest form; while, with skilled
labour as represented by Class VI, modifications directed towards making
the money-for-effort exchange somewhat less specific and short term are
more likely to be found.16
Secondly, we distinguish two classes which may be regarded as interm edi
ate in the sense that they comprise positions with associated employm ent
relationships that would appear characteristically to take on a very mixed
form. Class III covers the range of routine non-manual positions, usually
involving clerical, sales, or personal-service tasks* which exist, so to speak,
on the fringes of professional, administrative, and managerial bureaucracies;
and Class V takes in lower-grade technical and first-line supervisory positions
whose incumbents usually work closely with rank-and-file manual employees,
although being in certain respects differentiated from them . In both cases,
the extent to which a service relationship could be said to prevail over the
15 For d eb ates on t h e "class role o f th e intelligentsia in E astern E u rop ean n ation s, se e .
Konrad and S zelenyi (1979), Starski (1982), and F rentzel-Z agorska and Z agorski (1989).
16 For ex a m p le, w orkers in Class V I could b e regarded as m ore likely than th ose in C lass V II
to b e includ ed in internal; or craft-specific labour m arkets. It m ight b e a d d e d jh e re that
em ployers appear increasingly prepared to introd uce service ele m en ts in to th e em p loym en t
relationships o f various graces o f m anual worker in p osition s in volvin g high lev e ls o f skill or
responsibility, so that th e latter m ight then be m ore appropriately allocated to C lass V (se e
follow ing te x t). A n d , certain ly, w e w ou ld n ot regard th e m an u ality or oth erw ise o f work
p erform ed as b eing in itself relevan t to class allocation con sisten tly w ith th e p osition tak en in
n. 14 ab ove. H o w ev e r, oyer th e p eriod to which our data relate; w hich, as w ill b e s e e n , exten ds
several d ecad es b ack from th e m id -1970s, the m an u al-n on -m an u al division d o es in fact appear
to h ave corresp on ded rather clo sely w ith that w e have noted o f w orkers versus 's t a f f , e t c ., so
far as em p lo y m en t relations are con cern ed .

44

Concepts, Data, and Strategies o f Enquiry

presuppositions of a labour contract is problematic. EmpLoying organizations

appear to display a frequent uncertainty over w hether the roles performed
by such personnel are ones that w ould justify staff status or whether they
are to be treated essentially as labour. It is notable that difficulties arising
from this uncertainty in regard, say, to m ethods o f remuneration, tim e
keeping, prom otion chances, union representation, etc. have been a
recurrent them e in studies o f em p lo y er-em p loyee relations concerning the
groups in question.17
It should be added that the subdivision of Class III into Ilia and III6 that is
further indicated in Table 2.1 w as prom pted by the application of the schema
in studies of w om ens m obility, and is used only in analyses where women
are in volved . T he purpose of the subdivision is to isolate in Class III6 a range
of routine and very low-skill non-manual positions which are largely occupied
by w om en and to which (especially when held by wom en) very little
ambiguity in fact attaches. That is to say, these positions tend, in contrast to
th ose retained in Class I lia , to be m ore-or-less undifferentiated in their
conditions o f em ploym ent from those of non-skilled manual workers. Thus,
in analyses in which the subdivision is applied, Class IIIi> is usually collapsed
with Class V ila .
T here rem ains one other feature of the class schema on which some
com m ent should be made: nam ely, the sectoral division through which
proprietors and wage-workers in agriculture and other primary production
are separately identified as Classes IV c and Vllfc. This elaboration would
seem necessary in view o f various distinctive aspects assumed by class
relations within primary production in consequence, for exam ple, of the
m ajor form o f property ownership being in land, o f the organization of
production rem aining often fam ily-based, and of the frequent substitution,
even in the case o f labour drawn from outside the family, of payment-in-kind
for m oney w ages. Ideally, we w ould also have wished to introduce into the
schem a som e degree of differentiation am ong agricultural proprietors, in
regard to form o f tenure, size and value o f holding* etc. H ow ever, there
appeared little point in so doing, since relevant information, at least in! a
form that would permit cross-national'com parisons, is not sufficiently
available in the data-sets on which w e rely.
1
From what has already been said in this and the preceding section it will be
apparent that the class schem a is not constructed around any single hierarch
ical principle from which a regular ordering of the classes could be derived.
If, therefore, for particular analytical purposes, such an ordering would
seem desirable, it must be produced by reference to external criteria. A s we
have already remarked, we would doubt if any very ambitious m oves in this
!:

-- S e e , for ex am p le, o n clerical w o rk e rs , M ills (1951), L o ckw ood (1958/1989), C ro zier

(1965), K ocka (1980a, 1981); on fo re m e n an d su p e rv iso rs, T h u rle y and W ird en iu s (1973) and
C hild (1976); a n d on tech n ician s, R o b e rts , L o v e rid g e , and G e n n a rd (1972) a n d L ow -B eer
(1978).

45

Concepts, Data, and Strategies o f Enquiry

direction could be justified. H ow ever, good grounds w ould appear to exist

for introducing within the seven-class version o f the schem a a threefold
hierarchical division which could be m ore-or-less equally w ell taken as
ordering class positions in terms o f their prestige, so cio -eco n o m ic status,
or general desirability. W e w ould, in fact, regard scales purporting to rank
occupations according to such criteria as all reflecting in a rather sim ilar way
differences in, on the o n e hand, levels o f job rewards and, on the oth er, jobentry requirem ents (cf. G oldthorpe and H o p e 1972, 1974).
In Table 2.2 w e show the results achieved if w e assign scores to the classes
of the schem a on the Treim an international occupational prestige scale
(Treiman 1977a) and also on a num ber of scales available for individual
nations that will be represented in our subsequent analyses. O n each scale
we have obtained a score for every occupational and em p loym en t status
grouping com prised by a particular class to the ex ten t that a m apping
betw een the categories o f the scale and those o f th e class schem a proved

T a b l e 2 .2 . Scores f o r classes o f the sch em a o n d iffe re n t o c c u p a tio n a l sca les as a basis

fo r a threefold hierarchical d ivisio n

Class
Scale*

I+ II

III

IV a + 6

IV c

V+VI

V ila

VHb

Treiman

56

35

42

44

35

29

24

H o p e-G o ld th o rp e
(England)

63

36

39

47

40

29

31

W egener (F R G )

92

50

49

50

49

39

30

Irish O ccupational
Index (all Ireland)

58

30

42

42

37

24

-26

de L illo Schizzerotto (Italy)

71

41

51

48

34

20

11

N aoi (Japan)

62

41

37

37

41

33

30

27

46

25

33

17

14

Duncan (U S A )

66

V
.

D ivision

1 ,

Note:
* The international Treiman scale and those for the FRG, Ireland, and Japan are intended as
scales of occupational prestige, although constructed in different ways; the English scale and
also,it would seem, the Italian, areintended as ones of the general desirability of occupations in
popular estimation; and the US scale, while originally constructed as a proxy for a prestige
scale, is now generally interpreted as one of the socio-economic status of occupations. For
further details, see Treiman (1911a), Goldthorpe and Hope (1974), Wegener (1988) Boyle
(1976), de Lillo and Schizzerotto (1985), Naoi (1979), and Duncan (1961).

46

47

2 . Concepts, Data, and Strategies o f Enquiry

2 . Concepts, Data, and Strategies o f Enquiry

possible and we have then taken the median of these values as the overall
class score Although scores are not, of course, directly comparable from
scale to scale, it is still evident enough that Class T+II, the service class or
salanat consistently ranks above all others, and that Classes V ila and VIIb
those of non-skilled w orkers in industry and of agricultural workers, consist
ently rank below all others, while the relative positions of the remaining
classy are cross-nationally rather variable. A threefold hierarchical division
of the schema gu.ded by these results, and as indicated in Table 2.2, would
therefore seem well founded.

seen, these are rather diverse; and, secondly, it should be made in comparison
with what might be offered by the available alternatives.

T here is, how ever, one modification that we would wish to make in the
case of the class of farmers, IVc. In the course of industrialization, agriculture
typically undergoes a radical transform ation. Peasant, or other kinds of
largely subsistence, farm ing give way to m ore decisively m arket-oriented
forms of production, and then in turn commercialized family farming is in
part superseded by relatively large-scale agribusiness. In this process, a
substantia] decline in the num ber of farm ers goes together with a steady
increase in the average size of farms and in levels of capital investment and
sales values (see, e.g ., Renborg 1969; Newby 1978). As we have mentioned
above, we cannot, with the data at our disposal, differentiate among farmers
in such a way that w ould allow us to capture these changes directly in our
analyses. B ut in those cases where we utilize our hierarchical division, we
can, even if only crudely, still try to give some recognition to what is in effect
the collective upward mobility of farm ers by treating Class IVc differently as
a class of origin and of destination. That is to say, we can take Class IVc,
following T able 2.2., as falling in the interm ediate division when considered
as a destination class, but then relegate it to the lowest division, along with
Classes V ila and V ffij, when considered as an origin class.
In conclusion of this section, we would again wish to emphasize that the
schema we have presen ted is to be regarded not as an attem pt at providing a
definitive m ap of the class structures of individual societies but essentially
as an instrum ent de travail.18 As we have sought to make clear, its construc
tion and adaptation have indeed been guided by theoretical ideas but also
by m ore practical considerations of the context in which, and purposes for
which, it is to be used and of the nature of the data to which it is to be applied.
In turn, the crucial test qf the schema, as of any other conceptual device,
must lie in its perform ance: it must be judged by the value that it proves to
have in enquiry and analysis. But any such judgem ent, we would then
suggest, will need to m eet two requirem ents: first it must be made across the
entire range of applications in which the schema is involved and, as will Be
18
T hu s, it d oes not aim at respon d in g to th e q uestion p osed by R uncim an (1990) o f H ow
m any classes are th er e? th e o n ly.sen sib le an sw er to w hich is, w e w ould b eliev e, A s m any as it
p roves em pirically usefu l to distin gu ish for the analytical purposes in h a n d . E v en while
preserving the u nderlying idea o f the sch em a that classes are to be defined in term s o f
em p loym en t re lations, th e d ifferen tiation o f th ese , follow in g th e pattern o f Figure 2 .1 , could
ob viou sly b e m uch further e x te n d e d , w ere there g ood reason to d o so.

DATA

At the end of the previous chapter we argued that one m ajor reason for the
unsatisfactory state of comparative mobility research is that data of the kind
that have been typically utilized are seriously inadequate: in particular, they
lack comparability from one national case to another, and indeed to such an
extent that one might question the value of basing any analysis upon them.
In this section we seek first of all to amplify this point and then to describe
how, for the purposes o f the present work, we have attem pted to generate
data to which greater reliability might attach.
Until recently, most comparative research has been based on the published
results of national mobility enquiries: that is, on the results of these
enquiries as reported in books and journal articles. Such data are, of course,
readily accessed and, as their range has widened, they have offered tem pting
possibilities to the quantitatively oriented macrosociologist. It would, none
the less, seem clear that with such data severe, and in fact largely intractable,
problems of comparability are encountered in (at least) two respects.
First, national enquiries, as might be expected, differ in many of the
technical details of their design and conduct: for example, in their precise
population coverage, in their sampling and weighting procedures, in the
wording of questionnaire items, etc. In the literature to which we here refer,
such problems have in fact received little attention whether because it has
been bravely, but unjustifiably, assumed that they can be of no great
consequence or because, in working with already processed data, what can
be done fo overcome them is in any event very limited.
Secondly, as we have been concerned to emphasize above, mobility can
be studied from different conceptual standpoints, and attem pts to render
these operational in different national contexts have created great diversity
in the social categorizations by reference to which mobility has been defined,
observed, and measured. In the face of this difficulty, the standard practice

19
T he m ost ob viou s extant alternatives are the tw o different class sch em ata p roposed by
W right (1978, 1985) o f an explicitly M arxist character (and s e e also the further com m en ts o f
interest in W right, 1989b). W e do n ot o u rselves find the theoretical basis o f eith er o f these
sociologically convincing; a nd, fu rtherm ore, serious prob lem s arise in im p lem entin g at least the
earlier version in in tergen eiation al m obility stu dies, since class allocation in part d ep en d s on
interview data on individuals' su b jective assessm ents o f their d egree o f authority and autonom y
in w ork, w hich are u n likely to b e available for th e parental gen eration. H o w ev e r, com parisons
m ad e b etw een the results o f u sin g ou r ow n and W rights sch em ata across a range o f issues in the
field o f class analysis m ust form th e m a m basis o f any com p etitive evalu ation (cf. G oldthorpe
1990: 4 0 6 -9 ) . In this resp ect, the m ajor contrib u tion thus far is that o f M arshall and his
co llea g u es (M arshall et al: 1988; M arshall and R o s e 1990), w hich indicates that a version o f the
class sch em a h ere u sed perform s appreciably b etter in d isplaying a variety o f class effects than
d oes eith er o f W rights p rop osals.