Social Mobility in Britain -- the 'Oxford Studies' (aka the 'Nuffield Studies')


Social mobility research is of considerable significance in sociology. It is interesting in its own right to chart the amount, type and effects of social mobility (ie movements up or down the class system). This file pursues another main interest - social mobility studies as offering information about the class system - to see how "open" or "closed" it is. This is clearly a very important issue for a number of reasons - e.g. open societies are generally held to be more "fair" than closed ones. People seem more prepared to put up with a system of inequalities if access to the best positions is open and fair. For various reasons, a system open to merit is often seen as fairest  Social mobility studies are a good way to test how close Britain is to achieving an open stratification system. The test would be - is membership of upper group open to members regardless of the occupation of their fathers? (NB the Nuffield studies focussed entirely on males - this choice is critically discussed below)

Two contrasting possibilities arise:

(a) fully open - the "service class" (sc) contains members from a very wide spread of social origins. The social origins of fathers are irrelevant in crucial stages of selection - e.g. the education system.

(b) fully closed - There is little or no social mobility, and classes self-recruit (ie sons enter the same class as their fathers) as in a classic class system.

This is the sort of test commonly used and there are several ways to actually do it:

(i) predict what the social system would look like if it were perfectly open and fluid, and then measure the actual results to compare:

(ii) establish, by statistical means the relative influences of social origin and "merit".

The Glass Study

Until the Oxford Studies, the famous study by Glass was the key reference for evidence of social mobility. The survey was done in 1949 when about 4,000 males were studied to see what happened to them (i.e. they were asked about class of origin, their first job, and their present job to see if any movement up or down had occurred).

Glass found Britain was pretty "closed" with little mobility, and what there was limited in range, and mostly from manual occupations into a semi-manual "buffer zone". It turns out that Glass was actually pretty unlucky however - in 1949 he was still measuring the effects of the economic and political systems of the 1900's, one characterised by depression and slump. Other oddities in the study are discussed by Payne G (1977).

The Oxford or Nuffield Studies

The Nuffield Studies (Goldthorpe, J 1980 and Halsey, A et al 1980) were deliberately designed to research social mobility. About 10,000 males were researched to test a number of social "theses" about social mobility - for our purposes relating to issues openness and closure.

J Goldthorpe et al

The results are displayed in a special form below. Graphics don't look very good here, I'm afraid.You might wish to print them out separately anyway. An  electronic version is available to print separately. The best form of display in our experience consists of a number of small boxes, each with a symbol in them.  Compare this form of display with the original tables in Goldthorpe and Halsey (and with the interesting variant in Marsh, C 1988)


From J. Goldthorpe Tables 2:1 & 2:2 pp. 44-48 (errors averaged and rounded up)

KEY: (each symbol represents 1% of the sample)

Top row                                                                                      Bottom row
S -- service class origin                                                                        = destined to stay in class

I -- intermediate class origin                                                                  (single arrow up) destined to move up one

W --  working class origin                                                                    (double arrow up)destined to move up two

                                                                                                           (single arrow down) destined to move down one

                                                                                                           (double arrow down) destined to move down two

Interpreting the table

(a) Compare and contrast the sizes of service class with working class in the top and the bottom rows of the diagram. We can see that the service class has grown considerably over time. This is important because we would expect to find social mobility simply because this growth has made room for newcomers. 

(b) Take the top row of the diagram (the position in 1972). We can see in the service class lots of men with a working class origin (about 1/3 in all - 8 squares with W  in them). So there is a lot of real social mobility about, and it is not short-range, with no buffer zone. It is easy to see why experience is likely to lead you to think that Britain in 1972 was very open - you had a one chance in three of meeting a service class member from a working class background. 

Note there was even some downward mobility. Why is this important? To see if all this mobility is the result of a genuine change in the basic class allocation or just overall growth. There is in fact nowhere near enough downward mobility to indicate full "random" mobility. As other charts (not reproduced here) show, some of this downward mobility is only temporary.

(c) Now look at the lower row. Note that this is an abstraction (father's class at 14 - for some respondents, this will describe a picture of 1922, for others, a picture of 1952). More detailed work on different cohorts, shows gradually increasing mobility, in fact. We get a different picture of openness, however, by looking at "outflow" - the chances of going up or staying put.

We see that, in the service class, 8 squares are going to stay in service class - (8 equals signs out of 14 symbols in all). This gives a simple class chance then of about 1 in 2 of staying in service class for sons of s.c. fathers. 

For the working class group, 8 squares are going to end in service class (8 double upward arrows in the diagram) - but 8 out of 55, not 8 out of 14, a simple class chance of 1 in 9. There are considerable differences in these chances, then, viewed from the outflow end.

(d) Looking back to the top row, it is clear that those 8 squares in the service class with a W  in them are really in a very select group - 8 in 55. The 8 squares in the service class in 1972 with an S in them are rather less select since there were only 14 squares to select from in the first place (so to speak). 

From the top (inflow) the class system looks equal and open - 8 squares from each class. But from the bottom (outflow) there are very different chances of gaining social mobility.

So far there seems to be much more social mobility than most marxists think. There is no conspiracy to totally close off opportunity. The study specifically refutes Parkin's "closure" notion (which suggests the middle class manages on the whole to ensure its sons fill the upper positions by "closing off" entry from below). The "buffer zone" thesis is also rejected, as we have seen, and further arguments in the study reject the "counterbalance" thesis - that opportunities for mobility via one route (education, say) are counterbalanced by restrictions in opportuities via another route such as "on the job" mobility). Briefly, there are just too many men of w.c. origin in the s.c. to support any of these theses. There is enough absolute mobility to mean everyone knows a socially mobile proletarian.

But let's not get carried away and think Britain is totally open, and that your father's class makes no difference. Class still affects your chances of mobility a good deal. There is no evidence for meritocracy here, no complete mobility, the picture, as usual is complex. The complexities are not likely to be well-known, however, except to those (researchers) who know the whole picture. To use the terms of the study: there is much "absolute" but little "relative" mobility.

Now I have returned to this file (after some years!!), I can include Goldthorpe's brilliant diagram of a fluidity matrix (no scanners when I first did this). Here it is

fluidity matrix

What a lovely thing this is! it is drawn to scale, so it shows you the changes in the size of the classes over the period studied. Look down the left margin to see the size of the social classes in the past (when the respondents were 14). Now look across the top margin to see the picture in 1972. The service class, for example, is an oblong shape -- there were more people in it in 1972 than there were in the past. Class IV is almost a square, so ther has been little change in size. These changes are important, because expanding the service class will provide more opportunities for social mobility all on its own,without any actual change in the distribution of opportunites across the classes..

Now look at the different levels of shading. This indicates fluidity - that is, the extent to which the social class recruits from its own members. If the service class changed each generation, so that it was filled with entirely new people born in classes outside, it would indicate maximum fluidity. In this diagram, it would be shaded white, at level 8. As you can see though, Class I is the most 'unfluid' ('solid'? 'torpid', 'dense'?) social class and is shaded black (level 8). If you follow the diagonal on the diagram you will be looking at the self-recruiting qualities of each class. All classes self-recruit quite a lot. Note that Class IV is quite dense -- this Class is self-employed people who classically pass on the business to their sons and thus self-recruit quite a lot (not as much as the Class I though).

Now look at the columns. Each column (not row), shows the composition of the class in terms of entrants born inside and outside that class in 1972.  Each row (not column) shows the picture in the past. Things have inmproved in that some people at least born in  Classes VI and VII have made it in to the service class, compared to the position in the past.

Go back and read the discussion for yourself --  the table comes appears on p. 110 in Gol;dthorpe et al (1980)

Exercise 1

At this point you might like to pause and work through the interpretations of the tables above and try the following:

1. Check to your own satisafaction that the symbols match up in both rows of the diagram - eg that those destined to move up two classes in the bottom row can be matched with those of w.c. origins in the s.c.. in the top row

2. Calculate the class chances for yourself (NB the Nuffield study did not use such a simple method)

3. Use the data to comment on the well known disagreements between sociologists and lay persons on the topic of whether social class is still significant in affecting your chances of mobility in Britain.

Other findings

(1) the influence of educational qualifications is growing - but there are still many opportunities to work your way up once in a job (but start before you're 35!). This assumes the considerable economic growth in 1972 is still going (see Goldthorpe 198). 

(2) The effects on class consciousness are summarised thus:

(a) members of the working class are increasingly homogenous and generational - but NOT hostile. Nevertheless, there is a basis for a "class" view of society, social policy, and social change, and a chance for socialism (Goldthorpe concludes). 

(b) the intermediate classes are really in a mess and are fragmented, with lots of identity crises and problems. 

(c) the service class is surprisingly stable with easy and smooth integration of the socially mobile. It is relatively easy to make friends, (although there is a core of some "isolates" among the socially mobile - about 20% of those who went up two classes). Apart from that, there was no real evidence of strain (which functionalists had predicted as a consequence of high rates of mobility), and no disaffected, radicalised mobile proletarians who were prepared to denounce the service class from the inside (as some marxists had hoped). 

Basic Problems

Heath (1981) provides a convenient discussion. Briefly, there are problems with the study like these:

(a) definitional (how many classes, for example) - a problem since more classes mean more mobility. One real problem concerns the rather general and inclusive service class rather than a tighter notion like a ruling class or elite, which may be much less open. Maybe important distinctions exist within classes too - e.g. those of Hopper re status.

(b) Males only were studied. The social mobility of women is more complex and yet vital for understanding class. There has been much discussion of this point in subsequent articles (see here and see below).

(c) The research is dated. We've seen the end of the boom. There may be no more growth in the service class. 

A. Halsey et al

Halsey's volume in the series makes similar points about absolute and relative mobility. The data is summarised in the same format below.


From A Halsey et al Tables 4.4 and 4.5 pp.52, 53 (errors averaged and rounded up)

KEY (NB each symbol represents 1% of the sample. Symbols with slashes indicate fractions of 1%) 

Top half                                                                                              Bottom half
S -- service class origin                                                                                              A -- destined to go to sec mod

I --  intermed class. origin                                                                                          B -- destined to go to gr sch

W -- working class origin                                                                                           C -- destined to go to pub sch

                                                                                                                                   D -- destined to go to tech

                                                                                                                                   E -- destined to go to comp

                                                                                                                                    F -- destined to go to dir grant


From A Halsey et al Tables 10.2 and 10.3 pp. 180,182

KEY (NB each symbol represents 1% of the sample. Symbols in superscript represent fractions of 1%) 

Top half                                                                                        Bottom half
S service class origin                                                                                           A destined to go to no post-sec.

I intermed class origin                                                                                         B destined to go to p/t f.e.

W working class origin                                                                                       C destined to go to university

                                                                                                                           D destined to go to Coll of Ed

Comments on Halsey

In general, the picture is similar to the one discussed for Goldthorpe.

1.If we take grammar school entrance (TABLE 2), for example, it is clear that grammar schools seen from the inflow end were relatively open to entry from below. A third of grammar school boys were from, a lot more than critics thought. Many of these were first generation, many did quite well. There is no real evidence to suggest, from the inflow end at least, that grammar schools were merely "middle class enclaves".

2. From the outflow end, though, chances of getting to grammar school do reflect the influence of parental class. Using our simple calculations (NOT those used in the studies), the chances of a boy from an s.c. family getting to gramnmar school were about 5 in 13 (5 squares with B in them in TABLE 2, bottom row), or about 1 in 3. The chances of a w.c. boy were only 6 in 56 (5.75 squares with B in them, TABLE 2 bottom row, working class column), or 1 in 7.

Exercise 2

1. Go on to calculate the relative chances of s.c. and w.c. boys going to university, using TABLE 3

2. What evidence is there for the view that part-time f.e. is more open than any other form of post-school education?

3. Why might a teacher in a boys' grammar school be likely to think that Britain was a more open society than it really was?

4. How typical of the population were entrants to Colleges of Education? Why might they appear to be more open than universities? What were the chances of a w.c. boy ending up in one?

Additional findings in Halsey include: 

(a) The importance of the private sector emerges. The ability to buy a private education is, in fact, the main way in which the social class of parents affects the chances of the child. The study specifically tests (in a very controversial way)  and rejects Bourdieu's notions of "cultural capital" as the main vehicle of transmission of class advantages.

(b) School is uneven in its effect on working class kids. The social class of parents, for example, does affect the chances of staying into the 6th form - but if a child can survive, class no longer has an effect. Further Education actually helps working class kids once in at the expense of middle class students - but see Raffe (198 )

(c) Halsey argues that the elite tripartite system was about as open as the American comprehensive systems and is sceptical about British moves to comprehensive schools as doing any better.

(d) The tripartite system was close to exhausting middle class demands for schooling, leaving real opportunities for working class children - of course, that was in 1972 before the massive restructuring.


These are massive and important empirical studies of social mobility which seem to offer serious challenges to any simple account of class and mobility in Britain, whether from left or right. Class and merit interact (again), open-ness and closure take complex forms. The same data can be viewed from different directions - from inflow and outflow perspectives, in absolute and relative terms.

The studies have been heavily criticised too, and the debate has continued pretty regularly since the publication of the results,as many articles in the British journal Sociology testify. In brief, the main criticisms concern:

omitting women

being out of date

for operationalising class or cultural capital in controversial ways

I would be very pleased to discuss the criticisms, and the (pretty plausible) responses of the team (largely Goldthorpe) with you if you email me


Goldthorpe J et al Social Mobility and Class Structure in Modern Britain, Oxford, Clarendon Press, 1980 
Halsey A et al Origins and Destinations..., Oxford, Clarendon Press, 1980
Heath A Social Mobility , London, Fontana 1981
Marsh C Exploring Data..., Oxford, Polity Press, 1988
Payne G "A re-appraisal of 'Social Mobility in Britain' " in Sociology vol.11, 1977
 and see the subsequent files on social mobility (back to index) See also the RLO on social mobility