Classes or ‘microclasses’? The nature of occupational inequality

Bush Junior & SeniorFrom Bush to Bush. Miliband to Miliband. Kennedy to Kennedy. Churchill to Churchill. There’s no shortage of political dynasties either in the US or UK, where politically powerful parents beget politically powerful children. Likewise, we often observe how doctors’ children become doctors, or indeed, academics’ children become academics. (This is the moment for self-confession: my father and both my brothers also have PhDs. Tragically my greatest act of rebellion was becoming a social rather than natural scientist…).

When we come to look at class inequalities, we tend to look at ‘big classes’ like professionals. But what if society is instead mainly stratified by narrower occupational groups – ‘microclasses’? Spurred by the latest instalment of a series of papers in the world’s top social science journal (AJS), in these two posts I look at the role of microclasses in the US and beyond.

Different types of class structures

The study of microclasses is closely associated with two US academics, Kim Weeden and David Grusky. While not their first statement of their view, their 2005 paper in the American Journal of Sociology (AJS) was the first contact I had with their ideas, and they’ve since followed this up with two more papers in the same journal – most recently in 2012, from which most of this post is taken. All of these papers are wonderfully long and dense in the way that AJS papers can be, so inevitably I’ll be doing a gross injustice to their ideas in these short posts…

The conventional ‘big class’ view of the world is where “inequalities in life conditions are organized around such familiar sociological categories as professional, manager, service worker, craft worker, and laborer… members of a big class marry within their class, form confidant and support networks within their class, and interact mainly with one another. These within-class ties are facilitated by class-based segregation in the education system, in neighborhoods and communities, and in the workplace.

Their basic claim is that the forces that make people in one occupation different to another are primarily specific to that particular job (or microclass). It’s not that they’re arguing for fundamentally different processes to the big-class model, but rather that “they operate principally at the level of institutionalized occupations.” So you will take a route into that occupation which is different from other people in your big class; your working conditions will be occupation-specific; you’ll socialise mainly with people in that occupation; and your occupational environment will tend to push you towards particular lifestyles and attitudes.

Just to take one small example close to my heart: they (probably correctly) suggest that sociologists are generally “humanist,  antimaterialist, and left-leaning.” In a microclass way of thinking, they would explain this through “(a) the left-leaning reputation of sociology and the consequent self-selection of left-leaning recruits, (b) the inculcation of a liberal worldview through lengthy professional training and socialization, and (c) the reinforcing effects of interaction with like-minded colleagues.” So it’s not just about the general fact of being in the big professional class – else why would economists be different to sociologists?

They also differentiate this from the ‘gradational’ view that there is simply an occupational hierarchy of income and prestige. If this was the case then people in better jobs would be differentiated from people in worse jobs, but there wouldn’t be sharp differences between either microclasses or big classes.

From claims to evidence

That’s the claim, then – but does it hold up in practice? The best visual display of this is in the 2012 paper, looking at occupational stratification in the US in the 1970s across a whole host of different outcomes measures:

Weeden-Grusky 2012 figure 4

To interpret this figure:

  • Firstly look at the top bar in each category (the dark black bar on the left, the hatched bar on the right); let’s start with the ‘all outcomes’ one at the bottom. The bar to the right shows how much the association of class with ALL of these outcomes is due to microclasses (58%); the bar to the left shows how much is due to big classes (42%).
  • Secondly, look at the bottom bar in each category – the solid grey bar and the (small) white bars. This shows how much of the big class and microclass effects are gradational – that is, they go from the highest-income big/micro-classes to the bottom-income classes in a straightforward way. Most of the big-class associations are gradational in this way, but almost none of the microclass associations are.

In other words,  specific microclasses have a stronger association with outcomes than either big classes or occupational income in 1970s America. Which is a big finding! Only for ‘life chances’ – education, employment and home ownership – do they find big classes matter more than microclasses, and microclasses are still important here. They explain the lesser importance of big classes outside of life chances by saying that “this pattern is consistent with the claim that intermediary organizations no longer convert big classes into culturally coherent communities” [p1757].

Their 2005 paper looked across multiple other measures of stratification, and across 30 years of data (1972-2002), claiming that “trends proved to be relatively weak and do not alter our main conclusions here” (p155). But their 2012 paper goes into these trends in more detail, and finds that in several areas big class matters less in the 2000s than it did in the 1970s, particularly for lifestyles and attitudes. In contrast, microclass inequalities have barely shifted, and in the case of political attitudes may even have increased. So the idea that class matters less than it used to is – in their view – because people haven’t been looking enough at microclasses.

Grusky & Weeden are not saying that big classes and gradational models are completely irrelevant (although they have some provocative article titles…) – but rather that we can’t understand inequalities between people in different jobs unless we take microclasses into account.

To be continued…

Perhaps the most interesting part of their work, however, is when they look at intergenerational mobility – what parents pass on to their kids. In the follow-up post, I’ll look at their examination of microclass mobility in the US, Sweden, Germany and Japan, and the (typically) fierce critique made by one of Britain’s most notable sociologists, John Goldthorpe.

About Ben Baumberg Geiger

I am a Senior Lecturer in Sociology and Social Policy at the School of Social Policy, Sociology and Social Research (SSPSSR) at the University of Kent. I also helped set up the collaborative research blog Inequalities, where I write articles and short blog posts. I have a wide range of research interests, at the moment focusing on disability, the workplace, inequality, deservingness and the future of the benefits system, and the relationship between evidence and policy. You can find out more about me at
This entry was posted in Articles and tagged , . Bookmark the permalink.

13 Responses to Classes or ‘microclasses’? The nature of occupational inequality

  1. mel bartley says:

    Oh dear! This reminds me of the 1990s debate on how best to measure social position. And that was one reason I decided to use the term ‘social position’ as an umbrella for all kinds of inequality measures. “Social class” is not meant to be about attitudes, consumption patterns, intermarriage, co-mensality, or friendship. True that there may be close relationships between class and any of these other aspects of social existence; sometimes there is and sometimes not, and it varies between times and cultures. No wonder John Goldthorpe is getting agitated, he must think he and others like Gordon Marshall and Tak Wing Chan have already dealt with all this. Amanda Sacker and I compared the power of a class measure (based on employment relations and conditions) and the Cambridge Scale which measures something more like Grusky is talking about, to predict health risk. And we duly found that anything to do with consumption and lifestyle was better predicted by the Cambridge Scale. John G and Wing have now deveoped another measure using similar methods to the Cambridge but they call this a measure of prestige explicitly. It is based on clusters of occupations who mix socially and inter-marry. I don’t remember exactly but I think it related to health behaviours in the same sort of way (we never got round to looking at this properly). John and Wing wrote a series of papers that are very worth looking at which say far more than I could do here. But as Gordon Marshall pointed out, we need to avoid a ‘beauty contest’ between class and status. They measure different things and it is useful that they do.

    • Ben Baumberg says:

      Happy new year Mel! Sorry for not being clear about this – Grusky/Weeden’s point is slightly different. It’s not that they’re arguing for a status measure against a class measure. Instead, they’re arguing that stratification of whatever sort happens at the more fine-grained level of microclasses, rather than either as an overall gradational scale or by complicated patterns among big classes.

      I completely agree with you about the pointlessness of trying to reduce everything to either class or status though. I’ve previously blogged about Tak Wing Chan and John Goldthorpe’s wonderful paper on the usefulness of both, which you refer to – as I say there, it’s one of my favourite papers in recent years.

  2. Bill Gardner says:

    So Ben… your picture shows the two George Bushes (both Andover, Yale, and Presidents). In my Andover class of 250 or so boys, we had a future governor (Jeb Bush, R-FL) and a future Senator and Governor (Linc Chafee, Independent-RI). Jeb is the son of GHW Bush and Linc of John Chafee, former Senator from RI. They clearly had huge advantages of family. And the families were clearly part of a microclass of political professionals, which provided the sons with entry and advantage in politics. That said, I assume you are not denying that there was also a class advantage that transcends microclass? Jeb & GHW Bush, for example, are not just career politicians. They have moved back and forth between public office and private equity.

    • Ben Baumberg says:

      It’s a really nice example for why we need to consider all the different aspects of class, rather than restricting ourselves to one of them. When we come to ask ‘what do parents pass on to their children’, I guess the main answer is ‘a whole bundle of different things that defy easy categorisation’…

  3. Charlotte Cavaille says:

    when is the second post comming??? waiting with anticipation. I am just confused to be honest by this whole debate. especially their conclusion. take the increase in the relevance of micro-classes for explaining the distribution of political attitudes: they cannot say if it comes from an actual effect of micro occupations as such (microclass as context) or because of the change in the income homogeneity of members (microclass as containers). The whole thing is just confusing (didn’t they test for an increase in “gradationalism”)…I think I just don’t understand this piece…

  4. Charlotte Cavaille says:


  5. This is an example of really academic sociology – a nondebate. Not that much use to policy makers and of little interest to the general public.

    So what do they show here? Essentially when it comes to more attitudinal things, more detailed occupations are better at capturing them than coarse big classes. Furthermore, this explanatory power of detailed occupations does not seem to have much to do with income differences between occupations – unlike the exlantory power of social classes – so it must be other things correlated with occupation(s) which are are not correlated with class as much. Finally, the explanatory power of detailed occupations – or rather the things they proxy – is stable. (Excuse my use of term ‘explanatory power’, haven’t read the article, but you get what I mean).

    This is great if you want to use occupation as a control to capture hosts of things correlated with occupation you cannot readily measure, say, if your primary interest is in another variable – just used detailed occupations and not social class to ‘soak up’ as many other confounders as possible. But we knew this anyway. More detailed independent variables explain more variation in the data – common sense. But what are the research implications? What about when you want to use detailed occupations to explain things and come up with a story? The problem here is there are so many occupations, it will be difficult to coherently summarise findings – there are likely to be many factors at work for which occupations proxy – possibly too many to keep track of or come up with any fairly coherent story. The article alludes to many vague things occupations are proxying for. What do detailed occupations proxy for? There are so many things. One thing that springs to mind is personality traits. I would like to see an analysis that pitches occupations versus personality (e.g. big 5). Big classes themselves are supposed to proxy for employment relations according to Goldthorpe. If employment relations are so important, why not just measure those directly? In other words, why don’t we just measure the underlying constructs themselves? I’d much rather deal with 5 facets of personality than 120+ occupations. Perhaps all they are showing here is that personality is more strongly related to your occupation and attitudes than in the past – which is more interesting than 120+ occupations is better than a handful of more aggregated ones in a model.

    PS Brilliant post Ben – nicely summarised a long and dense article.

    PPS I’m obviously exorcising a few demons here as I got sucked into this “debate” at one time or another and my conclusion is as I state here – let’s try and not use proxies – and if we have to – let’s make sure we have a reasonable number of categories!

    PPPS Detailed occupations can be used to really identify which segments of the labour market are responsible for sweeping over time trends – like how in my own research I showed that a large proportion of growing inequality in Great Britain is due to just a handful of detailed occupations. But by and large, overall, I also show that general trends are more related to big classes. So both approaches have merits – but I’d rather be able to explain this by categorising detailed occupations in more meaningful ways. Merely saying detailed occupations are better than big classes is a nondebate! Which occupations and which classes and why is what’s important. Occupations are proxies and there are many.

  6. Pingback: Microclass mobility (and its critics) | Inequalities

  7. Charlotte Cavaille says:

    Mark these are refreshing words of wisdom, have you read this book: 2011 The Explanation of Social Action. Oxford University Press. ?
    you will like it!
    I am still confused on whether the three types are alternatives approaches to social stratification or if micro and gradationalism are two alternative ways of breaking down big class logic. Also wouldn’t mind any thought on my comment about individual vs context.

    • Ben Baumberg says:

      I’ll try and say something next week in response (ran out of time this week after trying to make sense of the mobility paper). But it won’t be as interesting and well-read as Mark’s comment above!

  8. Pingback: So should we bother with ‘microclasses’? | Inequalities

  9. Pingback: Inequality and civic morality | Inequalities

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.