1. *University of Sheffield †University of Southampton
Crime Through the Lens of Political Socialisation: An Analysis of Age, Period and Cohort Effects on Worry about Crime and Disorder in England and Wales. Crime, Punishment and Society CCR Conference 2017
Hello and welcome to this panel.
Emily Gray*, Maria Grasso* Stephen Farrall*, Will Jennings† and Colin Hay *
*University of Sheffield †University of Southampton

2. Introduction: long term trends in crime and politics
‘The crime problem’ (Garland, 2001) has been a key issue in UK political and social agendas and a hot topic in public opinion in various ways over time:
We have identified relationships between property crime and the economy (Jennings et al 2012) and shown how public punitiveness rises thermostatically in relation to crime rates which is translated into patterns of imprisonment in the UK (Jennings et al, 2016) .
Amongst friends in the UK and Internationally; Stan Cohen; New Left Realism; Jock Young; David Garland; Enns; Gottschalk; Miller.
Topics such as competition between political parties; punitivemess, fear of crime; victimisation. Rise in concern about anti-social behaviour.
So, myself Steve and Will Jennings have been looking at long term trends in crime and specifically how those trends can be explained by political and economic conditions. We’ve looked in particular at the impact of Thatcherism, but it would be fair to say Thatcherism predates and post dates Thatcher. So we’ve more accurately been looking at trends from 1960’s to the era of New Labour.
For example, we’ve looked into public punitiveness over time. Our analysis demonstrated that the public recognise and respond to rising (and falling) levels of crime, and that in turn public support for being tough on crime is translated into patterns of imprisonment. Very much akin to Enns work in the States.
Will was also able to show statistically significant relationships between the impact of unemployment and the economy on property crime in the 1980s.
We’re not the only people to be interested in socio-economic and political conditions of course. This has been a topic for academics in the UK and internationally who have looked at competition between political parties, public punitiveness, the antecendents of fear of crime, concern about anti-social behaviour, sentencing, to name but a few.

3. Figure 1. Property crime per capita (Home Office Recorded Statistics and BCS)
And the relationship between crime and politics is an understandable one. Crime has gone through some major changes in the last 50 years, it has risen steeply and fallen – even if we can’t really why. It has changed dramatically (we are now adapting to how theft and violence can take place in virtual spaces).
Of course, this graph shows us that Thatcher and Major in particular resided over period of steep rises in property crime that was identified in both recorded and self-report figures.

4. Figure 2: % ‘Worried about crime’ – BCS 1982-2004
And again, perhaps in response to the rise in crime, the public began to express a lot more concern about crime, again, peaking in the 1990’s, which became a social problem in and of itself.

5. Figure 3: Mentions of crime types in official debates in parliament – Hansard, 1950-2004
Notably if we look at Hansard we can see how often different crime types were mentioned in political debates in our British parliament over 50 years and we can see some very clear trends in relation to some offences – namely rises in mentions in antisocial behvaviour from the late 1990s, rises in discussion of burglary in the 80’s and smaller but clear spikes in mentions of mugging in the 70’s and later again in the 80’s.

6. Can long term data identify political generations?
Gaps in exploring long term trends in crime and politics. Little exploration of political generations.
Authors have sought to highlight the sponge-like qualities of crime; crime might ‘absorb’ measures of social stability and community cohesion. An appraisal of one’s peers, environment and political security (Ferraro, 1995; Girling et al, 2000; Jackson, 2006).
So, do Thatcher’s children differ from Blair’s babes? How do these populations differ from others that came earlier?
As mentioned, we’re not the first to be interested in the intersection of crime and politics. However we saw that there was a bit of gap in terms of an empirical analysis of the long term impact of key legislation and policy drives that impacted crime. Not just crime rates, but developments in housing and eduction. Which Steve will talk more about.
We were also interested in how different political generations may variously relate to crime. Certainly, for example, the fear of crime literature talks about crime as being an expression of ontological insecurity. So, if people fuse crime and disorder with a general unease about the state of social stability and the direction their society is perceived to be moving towards it is conceivable that the political context they grew up in will bare an influence on their beliefs about crime. Specifically for us, the question became did coming of age during Margaret Thatcher’s political ascendancy - a period in which crime rates soared and a ‘law and order’ agenda intensified - differentiate this political generation’s fear of crime from previous and subsequent political cohorts? Did a heavy emphasis on anti-social behavior (ASB) and penal populism in the late 20th century lead those who grew up under New Labour to view neighborhood disorder differently from older generations? How do generations who came of age in other political periods compare?
Lucky for us, a wonderful colleague in the politics department, Maria Grasso, was able to help up with those questions and we began to look at age period and cohort analyses.

7. Age, Period, Cohort Analysis
Longitudinal analysis of BCS/ CSEW data ( ) affords an APC analysis (Ryder, 1965).
Exploration of the unique contributions of three types of time-related variations - ‘age, period, and cohort’ effects.
Age effects - changes in the life course; period effects - arise via specific cultural and economic changes; cohort effects are the core of social change and represent the effects of formative experiences (Ryder 1965).
So, with Maria, we looked at data from the British crime survey, to conduct and age period and cohort analysis to see if we could identify political generations within it.
Age period and cohort effects are all temporal effects, they are all about the passage of time, but may well be unique from each other. So, age effects are changes that take place in the life course, we’re all familiar with the age crime curve, we don’t expect to see many pensioners robbing cars or vandalising fish and chip shops – these are the sorts of changes that take place during one’s life; period effects - arise via specific cultural and economic changes – such as for example big political events, 9/11 or the fall of the berlin wall etc, which may temporarily impact how people think; cohort effects are the core of social change and represent the effects of formative experiences (Ryder 1965).
Naturally, when you are looking at long term trends, it is not only interesting but important to see if you can assess the different temporal effects at play . Are we identifying age effects? Period effects? Or Cohort effects?
And it is sometimes possible to build statistical models that can disentangle these issues, that work well with repeated measures cross-sectional data such as BCS,
It is worth mentioning that there is some work to do with building these APC models. You need to structure the data in a particular way in order to avoid identification problems. That happens if two of the values are known (i.e. a respondent’s age and year of interview) the third value in the relationship (the birth cohort) is automatically known. To deal with this we categorise and constrain the data and use generalized additive models (GAMs) for plotting the identified smoothed cohort effects as a post-test of the data, which is arguably accepted in the field. (Grasso, 2014; Jiang and Carriere, 2014).
So, in an APC model we isolate and control for the various effects of ageing, time and period in order to identify ‘pure’ cohort, or generational effects using regression analyses.
How do we create political generations? In order to develop a categorisation of cohorts we have to select a key period of socialization, in general, scholars of political generations select the period of late-adolescence/early-adulthood. Political socialisation is not about the time you were born as such, but the political period in which you spent your formative years, the point at which you are most influenced by political discourses, which is considered to be We did play around with other age brackets, but there was no difference in the outcomes.

8. Table 1 Political generations:
Pre-Consensus Generation
Post-War Consensus
Generation
Wilson/ Callaghan Generation
Thatcher/
Major Generation
New Labour Generation
Formative period
(14 years)
(18 years)
(13 years)
Years of birth
Aged 15
Total N
(%)
8,435
(9.61%)
23,181
(26.41%)
21,653
(24.67%)
27,527
(31.36%)
6,980
(7.95%)
Because of the data we have built - we have linked together all of the British crime surveys/ CSEW from ( I say that as if it didn’t take the best part of 6 months), but we did put them together which resulted in a very large data base of people who were born from 1930 to This data is now available for public use on the UK data service.
We therefore distinguish between five distinct generations for the APC analysis conducted in this paper. The assumption is that the historical periods in which individuals have spent the majority of their ‘formative years’ are sufficiently different as to warrant the distinction of five generations. These are presented here.
So we have the five political generations, which are the pre-consensus generation; which we distinguished from the post-war generation. Followed by the Wilson and Callaghan Generation (economic decline and trade union disruption). The Thatcher and Major years of neo-liberalism. And finally the new labour generation.

9. Thatcher’s children? BCS/ CSEW data 1982-2012
Table 2. BCS indicators for analysis
Questionnaire item
Original response categories
Recode for analysis
Could you tell me how worried you are about...? Having your home broken into and being burgled
1 Very worried
2 Fairly worried
3 Not very worried
4 Not at all worried
1/2= 1 worried about being mugged
3/4= 0 not worried
Could you tell me how worried you are about...?
Being mugged and robbed
How common a problem (in this area) are...?
Groups of teenagers making a nuisance .
1 Very big problem
2 Fairly big problem
3 Not very big problem
4 Not a problem
1/2= 1 teenagers a problem
3/4= 0 not a problem
Noisy Neighbours.
Long term analysis such as this means you need long runs on questions – so we looked at fear of crime and other perceptions of crime and disorder which appear in the BSC/ CSEW from And these are the questions we used, which you may be familiar with.
The first two questions are direct questions about worry of particular offences, burglary and being mugged or robbed in the last year.
Meanwhile, the last two questions picks up on public sentiments around unruly, but not necessarily criminal groups of children or neighbours. These are more subjective assessments of one’s environment, which we heard a lot about from the 1990’s onwards.

10. Multivaraite models for analysis
Sex
Ethnicity
N Victimisation (in prior 12 months)
Married
Employment status
Education history
Income
Tenure
Since the literature demonstrates that fear of crime and perceptions of disorder can be linked to a broad range of social-demographics it was essential to build a multivariate model that controls for various personal and social circumstances.
Of course one of the issues of secondary data analysis is that you are limited to the decisions prior researchers have made about what data to collect. We don’t have reliable measures of religion, class or region.

11. Table 3. Results from multivariate identified APC models
Fear of Mugging
Fear of
Burglary
Problem –Neighbours
Problem -Teenagers
Pre-Welfare State (ref. cat.)
Post-War Consensus Generation
1.332***
1.504***
1.541***
1.561***
Wilson/ Callaghan Generation
1.441***
1.750***
1.997***
2.054***
Thatcher’s Children Generation
1.303***
1.777***
2.220***
2.324***
New Labour’s Children Generation
1.364***
1.609***
2.517***
2.620***
16-34 years of age (ref. cat.)
35-59 years of age
0.973*
0.966*
0.954*
0.952***
60+ years of age
0.959
0.871***
0.711***
0.761***
Year of interview
0.965***
0.996***
0.979*** N
347,512
348,622
429,631
429,197
These are the results from the regression analysis. All the models were identified and indicate some clear and interesting narratives and evidence of political generations.
We used the pre-welfare generation as the reference category.
So, lets turn first to fear of crime. All four later generations are more concerned than the Pre-Welfare State Generation about both burglary and robbery.
Thatcher’s generations were the most worried about property crime - which is salient, given the fact this generation ‘came of age’ during the dramatic rise in property crime during the 80’s. The Wilson/ Callaghan Generation followed quickly behind.
The Wilson/ Callaghan Generation also expressed the highest levels of worry about mugging (street-based crime) – which again, chimes with the moral panics about mugging which took place during the early 1970’s – which Stuart Hall wrote about in his very popular ‘policing the crisis’ title.
Then we turn to ‘common problems in the local area. All four later generations are more concerned than the pre-Welfare generation about teenagers hanging around and noisy neighbours. Thatcher’s Children are more than twice as likely and New Labour’s Children are almost two and a half times more likely to be concerned about teenagers hanging out than the pre-consensus generation. There is a similar pattern when one considers noisy neighbours. New Labours children, the youngest generation are much more likely to perceive these issues as problematic.
This suggests the public rhetoric around law and order that emerged during Thatcher’s time in office and most specifically during the New Labour administrations (Burney, 2005) is much more salient for the youngest political generation who demonstrate the most intolerance for these low-level public order offences.

12. Summary of results
Pre-welfare state generation least worried about burglary / robbery and anti-social behaviour.
Thatcher’s generation most worried about property crime.
Wilson/ Callaghan’s generation most worried about mugging.
New Labour’s children most concerned about anti-social behaviour
The multivariate analysis demonstrated that strongest predictors of fear of crime and perceptions of disorder were the political generations. However, there were also significant relationships for gender, ethnicity, recent victimisation and level of educational attainment.
Across all four indicators the patterns are unambiguous; there was an upward swing in worry about burglary and worry about mugging and perceptions of disorder once the Wilson/Callaghan generation is born after the 1940s.
These patterns suggest that anxiety about crime and disorder grew considerably and notably peaked for some generations around particular historical moments.

13. Enduring significance of crime
What we appeared to find - and this is a paper that is undergoing an R&R at the moment, so is subject to a little more tinkering - but our analysis confirms the notion that an individual’s socialization in a given political context is an important and formative experience in regards to their understanding of crime and disorder. It would appear that popular political discourses around crime can have a meaningful and lasting influence on the attitudes and beliefs of a political generation.
The example herein demonstrates the value of developing a close dialogue with political history, and perhaps, as criminologists we could profit further from creating even more opportunities to think politically about crime and perceptions of crime (c.f. Aviram, 2015; Enns, 2014; Gottschalk, 2006; Loader and Sparks, 2016).

14. Perceptions of crime across political generations
The political culture into which one is born, has a lasting effect on an individual’s political values, and their perception of particular manifestations of crime.
Are there other generational schema?
Fashions in crime will shift over time, in response to crime rates and popular political discourses.
Need to devise methodological and theoretical strategies for incorporating ‘slow moving relationships’.
So, crime-related concerns do appear to be related to political generations. It may be that there are other forms of generational schema’s which one could construct. These could include economic generations, based on periods of recession/expansion for example.
Our work suggests that the public’s preoccupation with certain crimes will shift over time, possibly in relation to crime rates, but also in line with socio-political discourses. Thinking about these conditions means thinking more carefully about criminal justice debates. It also means devising methodological and theoretical strategies for incorporating these slow-moving relationships – which are rarely captured in criminological research, and which we certainly hope to do more of in the coming future. Which Steve will talk more about in the last presentation.

15. Contact Details Email: emily.gray@Sheffield.ac.uk
Data available at the UK Data Service. Lodged in 2016 as SN7875
emily.graysheffield.ac.uk