Centre
for Research on Innovation and CompetitionCentre
for Research on Innovation and CompetitionArchive
In 'mainstream' innovation studies, there has been comparatively little investigation into the dynamic nature of consumption behaviour (relative to studies that have looked at the supply side). The main exception is the diffusion of innovation literature. However, we have argued that many of these studies have offered oversimplified accounts by treating populations of potential adopters as being socially homogeneous. One of the most well known models of adoption distinguishes between innovative adopters, the early majority, the late majority and laggards. In other words, they are simply defined by their propensity to adopt. The explanation is typically a combination of income and a notion of motivation towards things novel. Social factors and processes have been dealt with tangentially, if at all. Essentially, then, the model is individualistic, with no consideration of relationships between different social groups. This, we believe, is inadequate for understanding the diffusion of products in consumer markets. The extent to which social factors explain consumption patterns is central to debates in cultural studies and the sociology of consumption. One central question relates to the extent that individuals' tastes are shared with others in society (captured by the figure below).
This project has used insights from these literatures to examine the dynamic nature of consumption behaviour.
Empirical Work and Data Sources
Data from the General Household Survey (GHS) and the Health and Lifestyle Survey (HALS) were used to conduct the empirical analysis for these studies.
The GHS is an annual representative survey of around 10000 UK households, which is available in electronic form from the early 1970s. The survey has a large amount of data on household expenditures and assets as well as details on individuals within the household. This includes socio-demographic variables such as age, occupational class and education. There is detail about what sorts of technologies are present within the household such as microwaves, CD players, dishwashers and telephones.
The data for HALS were initially collected in 1984 and 1985 to form a random sample of 9003 respondents aged 18 or over and resident in households in Great Britain. The data collected included many variables related to the area of consumption and lifestyles. For example, detailed data on food consumption, smoking, alcohol consumption, hobbies, exercise etc. as well as socio-demographic variables including social class, household composition were gathered. The interviewees were traced and re-interviewed seven years later (the follow-up survey) and basically the same questions were repeated. Thus we have similar data from two points in time for the same people.
A variety of descriptive and multivariate statistical tests were carried out on these datasets to study the diffusion of household durables, the social basis of food consumption routines and trends in the consumption of cultural and leisure ('lifestyle') activities.
The analyses aimed to demonstrate that social factors such as occupational class will be significant determinants of routine consumption behaviour, whereas traditional neo-classical economic analyses of consumption typically take income as the primary determinant.
For the class variables, we have the basic Registrar General (RG) class of the household head as well as the more detailed socio-economic group (SEG). We are concentrating on RG class defined below. The rationale for using RG class is that it closely corresponds to the social class scheme used by advertisers in their marketing strategies (i.e. the familiar A B C1 C2 D E grouping). RG class is analogous to this and defined as follows:
I
Professionals
II
Managers
IIIN
Routine white collar workers
IIIM
Skilled manual workers
IV
Semi skilled workers
V
Unskilled workers
The regression models also analyse the importance of income, gender, age and education.
Socially differential rates of adoption for household durables
This analysis looked at the adoption profiles in 1986 and 1994 for dishwashers, microwaves and freezers. The results, comparing the adoption of different classes are presented in the figures below.
The adoption of household durables by class
The patterns of adoption in these charts clearly show the importance of occupational class as a determinant of consumer durable adoption. In addition, using a multivariate approach, we tested for the existence of class and income effects in the same model. The table below displays the results of such an analysis (in this case a logistic regression using dishwasher ownership in 1994 as a dichotomous dependent variable and including class of household head as dummy variables and household income as independent variables).
| Logistic regression predicting the adoption of dishwashers in 1994 |
||
| Variable |
Coefficient |
Standard Error |
| CLASS |
|
|
| (BASE CLASS IS V) |
||
This shows that, despite having controlled for income (which is of course highly significant) we still have highly significant occupational class effects. Therefore there is strong evidence that some other differentiating factors are at work in the consumers' tastes for dishwashers. These tastes are conditioned by factors related to class and therefore are not completely explained by occupational income differences.
Socially constituted consumption routines
This study examined the dynamic nature of socially constituted consumption routines. We defined a consumption routine as an executable capability for repeated consumption that has been learned or acquired by groups of consumers in response to social pressures or contexts. The analysis concentrated on the routine consumption of foods. There are several interesting findings:
First, we have shown occupational class to be a strong determinant of consumption routines. Modelling consumption by past social status is an unusual way to do this, and is probably an even stronger indication of its importance that simply using current social class. It suggests the longer-term persistence of occupationally constituted habitus in appropriately defining the groups within which routines are shared.
Second, we have shown that social mobility can have a significant effect on consumption routines. This resonates with our definition with respect to the notion of learning or adaptation. It suggests that consumers that have shifted from say blue collar occupations to white collar occupations will in some cases adapt their consumption behaviour accordingly. This they might do consciously, actively aspiring to behave in similar ways as their new colleagues, or through unconscious mimicking.
Third, the effects of these influences are different depending on the good that is being examined. This is particularly interesting since it implies that goods differ with respect to the way that they are perceived by different social goods; in other words some goods appear to carry more social significance that other goods. We have characterised three different types of good. Type 3 (socially ambivalent) goods appear to constitute little significant social meaning, whereas type 1 (habitus dominated) and type 2 (mobility adjusted) do. But the social significance of these goods is different.
The existence of consumption routines is particularly significant for those interested in the diffusion of innovative consumer products. The implication is that existing routines need to be modified or broken for innovations to succeed. This is reflected in practice, as advertisers and market research functions attempt both to reinforce routine consumption behaviour and to break it. This they frequently do through activities that are based on stratified populations of consumers. Marketers are well known for using the ABC1C2DE scheme (analogous to the scheme we have used here). Consequently product ranges are designed so that a hierarchy of products are offered to different social groups. Adverts too are created and presented in a manner to make clear the social significance of consuming a certain good.
We have demonstrated that consumption routines can change over time for certain goods and at different rates for different social groups. Understanding consumption behaviour as governed by routines and social mechanisms is essential to understanding the evolution of demand.
Lifestyles and Taste
This study looked at distinct consumption lifestyles identified with the PRINCALS (a version of principal component analysis) technique. This technique proved extremely fruitful in showing that not only are these lifestyles identifiable, but that they can be shown to coincide with different traditional social groups in British society. Furthermore, trajectories can be plotted to show how and to what extent other factors can come into play in modifying a social group's behaviour such as age or gender.
The results presented in figure 3 above show that distinct class-based lifestyles existed in the mid-eighties and early nineties and that the notion of post-Fordist lifestyle might have to be modified on the basis of these data. For example, there is little evidence here that there is a 'two-thirds society' split between skilled manual and white collar workers on one hand, and an unskilled underclass as Lash and Urry might suggest.
Significance of Results and Outcomes
The analyses described above successfully argue for attention to the dynamic nature of final consumption. The dynamics of interest are evident at a number of (interrelated) levels. First, there are changes in patterns of consumption that emerge through macro-social shifts (changing relationships between different social groups, demographic shifts such as the ageing population). Second, there are changes in the structure of consumption brought about by shift in the structure of production and retailing. Third, there are changes to practices of consumption with the innovation of new goods and services. These shifts are understood within a framework that sees consumption as stratified, relational and non-utility maximising. These relationships are seen as crucial for understanding the innovation and consumption of new goods and services.
Tomlinson, M. and McMeekin, A. (1998) "Does the 'Social' have a Role in the Evolution of Consumption?", CRIC Discussion Paper No. 14, Manchester University
McMeekin, A. and Tomlinson, M. (1997) "The Diffusion of Household Durables in the UK", CRIC Discussion Paper No. 4, Manchester University.
McMeekin, A. and Tomlinson, T. (1998) 'Diffusion with Distinction: The Diffusion of Household Durables in the UK', Futures, vol. 30, no. 9, p. 873-886.
McMeekin, A., Green, K., Tomlinson, M. and Walsh, V. (2001) "Innovation by Demand: Interdisciplinary Approaches to the Study of Demand and its Role in Innovation', Manchester University Press.
Tomlinson, M. and McMeekin, A. (2001) 'The Evolution of Consumption Routines', in McMeekin, A., Green, K., Tomlinson, M. and Walsh, V. (2001) "Innovation by Demand: Interdisciplinary Approaches to the Study of Demand and its Role in Innovation', Manchester University Press.
Tomlinson, M. 'Lifestyles and Social Class', CRIC Discussion Paper No. 9.
Warde, A., (1997) 'Consumption, food and taste: culinary antinomies and commodity culture', Sage, London
[Back to projects]
CRIC has combined with PREST to form the Manchester Institute of Innovation Research (MIoIR).
New book: Trust in Food, A Comparative and Institutional Analysis by Unni Kjaernes, Mark Harvey & Alan Warde.
CRIC Final Report to ESRC:"Main Report" and "CRIC Performance Indicators 1997-2006".
'Instituted Or Embedded? Legal, Fiscal and Economic Institutionalisation of Markets' by Mark Harvey
'Beyond Efficiency and Market Shares: Competition within the Finnish Games Industry' by Mirva Peltoniemi
'Accounting for Economic Evolution: Fitness and the Population Method' by Stan Metcalfe
'Innovation and Final Consumption: Social Practices, Instituted Modes of Provision and Intermediation' by Andrew McMeekin & Dale Southerton
'Alfred Marshall’s Mecca: Reconciling the Theories of Value and Development' by Stan Metcalfe