1Recent studies in sociology have shown that French cultural practices have changed during the second half of the twentieth century through the combined effects of mass education and the development of the cultural industries. Olivier Donnat (2011a) highlights four major trends between the 1970s and 2000s: a steady increase in audiovisual consumption, a decline in book reading, the growth in cultural practices by enthusiasts, particularly those related to using digital technology, and a rise in occasional visits to cultural institutions. New technologies and the media compete with school and cultural institutions in terms of their power to transmit culture (Octobre 2009). In this context we can examine the cultural practices and tastes of teachers, the key actors in cultural education in schools. How might they be considered the guarantors of traditional forms of cultural legitimacy? The available statistical data, based on surveys conducted in the 1960s and 1970s, show that primary and secondary school teachers read a lot and often visit places of culture, in particular in comparison with other graduate occupations (Berger and Benjamin 1964; Chapoulie 1987). However, the cultural practices of teachers during the recent period are little understood and this article attempts to analyse the changes underway.
2Research has highlighted the declining strength of the link between qualifications and legitimate cultural practices, particularly reading (Coulangeon 2011; Dumontier, Singly and Thélot 1990; Donnat 2011a), and we might expect teachers to follow this dynamic. However, if only because of their closeness to educational institutions, we can assume that the link between teachers and legitimate cultural practices has been maintained more than for other qualified categories, even if sociological changes among personnel and changes to their work activities during recent decades might have destabilized this link.
3In this analysis, I adopt a quantitative approach to change, in that I consider changes over time to cultural excursions and the amount of reading. I first explain my questions and hypotheses by emphasizing what unites but also what distinguishes, primary and secondary school teachers and other graduates from the point of view of their relationships to culture. With the aim of uncovering changes and understanding their roots, my empirical approach has two stages. In the first stage, I compare differences in the cultural practices of teacher categories to other graduate categories with the help of the “Pratiques culturelles des Français” (“Cultural practices of the French”) surveys (Ministère de la Culture and de la Communication) for the 1981–2008 period.  Then, secondly, I study the cultural practices of secondary school teachers in more detail, using a method based on the comparison of a “first hand” questionnaire survey conducted in 2008 with a survey conducted by Jean-Michel Chapoulie and Dominique Merllié in 1970 (Chapoulie 1987; Chapoulie and Merllié 1971), which investigated several lifestyle aspects of this population. This comparative method, based on a substantial and more targeted population, enables me to link the intensity of legitimate cultural practices in the two periods to certain sociological changes to the population of teachers in collèges [lower secondary education] and lycées [upper secondary education].
The legitimate cultural practices of the teachers in question
4In La distinction, Pierre Bourdieu ( 1984) states that cultural practices and tastes cannot be contemplated independent of levels of education and social origins, and are arranged in social space according to a principle of structural homology, from the most legitimate (the knowledge culture of the dominant classes) to the least legitimate (popular culture): “In cultural consumption, the main opposition, by overall capital value, is between the practices designated by their rarity as distinguished, those of the fractions richest in both economic and cultural capital, and the practices socially identified as vulgar because they are both easy and common, those of the fractions poorest in both these respects” (p. 176). The questions regarding the model of cultural legitimacy are now well known (Grignon and Passeron 1989; Lahire 2004; Peterson and Simkus 1992; Peterson and Kern 1996). In particular, recent research has shown that the link between qualifications and legitimate cultural practices seems weakened, above all regarding book reading, challenging the solidity of the relationship between the latter and teachers.
The weakening of the link between educational qualifications and cultural practices
5The decline in legitimate cultural practices among those with a baccalauréat and higher education graduates has been highlighted since the beginning of the 1990s. An article by Françoise Dumontier, François de Singly and Claude Thélot (1990) shows that between 1960 and 1980 theatre and concert visits have changed little, much as the proportion of heavy readers has remained substantially unchanged. Yet, in the context of an unprecedented rise in education levels in French society, the authors emphasize that this in reality represents a decline: “France is reading more but the French are reading less” (p. 65), “The French are going to the theatre and concerts as often, but those with the same education levels are going less often” (p. 73). Only museum and heritage site attendance seems spared (Manger 1993). Analysis of the reduced attachment to reading among graduates was taken up again by Philippe Coulangeon (2007), who shows, with the help of four surveys ranging over time, that there is a rising proportion of non-readers and a decline in the proportion of heavy readers among those with baccalauréats and higher education degrees that is much more pronounced than among those without baccalauréats. In the same way, the proportion of heavy television watchers has increased among those with baccalauréats and higher education degrees. The author highlights a generational dynamic: the behaviour of the best educated across the cohorts has approached that in the least educated categories, reading less and watching more television.
6These changes suggest various underlying causes. Firstly, the changes in the sociological characteristics of graduates through the expansion of education should be taken into account. In the second half of the twentieth century, the level of education of the population rose; secondary and university courses opened their doors to children from all social backgrounds (Merle 2009; Prost 1997), enabling the socialization of graduates in various forms of culture within educational institutions (Coulangeon 2007, 2011). The growing importance of technical and scientific education in the curricula leads at the same time to a weakening of literary socialization in the context of school (Mauger 1992; Dumontier, Singly and Thélot 1990). The link between a well-educated population and a “cultivated” culture has thus lost its meaning.
7Secondly, cultural transmission through school is in competition with the dynamism of the cultural industries that regularly bring new types of product to market. There has been a very large rise in the consumption of audiovisual equipment since the 1980s (Herpin and Verger 2008). These products expose users to new content and offer new means of access to content, differing greatly from the linear learning often taught in school (Octobre 2009).
8Thirdly, although cultural practices accumulate rather than compete (Coulangeon and Lemel 2009; Octobre 2009), an increase in cultural offerings reduces the time devoted to each activity, especially as the amount of free time decreases for the best educated as they enter the labour market (Chenu and Herpin 2002), in particular for time consuming practices such as reading books. Let us also bear in mind that, if we acknowledge that surveys provide declarative material rather than an objective measure, the response rate to surveys conducted over several decades may vary through the effect of changing representations associated with practices, in particular when they become commonplace (Mauger 1992).
9Moreover, legitimate cultural practices have not all experienced the same dynamic. In particular, the observation of a decline in theatre attendance (Dumontier, Singly and Thélot 1990) has been challenged by the availability of new survey techniques. Donnat (2011a), profiting from the possibility of comparing five waves of the “Pratiques culturelles des Français” survey showed an increase in cultural excursions in the area of live entertainment. This increase does not signify a parallel reduction in social distances: the gaps between senior executives/liberal professions  and employees/manual workers remain stable in terms of theatre attendance, even though attendance by both increased, and the gaps have tended to increase in the case of dance performances and classical music concerts. These studies are corroborated by Coulangeon (2011, p. 44), who highlights that the most visible or elitist practices, which can “go hand in hand with a fairly superficial relationship to their content,” are the most class structured.
10According to Donnat (2011a), the increase in theatre attendance is linked to more frequent visits by the over-40s, but also a growing propensity of young generations to attend (the exception however being classical music concerts). Educational encouragement comes into consideration here, as well as the cultural facilities available. The cultural offer has multiplied in Paris as in large and medium-sized cities (Menger 1993). Indeed, political enthusiasm for the performing arts, and the theatre in particular, has been particularly significant in the second half or the twentieth century, with the aim of democratizing culture, as evident in the creation of the Festival d’Avignon, for example (Ethis, Fabiani and Malinas 2008). Performing arts attendance has thus attracted “a proportion of those who have benefitted from the lowering of higher education access requirements, while conserving their power of attraction over the most educated sections of the population” (Donnat 2011a, p. 25).
11However, we should specify certain characteristics of graduates. Bernard Lahire (2002) has shown that there are variations in relationships to reading as a function of university course followed. Similarly, it is likely that graduates are differentiated by their membership of a profession. The occupational world of teachers, and their social trajectories, may heighten their awareness of the most learned forms of culture, at least if we accept the analyses written on the subject.
Teachers and legitimate culture
12The relationship between teachers and legitimate culture can be understood in two ways, neither of which rules the other out. Firstly, studies on attendance at cultural venues point to the presence of teachers among their publics. This is the case for museums, libraries, media libraries and cultural centres (Berthier 2003; Messu 1994). Teachers also form a large proportion of live performance audiences; research on attendance of the Festival d’Avignon has highlighted their significance among audiences, alongside those in information, arts and entertainment occupations (Ethis, Fabiani and Malinas 2008). Similarly, teachers can often be found among theatre season-ticket holders (Cibois 2003) and opera audiences (Doublet 2003). Note that they can be solicited by cultural establishments to educate young audiences in culture (Saint-Cyr 2003) and that they benefit from incentives granting them free access to some museums and national monuments. However, the representation of teachers among the publics tells us nothing of the penetration of practices among them.
13Although few, statistical studies on teachers’ cultural practices also point to a strong link. Ida Berger and Roger Benjamin (1964) point out that a large proportion of primary school teachers, male and female, assign a cultural and educational dimension to their free time, prioritizing reading, the cinema and theatre and rejecting television. According to J.-M. Chapoulie (1987), the proportion of leisure time devoted to reading is greater among secondary school teachers than among managers, as is the amount of daily newspaper reading, in particular Le Monde, and reading weekly current affairs magazines, such as Le Nouvel Observateur and L’Express. The author also shows that teachers owning televisions make much more selective use of them and that their theatre, museum and concert attendance is considerable, even when their place of residence does not facilitate this since they live and work in small or medium sized towns. J.-M. Chapoulie, however, makes it clear that this behaviour varies greatly depending on grade and sex, and that the same would undoubtedly hold true for those in other intellectual occupations able to organize their work schedules flexibly, and with limited financial resources.
14It should be added that teachers’ culture, at least until recently, was distinguished by its relationship to the school system. School has long been the main authority for socialization in cultural knowledge for teachers in secondary and tertiary education, and even more for primary school teachers from modest social backgrounds (Muel-Dreyfus 1983; Gerbod 1965). According to Bourdieu (1984 ), primary school teachers are part of the petty bourgeoisie in search of increased social standing; they exhibit “cultural goodwill,” recognizing legitimate culture more than knowing about it. Identifying culture with knowledge, they seek above all to hoard it, which leads them to a certain confusion of genres, and the accumulation of disparate knowledge and cultural tastes. Secondary teachers, graduates from the petty bourgeoisie, are in turn the “erudite,” custodians and guarantors of a learned culture, making “prudent” and “homogenous” cultural choices (Bourdieu 1984 , p. 265), disinclined to distance themselves from legitimate culture or to value the avant-garde insofar as their cultivated habitus is intrinsically tied to late learning acquired in school. In addition, secondary school teachers, a dominated group of the dominant class “hardly ever have the means to match their tastes, and this disparity between cultural and economic capital condemns them to an ascetic aestheticism” (ibid. p. 287). They are distinguishable from “socialites” who have a complicit relationship with the culture to which they are socialized from a young age within the family setting, and who are more able to dissociate themselves from the visible traces of the genesis of their cultural dispositions.
The potential effects of sociological changes in the teaching population
15Recent changes experienced by the teaching population have challenged the resilience of its consumption of legitimate culture. Without ignoring the multiplicity and complexity of these changes, I would like to focus on some of them in particular, whose effects on cultural practices are analysed in the empirical section of this study. Some changes lead to the hypothesis of a consumption of legitimate culture that remains high while others invite us to suppose the opposite.
Aging and demographic renewal
16The age structures of teaching personnel in the middle of the twentieth and the beginning of the twenty-first centuries are clearly distinguishable. Recruitment policies to respond to the growing demand for education led to the entry of many young teachers up to the end of the 1970s (Chapoulie 1987; Prost 2004). The “second school explosion” (Chauvel 1998) led to a new rise in teaching numbers in secondary education at the turn of the 1980s–1990s (Prost 1997), but at the beginning of the twenty-first century, teachers are on average rather old.  We can therefore investigate the effect of age on the intensity of cultural practices in the case of teachers, knowing that, in general, the most legitimate cultural practices are positively correlated with the age of individuals (Coulangeon 2005).
17But above all, this raises the question of a renewal of cultural practices and tastes alongside the generational renewal experienced by the teaching sector in the last few years (Baraton and Perronnet 2009). Knowing that the cultural practices of graduates undergo generational dynamics (Coulangeon 2007; Donnat 2011a), we might expect that young teachers questioned in different periods to have contrasting representations and habits with regard to culture, and that their consumption of legitimate culture declines. Moreover, we know that young generations of teachers are distinguished from older generations in their relationship to teaching. They are more distant from the school system and their work role, and, more than their elders, recognize themselves as teachers with more technical and pedagogical skills and less centred on mastery of a discipline or academic knowledge (Geay 2010; Rayou and van Zanten 2004; Maroy 2006), which can have similar effects on consumption of legitimate culture.
The effect of subject taught
18Additionally, we might ask ourselves to what extent the subject taught impacts on the cultural practices of teachers in collèges and lycées, and whether a subject effect persists in the recent period. Studies of publics in cultural places highlight the presence of humanities teachers (Ethis, Fabiani and Malinas 2008). Similarly, the importance of reading and cultural activities in the university courses followed by teachers of subjects in the humanities (Lahire 2002) suggests that the latter’s cultural practices are more frequent than those of colleagues in other disciplines.
19However, teaching in the humanities is not endowed with the same meaning at the beginning of the twenty-first century as it was in the middle of the twentieth century. The link between literary and school culture has eroded in parallel with changes to the brief of secondary education, no longer is it called on to cultivate the children of elites with no consideration of utility but to train qualified workers in response to the needs of the economy (Isambert-Jamati 1970; Prost 2004; Demailly 1985). The objectives set for teaching French encourage teachers to adopt a wider conception of culture aimed at the acquisition of techniques and methods rather than a literary culture (Baudelot, Cartier and Détrez 1999). The link between legitimate cultural practices and teaching subjects in the humanities may therefore have weakened.
20Furthermore, the structure of professional teaching bodies by subject type has been altered by the introduction of technical subjects in secondary teaching. J.-M. Chapoulie (1987, p. 17) reminds us that, in 1966, 92% of posts in the CAPES and CAPET competitive teaching examinations and 98% of posts in the agrégation competitive examination were in the humanities and “traditional” sciences, while in 1977 the proportions declined to 70% and 85% respectively, with the recruitment of teachers in artistic and technical subjects to these levels. As a result, the differences in subject composition of teaching bodies may vary the intensity of cultural practices in the long term.
Changes in lifestyles
21Other grounds for investigation concern the lifestyles of teachers, the single life being common among teachers until the 1960s. For female primary school teachers this was the result of their geographic isolation and intellectual solitude (Ozouf and Ozouf 1992). As for female secondary school teachers, until the middle of the twentieth century this seemed to be mostly a life choice (Cacouault-Bitaud 2007). The solitary life favoured intellectual pastimes, especially reading, greatly encouraged by the school system (Ozouf and Ozouf 1992; Muel-Dreyfus 1983). Marriage is now the norm, and these changes to marital status may have had an effect on the intensity of cultural practices. Married life leads to arrangements between household members around cultural practices (Singly 2003), although it does not seem on the face of it to affect the regularity of reading (Pharabod 2007). Here we find the idea that free time for cultural activities, and leisure activities more generally, depends on the division of social time including that devoted to domestic activities (Chenu and Herpin 2002). We know, moreover, that living in individual houses has increased in the second half of the twentieth century (Jacquot 2003). A house, which is often more spacious than an apartment, means more intensive use of domestic space in which leisure activities are concentrated, while collective living corresponds to a greater propensity to going out (Chenu and Herpin 2002; Herpin and Verger 2008). We thus need to determine what is the case for teachers.
A rise in the social status of recruits
22While the social backgrounds of those with baccalauréats and university degrees have on the whole diversified in the second half of the twentieth century with access to secondary and higher education for students from working-class backgrounds, those of teachers are at the top and middle of the social scale.
23Among successive birth cohorts, more candidates who are the children of managers, those in liberal and intermediate occupations, or of teachers have been recruited at primary level. There has also been a gross reduction in the proportion of children of manual workers, agricultural workers, craftspeople or shopkeepers (Farges 2011). This can be explained by a mechanical effect linked to changes to the structure of the active population (Vallet and Degenne 2000), but not only by this. At primary school level, the grades needed in terms of university education have risen in recent years and this works against those from backgrounds that are economically and culturally disadvantaged (Charles and Cibois 2010). Additionally, the more difficult conditions for entry into the labour market has made young graduates from advantaged backgrounds more interested in the competitive examinations to become teachers as an alternative to the private sector (Peugny 2009; Fougère and Pouget 2003; Fougère, Lixi and Pouget 2004).
24In secondary education, the gross changes show that the proportions of children of manual workers and employees has been quite stable across several cohorts of teachers, while the proportion of children of managers and liberal professions has declined slightly (since those born in the 1960s), and this is even more clearly the case for the children of the self-employed. These observations are similar to those of C. Thélot (1994), who pointed out the reduced attraction of teaching among those from the most advantaged backgrounds. At the same time the proportion of children of teachers (“teachers in secondary and tertiary education, scientific occupations” and “primary school teachers and comparable occupations”) and intermediate occupations has risen (Farges 2011).
25These social dynamics mean that the acquisition of legitimate cultural knowledge happens more in the family and less in school for a large proportion of teachers, challenging the singularity of a learned culture (Bourdieu  1984). At the same time, changes in social backgrounds imply a greater proportion of children of teachers, which might favour family socialization over school socialization (Tavan 2003). Yet, to leave it at this would be to ignore the fact that from one era to another, parents of teachers could have transmitted other cultural norms and values to their children, promoting more eclectic practices and tastes (Peterson and Kern 1996). It would also be to ignore the fact that peer groups and the media have broadened the cultural knowledge transmitted within family contexts (Lahire 2004, 2001). Thus, the effect of social background on the cultural practices of teachers deserves to be examined from a long-term perspective.
Are practices gendered?
26The differences between male and female teachers in terms of their cultural activities can also be examined. The feminization of teaching is a longstanding phenomenon in primary schooling. In secondary schooling, women became more numerous than men from the middle of the 1950s, although significant differences in status and subjects taught are observable and have continued over time (Cacouault-Bitaud 2007). Feminization, which has continued to rise slowly,  tends towards preserving intense cultural practices. Indeed, generally, we know that cultural practices, in particular the more legitimate, are more the preserve of women than of men, in line with different artistic socialization during childhood (Christin 2011). We should therefore check whether these gender-specific differences are perceptible among teachers as we might assume (Bihagen and Katz-Gerro 2000), and whether the gender effect varies over time.
27Before considering the possible effects on legitimate culture consumption of these sociological changes specific to teachers, we should understand how their cultural practices have varied over time, and to what extent these variations are comparable with those experienced by other categories of graduates.
The stability of the cultural practices of “teachers in secondary and tertiary education, scientific occupations” in comparison with other categories of graduates
28Drawing on the “Pratiques culturelles des Français” surveys from the Ministry of Culture and Communication, I look at cinema and theatre attendance and regular reading. These surveys have been conducted at least once a decade since 1973, but some editions do not allow us to work on categories that identify teachers. I have used the 1981 and 2008 surveys, which, based on the criteria of the classifications used, are the points in time furthest away from one another. For 1981, I selected categories 32 (“Teachers in secondary and tertiary education, scientific occupations” and 41 (“Primary school teachers, various intellectual occupations”),  for 2008, categories 34 (“Teachers in secondary and tertiary education, scientific occupations”) and 42 (“Primary school teachers and comparable occupations”).  These surveys thus provide access to a teaching population whose boundaries are unfortunately imprecise and that has changed between the two dates, but they allow us to compare the cultural practices of this population with other graduate categories. It should be noted from the outset that the 1981–2008 comparison only provides an incomplete overview of changes, since it does not consider populations at the same point in their life cycles but instead all ages are mixed together. However, this approach has been retained because of the low numbers for the different categories. 
New differences between “primary school teachers” and “teachers in secondary and tertiary education”
29While the levels of theatre attendance and regular reading by “primary school teachers” and “teachers in secondary and tertiary education” are comparable in 1981, the differences are much more pronounced in 2008 and the percentages are much higher for the latter than the former at this date (Figures 1 and 3). Regular cinema attendance has declined much more for “primary school teachers” than for “teachers in secondary and tertiary education” (Figure 2), such that the number of at least monthly cinemagoers among the former was proportionately greater than the latter in the 1981 survey, but lower in the 2008 one.
Have been to the theatre in the last 12 months
Have been to the theatre in the last 12 monthsField: Working people aged 15 or over (categories 91–99 from the CSP [socio-occupational categories] code were excluded from the 1981 survey, categories 71–78 from the PCS [occupations and socio-occupational categories] code from the 2008 survey).
Interpretation: In 1981, 27% of senior executives claim to have been to the theatre during the last 12 months (vertical bars represent confidence intervals).
Note 1: Concerning qualifications, the questions posed differ between the two surveys “Did you obtain any of the following qualifications at the end of your studies?” in 1981, “What is the highest qualification you obtained?” in 2008. For 1981, I considered qualifications noted as “other” to be equivalent to the level of the baccalauréat.
Note 2: In the 1981 questionnaires, the question was posed thus: “And since 1980, have you been to a drama or comedy theatre to see a performance by professional actors?” In the 2008 questionnaire: “Still from this list, which of these have you done in the last 12 months?” Response item: Go to the THEATRE to see a performance by professional actors.”
Have been to the cinema in the last 12 months
Have been to the cinema in the last 12 monthsField: Working people aged 15 or over (categories 91–99 from the CSP code were excluded from the 1981 survey, categories 71–78 from the PCS code from the 2008 survey).
Have read 20 books or more in the last 12 months
Have read 20 books or more in the last 12 monthsField: Working people aged 15 or over (categories 91–99 from the CSP code were excluded from the 1981 survey, categories 71–78 from the PCS code from the 2008 survey).
Note: In the 1981 survey, the question was posed thus: “About how many books have your read in the last year taking into account your holiday reading?” In the 2008 questionnaire: “In the last 12 months, roughly how many books have you read, taking into account your holiday reading?” For the 2008 survey, interviewers were asked to exclude reading for work and books read to children.
30The changes to the composition of the categories between 1981 and 2008 could in part explain the differences observed. In addition, professional reading was explicitly excluded in 2008 (advice having been given to interviewers). The omission of professional reading could help explain the lower number of books read by “primary school teachers” between 1981 and 2008, especially considering that the distinction between professional and leisure reading makes more sense for them than for “teachers in secondary and tertiary education,” in particular when the latter teach humanities (Bidou 1984; Lantheaume and Hélou 2008).
31Nevertheless, these differences can also be interpreted sociologically. Indeed, given that the school institution values legitimate cultural practices, the differences between “primary school teachers” and “teachers in secondary and tertiary education” could relate to the relationships they each have to that institution over a long period. According to François Dubet (2002), primary school teachers have accepted the changes to the role of schooling and dissociate themselves more from the institution of school. Moreover, intellectual motivations and a taste for a subject field habitually strongly contribute to the career choices of secondary school teachers (Rayou and van Zanten 2004), while access to primary school teaching presupposes instead a desire to teach or to care for children in a less academic way (Périer 2001; Farges 2010; Geay 2010).
The consistency of cultural excursions and regular reading among “teachers in secondary and tertiary education”
32Although frequent cinema attendance (Figure 2) is down among those with baccalauréats and higher education degrees and intermediate occupations, it has declined little between 1981 and 2008 for senior executives, liberal professions and for “teachers in secondary and tertiary education,” who remain faithful to a cultural practice that has become more of an occasional activity overall, presumably because of home audiovisual equipment and the growth in home film watching (Donnat 2011a). In contrast, it is notable that theatre attendance has risen significantly between the two dates among those with baccalauréats and higher education degrees as well as among senior executives and liberal professions (Figure 1). Yet, theatre attendance seems to have fallen among “primary school teachers” and risen slightly for “teachers in secondary and tertiary education,” a finding which could not be confirmed because of the confidence intervals. The difference between “executives” and “teachers in secondary and tertiary education” was reduced in 2008 in comparison to 1981, so that the confidence intervals clearly overlap. I mentioned above the reasons for increasing attendance at entertainment venues in the recent period, and it is surprising that theatre attendance by “teachers in secondary and tertiary education” does not follow a dynamic that is comparable to that of senior executives and liberal professions, although it would be useful to compare the variations for occasional and frequent attendance. 
33The differences in percentages of regular readers between 1981 and 2008 (Figure 3) do not move in the same direction as those observed for theatre attendance. Not only has the gap between “teachers in secondary and tertiary education” and “primary school teachers” widened, but the differences between them and senior executives and liberal professions are also much greater, as are the differences more generally between them and those with a baccalauréat and higher education graduates. Among “teachers in secondary and tertiary education,” the proportion of regular readers has remained stable, while it has declined significantly among other categories. Undoubtedly, as suggested above, the occupations grouped together in the “teachers in secondary and tertiary education” category figure among those that use more books professionally and do not always distinguish this from leisure use. It is also possible to see in these results the continued value that books represent for these occupations, corroborating the analysis proposed by Bourdieu in terms of cultural “asceticism” (1979, p. 325). The information, arts and entertainment occupations socio-occupational category also includes within it a significant proportion of regular readers for the recent period (Table 1), but this proportion also tends to be lower than that of “teachers in secondary and tertiary education,” a finding which could probably be confirmed if there were more respondents. I am therefore inclined to conclude that regular reading is becoming a practice specific to the occupations grouped within the “teachers in secondary and tertiary education, scientific occupations” category, which have stronger links to the academic milieu and the book culture this continues to propagate.
Have read 20 books or more in the last 12 months, breakdown for “Senior executives and intellectual occupations” in 2008
Have read 20 books or more in the last 12 months, breakdown for “Senior executives and intellectual occupations” in 2008Field: Working people aged 15 and over.
Interpretation: In 2008, 55.8% of “Teachers in secondary and tertiary education, scientific occupations” claim to have read 20 books or more in the last 12 months.
34While the category that includes primary school teachers is characterized by a decline in the rate of practice, including a decline in attendance at entertainment venues in contrast to other graduate categories, that which includes secondary school teachers remains apart from the changes to legitimate cultural practices. However, the results obtained using the “Pratiques culturelles” surveys are limited in many ways. On the one hand, they are based on aggregated categories, grouping together diverse occupations, such that alongside primary and secondary school teachers are occupations with varied cultural and economic resources, whose effect on the measurement of cultural practices cannot be neutral. On the other hand, the comparison allows us to consider changes in legitimate culture consumption and to situate teachers in this context, but the numbers involved are low, restricting the possibilities and scope of the analysis in terms of both statistical and heuristic approaches. The next section aims to profit from an original comparative method and to conduct further analysis on teachers in collèges and lycées.
The decline in legitimate culture consumption
35In 1970, J.-M. Chapoulie and D. Merllié conducted a questionnaire survey of secondary school teachers in France. The questionnaire combined several themes, so that data was produced both on work and social status as well as elements of lifestyle, in particular cultural activities. In a new survey, conducted in 2008, I sought to make a comparison with the former by repeating some of the questions (Box 1). The comparison of these two surveys presents a number of difficulties inherent to the use of data produced decades apart (Donnat 2011b). One particular limitation of this method relates however to the fact that the two surveys do not involve the same population selection criteria nor the same methods of implementation or administration. This explains, in particular, the very different breakdowns by subject, which constitutes a significant bias for the objectives of this article: the 1970 survey concentrated on the teaching of theoretical subjects, and few teachers from technical subjects were interviewed (Chapoulie and Merllié 1971; Chapoulie 1987).  Without ignoring them, I strive to overcome these construction limitations in my statistical approach.
Box 1.—The 1970–2008 comparative method
J.-M. Chapoulie and D. Merllié’s 1969–70 survey on secondary school teachers (Chapoulie and Merllié 1971; Chapoulie 1987) was conducted under contract to the Ministry of National Education (3,000 secondary teachers were interviewed), based on lists of teachers from a comprehensive list of lycées and collèges d’enseignement secondaire (CES) in mainland France. The sampling principle meant a greater probability of featuring newer institutions, CESs and large institutions. A random selection of 1000 questionnaires from this survey was subject to manual data entry in 2007 according to the original coding scheme. The data were weighted according to information provided by the authors on the characteristics of National Education between 1966–7 (according to grade—agrégés, certifiés, maîtres rectoraux— and sex).
The 2008 “Enseignants” questionnaire was designed so that it could be compared, as far as possible, with the 1970 survey, taking into account length and layout constraints. In the two surveys, teachers aged 59 and younger were selected. The age limit of 59 years was used as an indicator prior to the survey to distinguish between active and retired teachers in the construction of samples of teacher contacts provided by the MAIF. This age limit was retained thereafter and extended to the 1970 survey to ensure consistency of analysis. In the 2008 “Enseignants” survey, only secondary school teachers have been selected here. These population selection processes result in totals of 894 secondary school teachers in 2008 and 871 in 1970.
36The analysis that follows focuses on four practices: number of theatre visits per year, number of cinema visits per month, the quantity of books read in the last three months and the number of books bought per year on average. It should be noted that the questions asked differ between these surveys and the “Pratiques culturelles” surveys. Since there is no direct comparability, and the contrasts between the two sources are only reinforced by the fact that the survey years differ, it is important to begin by observing the variations in intensity of cultural practice using this second empirical tool.
Less intense practices in 2008 than in 1970
37Generally, the intensity of legitimate cultural practices by secondary school teachers is significantly lower in the 2008 survey than in the 1970 one (Table 2). In 2008, fewer teachers regularly visit the cinema and the theatre; in contrast occasional visits have increased, as has “non-attendance.” With respect to reading and book purchases, we note that there are fewer “non-readers”  in 2008 than in 1970. However, it is clear that the proportion of heavy readers has fallen in favour of more moderate options between the two dates.
The intensity of secondary school teachers’ cultural practices
The intensity of secondary school teachers’ cultural practicesField: Working secondary school teachers aged 59 or younger, weighted numbers, author’s calculations.
1970: N = 871, 2008: N = 894.
38Given the design characteristics of the 1970–2008 comparison, we might think that these contrasts are linked, at least in part, to the different methods of collecting data, and in particular to the differences in the representation of humanities teachers in each survey. However, we cannot exclude the possibility that they are associated with a change in teachers’ legitimate culture consumption, this change possibly being linked, completely or otherwise, to the changing social composition of the populations during the period studied. We are able to control the effect on the intensity of practices of several variables that potentially explain the changes (Table 3), according to the hypotheses made in the first section. Although the representativeness of the two surveys is not perfect from the point of view of socio-demographic characteristics, as a comparison with several public statistics surveys shows (Appendix, Table A7), it nevertheless enables us to pursue our analysis.
Explanatory variables, 1970–2008
Explanatory variables, 1970–2008Field: Active secondary school teachers aged 59 or younger, weighted numbers, author’s calculations.
1970: N = 871, 2008: N = 894.
39We will observe the effect of subject taught, the variable whose individuality I have highlighted, linked to the sampling criteria of the two surveys. We can also consider the effect of age. The age structures in the two surveys are contrasting, relating to secondary school teaching recruitment policies since 1970. Next, the effect of the father’s occupation can be analysed. Between the two surveys, the proportion of teachers with a father who is a senior executive, has a liberal or intellectual occupation or is a (primary or secondary) teacher has risen while the proportion of teachers whose fathers are agricultural workers, employees or manual workers has not declined. The differences in the proportion of “non-working” fathers or “non-responses” probably relates, in whole or in part, to the differences in the form of the questions posed, since father’s occupation was an open question in the 1970 survey but a closed one in 2008. We can also consider the effects of gender on the intensity of cultural practices. The proportion of women among collège and lycée teachers is greater in 2008 than in 1970, but the gap is not however that marked. Variables also allow us to control elements of lifestyle. Our comparison method reveals that more teachers in 2008 have partners than those in 1970. Additionally, in 2008, more teachers live in their own individual houses than in 1970,  which can be related to the changes in places of residence of this population. 
40With the objective of verifying that the finding of a decline is not the result of sampling characteristics and whether the decline is confirmed, considering which sociological interpretations are possible, I will seek to answer three questions. Firstly, which population composition characteristics have an effect on legitimate culture consumption that is likely to explain (in a statistical rather than causal sense) the changes to this between the two dates? Secondly, of these characteristics, which contribute most to reducing the effect of survey year, that is, best explain the decline in the rate of practice? Finally, are we able to fully explain the differences between 1970 and 2008?
The partial explanation of differences in terms of the population’s changing characteristics
41Logistic regressions were performed on dichotomous variables representing practices done frequently or rarely:  going to the theatre at least twice a year, the cinema at least once a month, buying more than six books on average per year, having read six books or more in the last three months. This method enables us to estimate the effects of the different explanatory variables on the variable we are seeking to explain. The results are expressed as a function of a reference category (Ref.), against which coefficients are interpreted.
42The first set of models (Table 4) is the logistic regression of Table 2 that crossreferences survey year with cultural practices; the same information is expressed herein in terms of odds ratios: the “odds” of frequent cultural practices are lower in 2008 than in 1970, as shown by the negative sign and the significance of B coefficients associated with the “year” parameter.
Logistic regressions, models 1
Logistic regressions, models 1Field: Active secondary school teachers aged 59 or younger, weighted numbers, author’s calculations.
Ref.: Reference category
*** P ≤ 0.01; ** P ≤ 0.05; * P ≤ 0.1.
43In a second stage (Table 5), we can observe the changes to the “year” parameter following the introduction of explanatory variables. In order to identify which elements of the population composition contribute most to reducing the effect of survey year, variables were introduced one by one.
Logistic regressions, models 2
Logistic regressions, models 2Field: Active secondary school teachers aged 59 or younger, weighted numbers, author’s calculations.
Ref.: Reference category. *** P ≤ 0.01: ** P ≤ 0.05: * P ≤ 0.1.
The effect of subject taught and sampling characteristics
44Teaching a subject in the humanities rather than a scientific subject increases the probability of engaging in cultural practices more intensely, all other things being equal. This effect, linked to humanities subjects, is very strong, which suggests that the differences in composition of our two surveys by subject explains a significant portion of the differences in levels of practice, and that the contrasts in Table 2 are more pronounced than if both samples had a structure that is more representative of subjects than that observed in the teacher populations of the two years. The effect of subject taught is particularly great in the case of reading practices since the introduction of this variable eliminated the significance of the “year” parameter. However, other composition characteristics have effects on practice intensity.
The effects of changes in social composition
45Since the advantage of the statistical method used is that it controls the effects of variables, we do not have to limit ourselves to subject taught to explain the differences in levels of practice. Thus, type of housing contributes to reducing the “date” parameter for number of books read. Living in an apartment rather than a house favours reading, yet more teachers live in an individual house in 2008 than in 1970. The same is true with respect to having a partner. Conversely, the proportion of teachers aged over 45 is greater in 2008 than in 1970, which appears to be mostly favourable to reading practices, since the 34-years-and-under are less likely than the former to have read at least six books in the last three months. Similarly, the rise in percentages between 1970 and 2008 of teachers having a father who is a teacher limits the fall in number of books read, since having a father who is a teacher has a positive effect itself on reading volume, reflecting a continuity between school norms and educational norms conveyed by teacher parents.
46Regarding cinema and theatre going and book purchasing, the effect associated with the “date” parameter diminished from the first to the second stage but remains significant. Beyond the effect associated with subject taught, the other explanatory variables introduced also have effects themselves on these practices. Firstly, living in an apartment favours them as it does reading. Concerning the theatre and cinema, being a teacher aged over 45 years implies a greater likelihood of going relatively frequently (particularly to the theatre), as does being a woman (similarly, above all to the theatre), while having a partner implies the opposite. On theatre visits, a specific effect of social background is evident. Coming from an advantaged social background (father with a liberal profession or a senior executive) increases, ceteris paribus, the probability of going quite frequently (conversely, having a father whose occupation is not mentioned diminishes this probability), which reflects the unequal distribution of cultural capital depending on social trajectories and indicates the position of theatre visits in the cultural practices hierarchy. Regarding the number of books bought, fewer of the explanatory variables have a significant effect, but the effect of another social origin indicator variable should be noted for this practice, which reveals that book purchasing is a practice that is sociologically distinct from reading itself. Although there is a “teacher father” effect evident for reading, book buying is more common among teachers with a father who is in an intermediate occupation, a craftsman or a shopkeeper rather than an agricultural worker, employee or manual worker.
47For these three practices, the introduction into the models of social characteristics favourable to more intense cultural practices but the proportions of which have declined in the population (teaching the humanities, living in an apartment, having a partner) leads to a decline in the “year” effect, that is to say, helps to explain the observed differences between 1970 and 2008 in Table 2. However, the “year” effect does not disappear, which means that temporal variations in rates of cultural practice are not reducible to the changing composition of the teacher populations between the two surveys.
Changes over time to the effects of sociodemographic characteristics on legitimate culture consumption
48Models 2 present coefficients that do not take temporality into account. However, it is very probable that the effect of explanatory variables on cultural practices will not be the same at the two dates, which could help explain the fall in rates of practice. To estimate these changes, at the third stage of analysis, interactions between the explanatory variables and the survey year were added to the models.
Interpretation of interaction effects
49Interpretation of the effect of the explanatory variables on the variable to be explained is made more complicated because of interaction. We can no longer interpret the net effect of variables “in general” (as in models 2) since these need to be specified depending on survey year. In Table 6, in which the models constructed with an omitted category act as a reference category, the B coefficients for the variables with no interaction are read with 1970 as the reference year. The interaction terms indicate that there has been a change in the effect of the variable between 1970 and 2008; this change may be significant or not. Added to the B coefficients without interaction, the interaction terms enable the coefficients relating to 2008 to be calculated. To estimate the significance of the latter, I constructed the same models, the results of which are presented in Table A8 (Appendix), using 2008 as the reference year. I will refer below to both Table 6 and A8 and comment only on the significant changes, that is those that reflect changes whose validity can be extended beyond the survey samples alone. 
Logistic regressions, models 3 (1970 as reference year)
Logistic regressions, models 3 (1970 as reference year)Field: Active secondary school teachers aged 59 or younger, weighted numbers, author’s calculations.
Ref.: Reference category. *** P ≤ 0,01: ** P ≤ 0,05: * P ≤ 0,1.
50An initial result emerges: the introduction of interactions removes any significance for survey year for cinema and theatre visits and book purchases. Following on from models 2, this third stage thus enables us to move forward in the explanation of the observed decline. While the stable effect over time of subject taught should be emphasized (the coefficients being significant and positive for both 1970 and 2008 without any change being found), models 3 above all indicate that the intensity of legitimate cultural practices varies more under the influence of certain demographic characteristics in the recent period than in 1970, and, moreover, that some effects were reversed. The discovery, for 2008 only, of the lower likelihood that young teachers have consumed legitimate culture quite intensively, is revealed as the second significant result of this third stage.
The striking effect of interactions between age and survey year
51In the case of theatre trips, variations in the effects of several sociodemographic characteristics over time can be observed, such that the “odds” associated with this type of excursion are more a function of housing type, social origins and age in 2008 than in 1970.  Thus, in 1970, type of housing has no particular effect on this type of excursion, while in 2008, when more teachers live in houses, the fact of living in an apartment increases the probability of going to the theatre significantly. Similarly, whereas in 1970 the gap between teachers with working-class social backgrounds(father who is an agricultural worker, employee or manual worker) and those with a middle-class social background (intermediate occupation, craftsman, shopkeeper) is not significant, it becomes so in 2008, the interaction term clearly indicating the direction of change between the two dates. This social background effect particular to the recent period is an echo of the “generally” positive effect on theatre trips of fathers who are senior executives or in liberal professions, evident in models 2. Teachers’ theatre trips appear to be more dependent on social trajectories in 2008 than in 1970. This could indicate that familial socialization to this rather “sophisticated” cultural practice tends to be substituted for a more collective socialization (by the school institution, peer groups, work environment), having the effect of distancing sections of teachers from these practices, in a context in which cultural offerings have diversified and turned increasingly towards audiovisual and digital consumption. In addition, within middle and upper-class families, parental education regarding elitist cultural practices is probably more pronounced than in the past because of the central role played by cultural resources—in a period of rising education levels (Duru-Bellat 2006)—in the acquisition of social status (van Zanten 2009), which can reinforce the differences between teachers in terms of their legitimate culture consumption.
52Moreover, models 3 show a varying age effect depending on period, which is true for theatre trips and also for cinema trips. For both these practices, the interaction between age and survey year is particularly decisive since its introduction to the models removes the significance of the “year” parameter. In 1970, there is no clear effect of age on theatre trips, which suggests that the youngest teachers have no greater or lesser “odds” of going quite often than the oldest teachers. However, in 2008, teachers in the middle age group (35–45 years) and young teachers (34 years and under) are less likely to go to the theatre quite often in comparison to those over the age of 45. The gap between those aged 34 years and under and the over-45s is particularly large. Regarding cinema trips, in 1970, teachers aged under 45, and among them those aged 34 and under in particular, had greater “odds” than older teachers of going to the cinema at least once a month. However, in 2008, this finding is reversed.
53This significant effect of interaction between age and survey year is also evident regarding the purchase and reading of books. In 1970, none of the three age categories had greater “odds” of reading than the others. At this time, the youngest were more likely to buy more than six books on average per year. However, the book-related practices of the youngest teachers have changed, since in 2008 they have significantly lower “odds” of reading or buying books than the oldest teachers. Note that in the case of buying books, the interaction between sex and survey year should also be taken into account. While men in 1970 were more likely than women to buy more than six books on average per year, in 2008 this result is reversed and it is women who have greater “odds” of being book consumers. 
Less intense legitimate cultural practices among young teachers
54Thanks to models 2, we have established, for three of the four practices, that the decline applies to all age categories (which was shown by the persistence of the “year” effect). With models 3, it can also be argued that the age effect has changed over time for the four practices: the finding, with other characteristics controlled, of lower legitimate culture consumption by young teachers in comparison to older teachers is true for 2008, but not for 1970. One may wonder which social or occupational characteristics, hidden behind the “age” variable, distinguish young teachers from older ones in the recent period and lead to a decline in their cultural practices. Some can be envisaged, relating to the ways in which they were socialized to culture as well as their occupational socialization.
55Firstly, the interpretations provided to explain the decline in legitimate cultural practices by graduate categories could also be true for young teachers. The declared intensity of practices may have varied as a result of the way in which they have become everyday activities (Mauger 1992). Like their peers, young teachers were educated in schools that had become unitary, where they were socialized in a variety of cultural practices and tastes in a context of expanding choice. At university they were in contact with students forging different careers with possibly distant relationships to “cultivated” culture, access to universities having been widened over the past decades (Merle 2009).
56Secondly, teachers’ relationship to the profession should be emphasized. Since the beginning of their careers, young teachers bear witness to a dissociation with the school institution and the different collective dimensions of their profession, showing more individualistic behaviour that can be explained sociologically. Indeed, according to Bertrand Geay, “with relatively elevated social backgrounds, these young people thus join the profession at the end of educational trajectories marked by relative success but also increasingly through the need to enter the labour market, without downgrading their education and with a certain cultural proximity to the world of teaching” (2010, p. 75). One may therefore conclude that the new forms of socialization to the teaching profession, although more visible in primary school teaching, reinforce the distance that separates secondary school teachers from the most academic professional identities, and thus the more traditional forms of cultural legitimacy.
57In fact, these results, are in line with work that has shown the gradual shift in professional identities during the twentieth century, from the mastery of knowledge in the “schoolmaster” model in favour of teaching based on a “pedagogical” figure (Lang 1999; Maroy 2006). A clearer separation between social times and spaces is also observed, in so far as, in the private sphere, more time is set aside for the family than in the past (Farges 2010; Geay 2010), limiting the “hold” of the profession (Lantheaume and Hélou 2008, p. 73). It is thus understandable, for example, that young teachers make a firmer distinction between professional reading and leisure reading than their elders, leading to an even greater reduction in the stated amount of reading.
58Moreover, the teaching profession has not just changed in terms of its conception; working conditions also raise new issues. In terms of time resources, we should remember that young teachers, in particular when they are appointed to posts far from where they want to be, are less able to exploit the experience accumulated with age. Faced with the increasing intensity and complexity of teaching (Maroy 2006), some adopt “survival strategies” (Rayou and van Zanten 2004; Rayou 2009), with the likely corollary being a reduction of time and availability outside work, and dissociation from cultural activities requiring the dedication of time and/or travel (Lantheaume and Hélou 2008; Chenu 2002; Bozon 2009). As such, the decline between 1970 and 2008 in “opportunities” to go relatively often to entertainment venues for the 35–45 years category, as observed above, may reflect a marked reduction in time available for teachers of this age group, who often have family commitments.
59Finally, we might consider material resources, in particular when interpreting the gap between the youngest and oldest teachers from the point of view of theatre trips, a gap which is wider than for other cultural practices. It has been shown that the net salaries of teachers recruited in the 2000s are less generous than those of teachers recruited at the beginning of the 1980s (Bouzidi, Jaaidane and Gary-Bobo 2007), while young teachers, as are all young workers, are more constrained than in the past by their living expenses, particularly by expenditure relating to housing (Accardo and Bugeja 2009). In sum, since the prices of cultural products have not dropped between 1970 and 2000 (Herpin and Verger 2008), the most expensive cultural practices have become less accessible.
60* * *
61In the absence of survey material which would enable us to analyse them, variations in teachers’ legitimate culture consumption have been little studied, although a transformation in cultural practices, primarily the decline in book reading, has been shown among graduate categories, and the cultural knowledge transmitted by teachers to their students is greatly challenged, at the current time, by audiovisual consumption. In the first stage of my empirical approach, fairly stable legitimate culture consumption by “teachers in secondary and tertiary education, scientific occupations” in comparison to other graduate categories emerged from the results of the “Pratiques culturelles” surveys. These results have been supplemented and qualified with the help of two surveys of secondary school teachers as a second stage of analysis. Separated from other intellectual and scientific occupations and taking into account their sociodemographic characteristics, teachers in collèges and lycées recorded a decline in reading practices and frequency of cultural excursions between the 1970s and 2000s, thus appearing to be unreliable guarantors of traditional forms of cultural legitimacy as understood here through four types of practice.
62This analysis, in turn, calls for a study of the dynamics that may be observable for other cultural practices, which my material does not allow me to offer. Thus, it would be desirable to extend the analysis to other cultural forms, such as television consumption, use of new technologies and musical practices, to assess, in particular, the distance that separates the cultural repertoires of teachers from those of their students, to study the relationships between different cultural forms and understand to what extent teachers compromise between different styles of practice and legitimacy. An analysis of tastes over time—more difficult to carry out—would be useful to complement that of the transformation of practices.
63Finally, an analysis of the way in which cultural materials are used in the classroom is warranted for the recent period (similar to the study done by Lise Demailly 1985, for example). In line with the characteristics of the relationship between young teachers, the teaching profession and the school institution, it is entirely possible to imagine a duality in their relationship to culture, taking a sort of middle ground (Geay 2010) combining a proximity to norms and values characteristic of the setting and a distance, particularly visible outside of the school aspect of teachers’identities. Out of respect for the school institution and its operation, teachers might promote legitimate cultural practices in class that they themselves rarely subscribe to outside of their time dedicated to schoolwork, indicating a dissociation between their practices in class and private practices.
1970–2008 explanatory variables and sample representativeness
1970–2008 explanatory variables and sample representativenessField: Working secondary school teachers aged 59 or younger, weighted numbers, author’s calculations. 1970: N = 871, 2008: N = 894.
Public statistical surveys for 1970
64The distribution by subject type was calculated based on numbers provided in the 1971 edition of La Tableaux de l’éducation nationale (Ministère de l’Éducation Nationale, Service Central des Statistiques and de la Conjoncture, pp. 60–1). These are full-time secondary school teaching personnel in the public sector for the year 1969–70. Personnel in general teaching and theoretical and practical technical teaching were included.
65The 1970 “Emploi” survey (INSEE) was used for the partner, housing, sex and age variables.
66Field: Active secondary school teachers aged 59 and under (the codes selected from the job codes were 9095 “Professeur d’enseignement du second degré,” 9073 “Adjoint d’enseignement,” 9079 “Professeur s.a.i,” 9092 “Professeur technique de collège d’enseignement technique and assimilés,” 9096 “Professeur technique de lycée technique and d’école normale d’enseignement technique”).
67N = 749. Numbers weighted with the help of the “extr2” variable, author’s calculations. The indicator used for having a partner is legal matrimonial status.
68The source used for father’s occupation was the 1970 “FQP” survey (INSEE).
69Field: See above (N = 447). Numbers weighted using the “ponder” variable, author’s calculations. Teacher fathers combine the INSEE socio-occupational categories 32 and 41.
Public statistical surveys for 2008
70The distribution by subject type was calculated based on numbers provided in the 2008 edition of Repères et références statistiques sur les enseignements, la formation et la recherche (Ministère de l’Éducation Nationale and Ministère de l’Enseignement Supérieur et de la Recherche, p. 293). Public sector secondary school teachers were taken into account.
71The source used for the father’s occupation, partner, sex and age variables is the 2008 “Emploi” survey (INSEE).
72Field: Active secondary school teachers aged 59 and under (the codes selected are 341a “Professeurs agrégés and certifiés du secondaire,” 422a “PEGC,” 422b “Professeurs de lycée professionnel,” 422c “Maîtres auxiliaires du secondaires”).
73N = 2,787. Weighted numbers using the “extri09” variable, author’s calculations. The indicator of having a partner is constructed by combining legal marital status and life as a couple.
74For the housing variable: the source used is the 2008 census (INSEE, file detailing individuals located in the region).
75Field: See above (N = 166,789). Numbers weighted using the “ipondi” variable, author’s calculations. Teacher fathers combines the INSEE socio-occupational categories 34 and 42.
Notes on the differences observed
76The differences in distribution by subject type between the 2008 “Enseignants” survey and the public data stem in part from the coding of language teachers. The 2008 questionnaire offered “English,” “German,” and “Spanish” responses but teachers of other modern languages had to answer “other.”
77The two surveys on teachers overestimate the proportion with higher and teaching social backgrounds, while they play down working-class social backgrounds. The sampling characteristics do not seem to be the only possible explanation for these differences. Some of the differences between the 1970 secondary school teachers survey and the 1970 “FQP” survey, could stem from the question formulation (in the former, the occupation of the father given is that which was current at the time of the survey, in the latter the father’s occupation given is the one he had when the interviewee finished his/her studies) and the way in which the questionnaires were administered (self-completed questionnaire versus face to face interviews in the “FQP” survey). The same limitations are true for the 2008 surveys: the form of the questions can also explain some of the differences for this (closed question in the “Enseignants” survey versus open question in the “Emploi” survey).
78There are large differences in the proportion of individuals in couples between the surveys on teachers and the public statistical surveys, although in all cases the teachers in couples are very much in the majority. In the 1970 survey on teachers, as in the 2008 one, the interviewees themselves gave the information on their partners. In the 1970 “Emploi” survey, having a partner in one’s life or not was seen as only possible though legal matrimonial status. In the 2008 “Emploi” survey, having a partner or not was estimated by constructing an indicator combining life in a couple at home and legal marital status. These differences in the nature of the indicators could explain the observed differences.
Logistic regressions, models 3 (2008 as reference year)
Logistic regressions, models 3 (2008 as reference year)Field: Working secondary school teachers aged 59 or younger, weighted numbers, author’s calculations. Ref.: Reference category. *** P ≤ 0.01: ** P ≤ 0.05: * P ≤ 0.1.
I would like to thank the Réseau Quetelet and the Centre Maurice Halbwachs (ADISP) that allowed me to work on the data from these surveys as well as on the data from the “Emploi” [“Labour Force”], “Formation et qualification professionnelle” (FQP) [“Occupational training and qualifications”] surveys, and INSEE censuses, also used in the course of this research.
[Translator’s note] The liberal professions are occupations requiring special training in the arts or sciences, and whose activities are usually closely regulated by national governments or professional bodies. They include, for example, lawyers, notaries, engineers, architects, doctors, dentists, and accountants.
In 1982, the average age of “agrégés, certifies” teachers (“general secondary education teachers”) was 40, rising to 43 in 2009. For those in the “professeurs de lycée professionnel” (“vocational high school teachers”) category the average ages on these dates were 39 and 45; for those in the “PEGC/maîtres auxiliaires” (“junior high school teachers/unqualified teachers”) category they were 37 and 42; and for “instituteurs/professeurs des écoles” (“primary school teachers/schoolteachers) they were 37 and 41 (sources: “Emploi” survey, INSEE, 1982 and 2009 series, active working population, microdata from the Centre Maurice Halbwachs (ADISP), author’s calculations, weighted numbers).
In 2000, women represented 77.8% of primary school teaching personnel (public sector), 81.5% in 2010. They represented 56.7 % secondary school teaching personnel (public sector) in 2000, 57.6 % en 2010 (DEPP 2010). Note that the proportion of women varies greatly depending on teaching status, educational tracks (academic or vocational) and also subject taught.
By using the INSEE “Emploi” survey we can get an idea of which occupations make up these categories, and in what proportions. According to the 1981 “Emploi” survey, “secondary education and extended technical education teachers, and comparable occupations” make up 60% of category 32, “primary education and short course technical education teachers and similar occupations” make up 68% of category 41. Categories 32 and 41 each comprise around twenty different occupations. After “secondary education and extended technical education teachers, and similar occupations” in category 32 come “doctors, surgeons, within and outside of hospitals” who are the most represented, followed by “assistants and higher education teachers and comparable occupations.” After “primary education and short course technical education teachers and comparable occupations,” the “specialist educators, social centre counsellors” represent a significant proportion of category 41.
According to the 2008 “Emploi” survey, by grouping together occupations to be closer to the 1981 classification, the “qualified secondary school teachers” and “heads of secondary education institutions and inspectors” make up 55% of category 34 and “teachers,” “primary school teachers,” “general junior high school teachers,” “technical high school teachers, “teaching assistants and supply teachers in secondary education” represent 61% of category 42. “Higher education teachers,” “public research researchers,” “hospital doctors with no private practices,” represent a significant proportion of those in category 34. This is the case in category 42 of “trainers and continuous education teachers” as well as “supervisors and teaching assistants in educational institutions” and “instructors, sports trainers, professional athletes.”
In the 1981 survey, there were only 73 “teachers in secondary and tertiary education” and 99 “primary school teachers” and 86 and 112 respectively in 2008. The results should therefore be viewed with caution taking into account confidence intervals. In all the analyses that follow I have adopted a significance level of 10%.
This comparison was not made because the numbers of “teachers in secondary and tertiary education” and “primary school teachers” were too small.
Music, drawing, manual work and physical education teachers were excluded from the survey (Chapoulie 1987, p. 355). The authors also make it clear that the sample was not representative in terms of grades and included too large a proportion of agrégés and unqualified teachers as a result of the over-representation of big institutions and collèges d’enseignement secondaire [lower secondary schools that existed in France between 1963 and 1976, educating 11–15-year-olds] (Chapoulie and Merllié 1971).
Here I would like to make a clarification in relation to “non-practitioners,” that is individuals in the “never” or “none” categories. The distinction between the no responses and the responses “zero” or “none” poses a question that is difficult to answer: has the person not responded because he/she has lost interest in the question in favour of another, skipped the question, or does he/she wish to indicate that for him/her the answer is zero because he/she does not practice the activity in question? As the length and layout of the two questionnaires was very different, providing an answer to the previous question is all the more difficult. We might assume that the layout of the 1970 survey, which has more white space, lends itself more to questions not being answered or to being crossed-out to indicate “zero.” Moreover, although the questions relating to reading (books bought and books read) in the two surveys are open questions, the questions on cultural excursions are open in the 1970s survey but closed in the 2008 one. All these survey biases prevent us from distinguishing the true non-responses from “zero” responses. Hence, in the statistical treatment of the two data sources, individuals who have explicitly declared they did not read or had not bought a book, or never went to the cinema or the theatre have been grouped with those who did not respond to the questions. Thus speaking of “non-readers” is misleading since within this group are individuals who have claimed not to read as well as others who for various reasons did not answer the question.
Cross-referencing the “age” and “type of housing” variables shows that, while living in a house is more common in 2008 for all age categories, more of the over-45s and those aged 35– 45 years live in an individual house than those 34 and under in 2008, as was the case in 1970.
Sadly, we do not have a variable indicating the size of commune of residence that is strictly comparable between the two surveys, which is a flaw for our analysis given the influence of this lifestyle element on cultural leisure activities (Coulangeon, Menger and Roharik 2002). The 1970 and 2008 “Emploi” surveys allow us to suggest that the proportion of secondary teachers living in communes with fewer than 10,000 inhabitants has risen from 19% to 31%, to the detriment of those living in medium-sized towns (29% against 18.5%); the proportion of inhabitants of urban areas with more than 100,000 inhabitants remains stable at around 50% (for information on these data refer to Table A7 in the Appendix). The rise in those living outside of medium-sized urban areas could be linked to a decline in legitimate culture consumption. Indeed, according to Donnat (2011a), between 1973 and 2008, the differences in the cultural practices of urban and rural dwellers have persisted, although the rise in attendance at entertainment venues has risen much faster than average between 1970 and 1980 for rural dwellers. The decline in book reading has affected all categories, whatever the place of residence. The author notes furthermore that the Parisian exception has increased in the period studied because of the “elitization” of Parisians in inner-Paris.
For the percentages of individuals in each category refer to Table 2 and remember that more teachers engaged in cultural practices frequently in 1970 than in 2008.
As is shown by a comparison between Tables 6 and A8, several coefficients, insignificant for 1970, become so for 2008. However, this does not mean that the change between the two dates is statistically significant, which only the significance of interaction terms can establish.
The effect associated with teachers’ fathers has changed for the frequency of trips to the cinema: while in 1970 having a father who is an agricultural worker, employee or manual worker, does not increase the “odds” of quite often going to the cinema, the interaction term is positive and significant, indicating that there has indeed been a change between 1970 and 2008. However, while the change is significant, the coefficient for 2008 is not, which means that it cannot be said that teachers who are children of teachers have greater “odds” in 2008 of going to the cinema quite frequently than children of agricultural workers, employees or manual workers. This is why I will not comment on this effect.
It can be noted that contrary to existing knowledge on the sociology of readers, neither models 2 nor models 3 bring to light differences between men and women in terms of book reading.