• Firstly, and most importantly, come linguistic usages. These are complemented with other variables with which high correlation is maintained. Informal usages are particularly highlighted, such as language usages with peers and in the home, which would be the key factor for ordering the socio-linguistic diversity of pupils. As well as informal usages, a high correlation with three more variables is maintained: self-definition as belonging to a linguistic group, the identification language and the initial language. These are finally completed, less strongly, with representations on linguistic usages, the language of consumption of the communications media and oral usage with teachers, both formal and informal.
  
• The second key point shaping diversity, not so closely related to the first, concerns written usages: the language used in exams, the language of textbooks, classroom materials or compulsory readings. These variables provide a different focus but with less explanatory power. They do not completely correspond with linguistic usages, as these fit into the first factor, and they make it easier to read this socio-linguistic complexity.
  
• Thirdly, there is a much less important factor. This concerns linguistic competences in Catalan, Spanish and English.
  
• Fourth comes the geographical origin of the student and his/her parents.
  
• And a far-off fifth comes competence in English, which, together with the parents' level of education, would determine a small but appreciable differential factor.

3.1. Family language and language with friends

"We've won the school, but have we lost the playground?" (Vila 2004). This has been quite a widespread comment concerning Catalan and Catalan schools, putting the effects of the school on the inter-personal uses of the language in doubt. But different studies have gone into depth extensively on this and have drawn conclusions utterly opposing this argument. At least, in their strict interpretation (Querol 2001; Rosselló 2003; Vila and Vial 2003; Vila 2004; Galindo 2006). They indicate that the institutional use of the language in the classrooms has grown. And, in playgrounds, Catalan is not the language of a tiny minority. Although its presence is lower than Spanish, it is considerable, as the results of the study we have dealt with up to now indicate. And the presence of Catalan in playgrounds is not the consequence of a systematic change of socio-linguistic habits by Catalan-speakers or Spanish-speakers – rather the contrary. Possibly the socio-linguistic dynamics resulting from the great migrations of the '40s and '50s are broadly continuing. In all cases, what is perceived is the huge expansion of spaces of contact between the native population and the old immigrant population and, still further, the dilution of the limits between the two, with the birth of new generations. These changes are interacting with the extensive exposure of the new generations – children of Spanish-speakers – to Catalan in the classrooms.

As the Language with friends section indicates, Catalan-speakers have not abandoned the use of Catalan with their friends and Catalan is not the language of a small minority in interactions between peers (see Table 3). It is quite true that, in interactions between friends, Catalan-speakers use Catalan less than Spanish-speakers do Spanish. And the difference is not a small one: 11.7 percentage points. See Table 5. Clearly, these lower rates of maintenance by Catalan-speakers continue to erode the social use of Catalan. But in no case does it amount to losing the playground.

Table 5. Language usages with friends from school and from outside school, according to the initial language