1.2. Size of the universe, sample selection and design of the questionnaire

Lists provided by the Department for Tourism, Commerce and Industry of the Generalitat of Catalonia were used to determine the universe of the fuel sector. The sample design was based on a proportional distribution of the list of service stations according to province, district, town and brand, and the stations making up the sample were selected at random; the planned size was 533 stations. This number was deemed adequate for a margin of sampling error of 3% and a confidence interval of 95%. Table 2 below reveals the provincial forecast of the sample:

Table 2. Provincial structure of the sector and sample distribution

Source: Language in service stations in Catalonia. Sociolinguistic contribution. 2000

1.2.1. Creation of the questionnaire

The research team created a protocol or observation questionnaire based on the notion of the routine that we mentioned earlier and by observing the movements involved in this routine.

Firstly, they needed to gather linguistic information from signage and brand advertising at the entrance (see the Campsa and Shell canopy pillar photographs; at Campsa, the linguistic ambiguity of the logo and product names is clear, not to mention the Spanish monolingualism of the opening hours notice. Shell's bilingual solution is also clear).

Secondly, they had to focus on signage on the canopy, pillars and petrol pumps in the service station: names of different products, volume and cost and position of pictograms (three photographs of this area).

The office and cash register area formed the third observation block; this included the presence/absence of Catalan in the shop area and the bar/cafeteria, restaurant and/or hotel (note the sign written exclusively in Catalan alone above the pump, and those of Shop and Night Till also in Catalan alone at BP).

Before leaving the service station, the team also had to observe other services such as water/air pressure machines and whether or not there was a car wash, garage or other services. (This photograph reveals how a sign in a service station such as Cepsa uses bilingualism in its notices regarding sets of measuring instruments, complaints and safety regulations)

The questionnaire was completed by noting down on forms the language or languages used in notices, signs, brands and the names/prices of the products. This observation was made in an average of 20 minutes per area because service station staff did not need to be asked for information. The aim of the second part of the observation was to find out the language used orally by service station staff in their dealings with customers. This part was based on spontaneous oral language used to interact with the observer. The interview ended with a number of questions to staff in charge of the premises and a few observations about the condition of the surroundings of the service station.

Before any fieldwork was undertaken, a number of checks were carried out at a variety of service stations in order to verify and correct the protocol; a series of photographs were also taken and an expert in the sector was consulted to ensure that important elements were not left out of the aims of the investigation. Immediately afterwards, the team of investigators attended training meetings to standardise and clarify the observation criteria.

1.2.2. Sample extraction and fieldwork

The sample was extracted at random by stratifying it according to comarcas (districts). Once the definitive list of service stations to be visited had been drawn up, a map was used to work out a system of routes to ensure that fieldwork was carried out as efficiently as possible. All motorways, dual carriageways, national and district roads, the Eix Transversal and service stations located in urban suburbs shopping centres, co-operatives and marinas were covered (3).

A total of 533 service stations were visited, of which information was obtained from 513, including points of sale in certain shopping centres, co-operatives and ports.

Some incorrect addresses, closed or abandoned establishments and in some cases, staff who refused to answer the questions and/or made observing the establishment difficult, account for the twenty cases in which we did not receive a response.