Methods
Data collection and procedure
The methodology applied was a cross-sectional survey conducted via self-administered questionnaires. The measurements instruments were pretested with five professionals from the sector being studied to ensure respondents would understand the questionnaire correctly. The study target population comprises firms managing hotel establishments in Spain with three or more stars and at least 50 employees. As the second largest tourism destination in Europe and the third largest in the world, Spain received 65 million visitors in 2014 (World Tourism Organization, 2015). Companies running establishments with this rating were chosen because they were better suited for testing the model proposed. Higher rating hotels are run by more professional firms that are concerned about innovation activities as a way to maintain the level of quality necessary to keep their rating (Ordanini and Parasuraman, 2011; Pikkemaat and Peters, 2005). Firms with fewer than 50 employees were excluded from the population to avoid the inclusion of very small organisations with no formal HRM procedures in place (Collins and Smith, 2006; Huselid and Becker, 1996; Lepak and Snell, 2002; Youndt and Snell, 2004). Information on the hotel accommodation firms was taken from the 2011 Hostelmarket Annual Report on hotels, restaurants, tourism and leisure. A population of 523 firms remained after the database had been refined. It was decided that all the sampling units would be studied and the questionnaire was sent to all the firms in the population. Data gathering was conducted from September 2011 to March 2012. Firms in the population were contacted six times during this period, both by email (four times) and post (twice), until an acceptable number of answers was obtained. Of the 112 questionnaires received at the end of the data gathering process, 109 were considered valid. This corresponds to a response rate of 20.84%, indicating a sampling error of 8.4%. The professional and demographic variables contained in the questionnaire show that most respondents held general management positions, had been employed in the firm for more than 10 years, and had received higher education. The Kolmogorov–Smirnov test was used to compare the cumulative frequencies of the number of accommodation units, the number of employees and the sales figures in the population and the sample. The results lead to the conclusion that no significant differences exist between the population and the sample, indicating that the sample accurately represents the population being studied.