The network tries to provide rigorous methodology in order to manage this complex international comparative survey. To do so we provide systematic assessment of the development of an international comparative questionnaire; data collection procedure; the process of generating a systematic dataset for analysis and the guidance of network partners in analyzing the data. For any details on the survey and the network see also Brewster et al. (2004). The data collection is coordinated by Cranfield School of Management, UK.
The items within the survey were initially derived from a comprehensive review of the literature on HRM practices. The questionnaire was developed in 1989 based on the literature available at that time and on discussions within a network of academics with expertise in HRM. The survey has been repeated several times since 1989 and has been revised on each occasion, based upon up-dated literature reviews and upon discussions within the research team and with senior practitioners.
For each survey round, the questionnaire is designed by a sub-section of the Network in cooperation with the coordinator at Cranfield School of Management, UK; and is then approved by the rest of the Network. The questionnaire is developed in English and then translated in the national local language where appropriate. Translations of questions are slightly amended between national questionnaires to capture the nuances in meaning between languages. The questions are blind-translated into the national language by a translator with some familiarity with HR and then translated back into English by a different translator. Any differences are discussed in order to capture the universal international comparative meaning of the terms as closely as possible. As much as possible the questionnaire items are held comparable over the years, although insights and developments in HRM have led to a number of changes to the questionnaire. A number of countries developed a panel throughout the several waves.
The survey is focused on factual questions about practices in the organizations and not so much about opinions. This is to be able to capture the more objective elements of the HR architecture and to be able to create a comparable base.
In most cases the survey is conducted using a postal questionnaire. In recent rounds the use of web-based surveys or computer aided interviewing is introduced by several partners in order to increase the response. In most cases the response rates range from 12–25%. In each instance, the respondent was the person with responsibility for human resource management in the business organization. The survey provides questions about the background of the respondent. The unit of response in each country is therefore the ‘organization’: this can mean the firm in its entirety or a relatively self-contained unit within a company, such as a subsidiary. There are questions in the survey that control for the nature of the organization.
The sampling frames used in each country were designed to produce stratified representative samples (by sector and size) and do so in the main for all the countries involved. However, due to slightly different sampling procedures in each country, (descriptive) analysis of the Cranet data cannot claim to provide a representative global overview. In analysis prospective authors will introduce control variables to overcome possible biases.
The data collection is done under the responsibility of each country partner in coordination with Cranfield School of Management and the Network as a whole. The intellectual ownership of the data resides with each country partner. The data is checked and cleaned by Cranfield in order to develop a clean data file that is made available for the partners in the Network. Any processing, analysis and reporting is done by partners under the consent of all partners that are part of the survey round in a particular year.
Monitoring and Guidance in analyzing
The Network meets two times per year to discuss important issues concerning the survey, analysis and prospective publications. We discuss also experiences in validating and analyzing items from the questionnaire. This guarantees a constant monitoring of the quality of the survey, of the analysis and of the comparability of the outcomes and subsequent publications.