Description of node <>

Subject Predicate Object Data are controlled and corrected at an individual level during the collection by the team managing the survey: this team contacts by phone or by e-mail the units which presents some inconsistent or implausible data in order to validate or to correct these data. After the end of data collection, some more complex checks are applied. This step, called "data cleaning step", includes some individual correction. Then comes the step of "non-response correction and calibration" which includes the total non-response correction (changing the weight of units in order to grant that the exploitation of the results will be precisely representative of the survey scope) and the partial non-response correction. This step also includes a calibration process of units weights on framing variables (known apart from the survey) in order to enhancing the accuracy of estimations done with the survey results. Eurostat has transmitted to each country carrying out the survey the list of indicators required at the European level, as well as the expected levels of aggregation for these indicators. An example of required indicator is the number of companies which have totally or partially sourced internationally some activities between 2009 and 2011, by activity and by geographical area of international sourcing; This indicator must be delivered not only for the whole survey scope but also for two aggregated sectors (industry, services) and for several less aggregated categories of sectors (for example: High and medium high-technology manufacturing).