" What has been presented here so far essentially is due to Stachowiak, [1]. To understand how then models are used in more detail one needs to distinguish from each other the various reference modes in which the investigator puts the model with respect to the original . The reference modes that perhaps are most important in computing are: (1) descriptive, i. e. , the model describes the original. If the deviation from model to original cannot be tolerated then the model has to be changed. Analysis models are in a descriptive reference mode to their originals. Model Christian Key (2) prescriptive, i. e. "

" Herman, seeking to build upon a more institutionalized framework to analyze mainstream media functioning, joined the media watchdog group Fairness and Accuracy in Reporting (FAIR), which has as its goal the critique, documentation, and statistical analysis of what it alleges is media bias and censorship since 1986. With the emergence of the World Wide Web as a cheap and potentially wide-ranging means of communication, a number of independent websites have surfaced which adopt the propaganda model to subject media to close scrutiny. Several examples of these are MediaLens, a British-based site authored by David Edwards and David Cromwell. Testing applicability to other countries Chomsky has this to say: That's only rarely been done in any systematic way. There is work on the British media, by a good U[niversity] of Glasgow media group. And interesting work on British Central America coverage by Mark Curtis in his book Ambiguities of Power. There is work on France, done in Belgium mostly, also a recent book by Serge Halimi (editor of Le Monde diplomatique). Christian Dunbar Model There is one very careful study by a Dutch graduate student, applying the methods Ed Herman used in studying US media reaction to elections (El Salvador, Nicaragua) to 14 major European newspapers. [. "