In order to investigate this problem we have studied a sample of editorials in three leading Swedish newspapers during the last fifty years. We chose to examine three newspapers with different political outlooks. Dagens Nyheter is the largest morning paper in Sweden and has traditionally been a liberal paper. In the 70's, however, it has claimed that it is a politically independent newspaper. Svenska Dagbladet is the leading conservative organ in Sweden and it is, like Dagens Hyheter, published in Stockholm. Arbetet, our third newspaper, is published in Malmo and is the largest Social Democratic paper in Sweden. The editorials in these newpapers one month before every election since 1928 have been examined. The material covers sixteen election campaigns and some 1,3 editorial pages. Our main purpose has been to illustrate how the political debate has been affected by opinion polls.
For editorial policy it is important to estimate readers' appreciation of the contents. The panel method is one way of sounding the readers and getting a contribution to the editorial policy. In this paper we will present experiences and examples from various media.
The aim of the analytical evaluations presented here is to make better use of the information contained in the copy-test data than it has been happening thusfar. Better use must be made of such information in order to obtain more exact indications for an optimal combination of the editorial offer. The principle of this new method of approach consists of coordinating in a proper manner the evaluation per article with the evaluation per person. A three stage evaluation system was formulated on the basis of these considerations. It consists of a table, simulation and optimization program that can be adapted to the demands of actual situations in a flexible manner.