This thesis is an examination and a critical account of the Swedish Election Studies, carried out at the Department of Political Science, Gothenburg University in collabo­ration with Statistics Sweden. The primary sources are the fifteen main reports delivered to the public through the years 1977 – 2006. The method is to reconstruct some central blocks of thought and ideas in these reports, on the one hand, and to scrutinize and criticize them, on the other. The overarching result is that the Swedish Election Studies, in spite of its proclaimed scientific objectivity and value-neutrality, in fact can be characterized as permeated by normative political thought, explicitly as well as implicitly.

Concerning the view of science, the Swedish Election Studies proclaims objectivity and value-neutrality. This is criticized in the thesis as impossible, from the point of view of philosophy of the social sciences, and as de facto wrong, since normative political thought is shown to be both present and important in all the reports.

Concerning the Swedish democratic system, the Swedish Election Studies: 1) use Sweden as its immediate starting-point for description and explanation, and not some specific ideal or normative perspective, 2) maintain that Swedish democracy, empirically seen, equals electoral and representative democracy (swed. “valdemokrati”), 3) and that Swedish democracy, equally empirically seen, is in good health. This is criticized in the thesis as containing a subtly masked and deviously defended implicit normative ideal, i.e electoral and representative democracy, as a democratic model, and for leading to an equally implicit and masked guardianship of the existing democratic system in Sweden.

Concerning the enlightened citizen, the Swedish Election Studies formulate an explicit ideal. Citizens in a democracy should be politically enlightened in order to be self-governing; and according to the accompanying operational criteria, Swedish citizens are. This is criticized in the thesis as an ideal that is reduced to an utter absurdity, where the operational criteria chosen are both narrow and shallow, and the interpretative thresholds low. The accompanying empirical results, and the continuing interpretation in all the reports, are that Swedish citizens always are sufficiently enlightened. This is regarded in the thesis as just another block of thought in the guardianship and normative bias inherent in the Swedish Election Studies.