Universities lead the global science enterprise (Royal Society, 2011) and scholars have attempted to find recipes for performance. If studies have effectively shown how certain individual (Cummings, 2012) and institutional behaviors (Aghion et al., 2009b) are correlated with research outputs, few have highlighted the importance of higher education systems (HES). Studies based on measures of central tendency (Aghion, 2010) and mainstream observations (Marginson, 2006) also failed to explain the performance of outliers. In fact, a per capita analysis reveals that Nordic HES also achieve comparatively high results in terms of world-class universities, publications, citations and patents.
This presentation introduces an innovative research design to identify systemic factors contributing to academic knowledge production and discusses its inherent possibilities and pitfalls. First, following Holmes’ (1981) hypothetical-deductive problem approach and Popper’s (2005) falsification process, previous studies were taking as point of departure leading to the formulation of six hypotheses/factors (beliefs, academic structure, governance, funding, networking and internationalization) in four settings (Denmark, Finland, Norway and Sweden). Since it is impossible to perform regressions based on four national aggregates, it was decided to rely stakeholders’ perception of the six factors. To achieve systemic representativeness, a multi-level and vertical scheme was designed. It included 13 strata (e.g. NordForsk, ministries of higher education, research councils). Based on 400 questionnaires and 60 interviews, factors were considered tentatively true if was achieved saturation in interviews, significant statistics and coherence in all four countries.
There are however pitfalls to this design. First, tensions between quantitative and qualitative ethos emerge when considering reliability and validity. Second, solely focusing on commonalities may fail to explain differences between Nordic and Anglo-Saxon countries. Third, relying on perceptions might lead to circularity and rhetoric of “good research environments.” Finally, the aggregation of “noise” (i.e. institutions and academic disciplines) might prevent a finer evaluation.