Each material has some limiting factors for its massive use in structural applications. It holds in the case of brittle materials all the more. The main limiting factor for ceramic based materials is the brittleness and its natural scatter. It is necessary to have reliable fracture toughness characteristics when ceramic based material tends to be used in industrial scale. There is a few methods used for fracture toughness determination; straight notch and chevron notch technique are the most frequently used. Both methods have advantages (i.e. straight forward evaluation in the case of straight notch and easy preparation in the chevron notch technique) and disadvantages (i.e. slow preparation stage in the straight notch and complicated evaluation in case of the chevron notch). There is an increasing need to test geometrically non standard (i.e. sub-sized) test pieces. This paper is focused on the statistical aspect of size effect at fracture toughness evaluation having used both techniques. Sets of test pieces tested at various geometrical conditions were used for this study. The obtained fracture toughness data were statistically evaluated with the aim to describe and explain the influence of the geometrical parameters on the fracture toughness and scatter of measured data. Fractographical analyses of fracture surfaces were used to identify possible sources of the scatter and to validate the tests.