Prevention of soil compaction requires examining soil mechanical stability. This is usually done by applying the basic concept of the pre-compression stress σp, assuming a fully elastic behavior for soils being loaded by stresses lower than σp and a mixture of elastic-plastic behavior being loaded by higher stresses. However, a small cumulative plastic deformation may occur even when the applied stresses is not beyond σp being indeterminable by current pe-compression test procedures which is not accounting for unloading steps. Therefore, we investigated an alternative approach to simultaneously determine σp and mechanical elasticity of soil by introducing unloading steps (referred to as release compression, RC). Results were compared to the conventional procedure with continuously increasing loads (continuous compression, CC). In the case of RC, soil settlement was measured both after each loading (RC-L) and after each unloading (RC-U) phases between two consecutive pressures providing two load displacement curves. Oedometer tests were conducted with three different soils comprising different clay contents ranging from 75 to 454 g kg-1. Remolded soil was filled into steel rings and compressed under pre-determined loads in a loading frame providing test samples with known pre-compression stresses of 60 and 120 kPa. The results revealed that σp values determined under two different applied loading paths (RC vs. CC) were significantly different. Both testing approaches resulted in an overestimation of σp values based on stress-strain curves which was more pronounced for RC-L compared to CC paths. However, the results from RC-U paths were closer to the CC method than the RC-L path suggesting that if an unloading step is used, RC-U curve should be used to determine σp values. The results also showed that soil with higher clay contents comprised the lowest σp value determined from the oedometer tests among all examined soils. From release compression test, we were also ab
