Two experiments were conducted to evaluate the accuracy of a conventional-type strain probe, whose section was 10 cm long. The strain gauge that was attached to the section could detect only one bending strain. The first experiment tested three different thicknesses of the shearing deformation layer at different heights of a conventionaltype strain probe. The difference between the actual and calculated displacement was large when the thickness of the shearing deformation layer was smaller than the length of a single section, but it decreased with increasing thickness. This difference was caused by a strain propagation from the upper to the lower section of the conventionaltype probe that was being tested. The second experiment reproduced the distribution of bending strains in a conventional-type strain probe, for seven cases. In which, the conventional probe showed unrealistic calculated displacements during our field survey on an experimental slope in Asuka Village, Central Japan. Two types of deformations of the probe can explain the field data: a single bending deformation in the downslope direction at a certain depth with a small deformation in the upper part, and a convex deformation in the downslope direction at a certain depth with scarce upper deformation. Since the former type showed excessive displacement values compared with our field observations, convex-shaped deformation had indeed occurred in the field. In addition, convex-shaped deformations were detected using a flexible printed-circuit (FPC) type strain probe, at the same time in the field. However, the displacements reproduced were larger than those obtained using an FPC-type probe. The two experiments revealed that in detecting shearing deformation of the soil, calculated displacements obtained from conventional-type probes are untrustworthy, without tests for careful laboratory reproducibility.