Statistical analysis for solar cell research

The use of statistical blocks is important for larger experiments with multicrystalline silicon wafers and in cases where the cell parameter depends on the material quality (such as V_0c). Neighbour wafers have to be used for these kind of experiments and each neighbour type has to be distributed over all the treatments. Each neighbour type represents one block and each block contains all the treatments. Cells that had a certain treatment represent an experimental group. With the use of the statistical blocks one can separate the two sources of variance within one group: variance due to the difference of different neighbour types and variance due to the experimental random error. This will result in a reduction of the error. The use of statistical blocks is not important for cell parameters that are not or only slightly dependent on the material quality (such as FF)