Skip to content

K1K2 Design

  • Experimental units are grouped into k1 blocks, each containing k2 units.
  • Blocks are arranged so every pair of units from different blocks is included together the same number of times, which helps separate block effects from factor effects.
  • Common instances include the Latin square (used in agriculture) and the Balanced Incomplete Block design (used in psychology).

A (k1, k2)-design is a type of combinatorial design used in the design of experiments. The experimental units are divided into k1 groups called blocks, and each block consists of k2 units. The blocks are arranged so that every pair of units from different blocks is included in the same number of blocks.

In a (k1, k2)-design, grouping experimental units into k1 blocks of k2 units provides a structured way to control and distribute block-related variation. The defining property—that every pair of units from different blocks appears together the same number of times—ensures an even distribution of block effects across comparisons, which makes it easier to determine the effect of each experimental factor on the outcome.

The Latin square design, often used in agricultural experiments, arranges experimental units in a k1 x k2 grid. In this arrangement each row and each column represents a different block, and every pair of units from different blocks is included in exactly one block, facilitating determination of factor effects.

The Balanced Incomplete Block design, commonly used in psychological experiments, divides experimental units into k1 blocks with each block consisting of k2 units. Unlike the Latin square, units within each block are not arranged in a specific pattern, which allows greater flexibility in design but can make it more difficult to determine the effects of different factors.

  • Agriculture (example: Latin square)
  • Psychology (example: Balanced Incomplete Block design)
  • Other areas of science and engineering
  • Balanced Incomplete Block designs allow greater flexibility because units within each block need not follow a pattern, but this flexibility can make it more difficult to determine the effects of different factors.
  • Latin square design
  • Balanced Incomplete Block design (BIB)