Complete Randomized Design (CRD): Easiest to use design. When experimental units are essentially homogeneous, they can use CRD design. Due to the homogeneity requirement, it might be challenging to use this design for field studies. For experiments with a limited number of treatments, the CRD works best.
Treatments are entirely randomly assigned to experimental units. Each experimental unit has an equal chance of receiving any treatment. A random number table, computer, program, etc., is used to perform randomization.
CRD is a Very adaptable design (i.e., the number of replicates and treatments is only constrained by the number of experimental units readily available).
In comparison to other designs, statistical analysis is straightforward. Information loss as a result of missing data is minimal in comparison to other designs because of the greater degree of freedom for the source of variation in error.
There could be a loss of precision if the experimental units are not homogeneous and you don't use blocking to reduce this variation. Unless experimental units are homogeneous, this design is typically the least effective. Unsuitable for a wide range of treatments.
Complete Randomized Design Block (CRBD): A random assignment of treatment combinations to the experimental units within a block is known as a randomized complete block design, or RCBD. Blocks typically can't be randomized because they stand in for factors restricted from randomization, like location, place, time, gender, ethnicity, breeds, etc.
It is not possible to choose a person's gender at random. Picking one country and dialing "X country" is not possible. But their presence—also referred to as nuisance factors—will cause the study's results to vary consistently. For instance, the crops grown in the northern vs. southern regions will experience various climatic conditions.
As a result, whenever possible, they should be controlled. By limiting these obtrusive factors through blocking, one can improve the experiment's precision and reap various other advantages. The experiments in a completely randomized design (CRD) can only influence random, unknowable, and uncontrolled factors (also known as lucking nuisance factors). If there are systematic and well-known sources of variations (also known as nuisance factors), the RCBD is used to manage and control them.
The most popular type of experiment design is arguably the randomized complete block design (RCBD), used in various fields such as engineering, medicine, and agriculture. In addition to minimizing experimental error, the design broadens the generalizability of the research results.
For instance, if the study uses the location as a limiting factor, it might extrapolate the findings to other locations. A fertilizer manufacturer can only assert that their product is effective when tested in various climates regardless of the climate.
The "complete block" component denotes that every treatment combination is used in every block. The experiment would be known as a Randomized Incomplete Block Design if a block missed one or more treatment combinations. Because the treatment combinations are distributed randomly to the experimental units within the blocks, the design would still be randomized. The experiment will still be referred to as a randomized complete block design (RCBD) with missing data, even if a block only has a few observation units with missing data, as opposed to an "incomplete block design."
