Experimental design was an experimental design (with each action step is exactly defined) to obtain information that is related or necessary in discussing the issues at hand.
That is, the design of an experiment is complete steps need to be taken before the experiment is conducted so that the required data can be obtained so that an objective analysis obtained and appropriate conclusion over the issues that are being discussed.
An experimental design aiming to obtain or gather information as many necessary and useful in the investigation of problems that will be discussed. But in collecting information, the design should be made as simple as possible. This is supported by the fact that such a design would be quickly analyzed in addition to economical. Some of the terms commonly used in the design of the experiment are:
a. Treatment is of experimental conditions to be used against the experimental unitsthat will be used against the experimental units in the scope of the selected design.
b. experimental Unit is a single unit against which treatment (which may be a composite of several factor) charged in a basic experiment replication.
c. errors of experiments stated the failure of two units identical experiments that arepreferential treatment to give the same results.
The principle – the principle of design of experiments:
a. replication, is the repetition of the fundamental experiment aims to:
1) Provide experimental errors of estimates that can be used to determine the length of the interval konfidensi or as a basic unit of measurement to determine the levelof significance of differences observed treatise.
2) produces a more accurate estimates for the confusion of experiments.
3) Allows getting a better estimate on the average effect of a factor.
b. Randomization or Redemisasi.
Randomization test valid cause that cause the data analyzed, it‘s assuming it is as though the assumption of independent observations – are met. Randomization allowsresumption of the next steps in the experimental randomization means not to guarantee the independence of the occurrence, but not only minimize the presence of correction between observations. Another benefit of randomization is to remove bias from the data.
c. Local Control
Is steps-steps that form a balancing, pengkotakan, and grouping of units used in the design of experiments. If replication and randomization would essentially allow the enactment of significance test, then map the local control led to local causes, more efficient design.
Grouping is the placement of a group of homogeneous experimental units into different groups got the same treatment.
Pemblokan is allocated to the experimental units in the block, so that the units in the block are relative are homogeneous whereas large variations that can be alleged to be part of the experiment so that design errors more efficiently.
Balancing is acquiring experimental units, the business grouping, pemblokan, and the use of the treatment of experimental units so that the resulting balanced configuration.
Steps-steps to make experimental design depends on the problem at hand, but all contains experimental design staple things, among others:
1. Statement of the problem or the issue discussed.
2. Formulation of hypothesis.
3. determination of eksperimenyang engineering and design is required.
4. Examination of all the possible outcomes and the background or base an-excuse so that experiments yield the required information.
5. Consider all the best results possible in terms of statistics the expected procedures apply.
6. Perform the experiment.
7. The use of statistical techniques against the results.
8. Deduction using the degree of confidence in the things tested.
9. Research throughout research, compared to other research studies – about the same problem.
To investigate a particular combination of factor then performed a combination treatment. When experimental design only discuss one factor only factorial analysis, discuss about all combinations between different degrees of each factor. In other words, the factorial experiment is an experiment where all levels of a particular factor is combined with all the other faltor levels contained in the experiment. Based on the presence of many degrees of each factor in falam, experimentation is often given the name by adding a multiplication factor levels among many that one with many other factor levels.