Sidebar Menu

Multiple Comparisons

Description on the average comparison sub topic:

  • Hypothesis Test
  • Error Type
  • Level of significant
  • Analysis of variance (Anova)
  • Average comparison
  • The number of possible combinations of pairs
  • Problems in Multiple Comparison

Duncan's test is based on a set of significant difference values that increase in size depending on the distance between the powers of the two mean values being compared. Can be used to test for differences among all possible treatment pairs regardless of the number of treatments.

The LSD test is the simplest and most commonly used procedure for testing the difference between the treatment averages. This method was introduced by Fisher (1935), so it is also known as the Fisher's LSD method [Least Significant Difference]. To use the BNT test, the attributes we need are the mean square of error (KTG), significance level, degrees of freedom (db) of error, and a student t-table to determine the critical value of the comparison test.

The steps for performing comparisons between treatment means are similar to using Duncan's Test. The difference lies only in the comparison value used.

Tukey's test is often also called the honestly significant difference test ( HSD ), introduced by Tukey (1953). The testing procedure is similar to LSD, which has one comparison and is used as an alternative to LSD if we want to test all pairs of treatment means without a plan. Tukey's test was used to compare all pairs of mean treatments after the Analysis of Variety test was carried out.

In certain experimental cases, we may only be interested in comparisons between controls and other treatments. For example, comparing a local variety or standard chemical with a new one. For this case, we can use Dunnet's test. Dunnet developed this test and popularized it in 1955. Dunnet's test maintains the MEER at a level that is no more than the specified level of significance, eg = 0.05. In this method, only one comparison value is needed to compare the control with other treatments. The formula is similar to LSD, but in this test, the t-value used is not the t-student used in the LSD test. Dunnet uses a separate t table, which is usually attached to experimental design books.