graybill_f function

Graybill F Test

Hypothesis test as described by Graybill (1976).

graybill_f(df, Y1, Yj, signif = 0.05, output = "simple")

Arguments

• df: A data frame.
• Y1: Quoted name of the standard variable.
• Yj: Quoted name of the proposed variable.
• signif: Numeric value for the significance level used in the test. Default: 0.05.
• output: Defines the type of output. If "simple", a simple data frame is created, with only essential information about the test. If "table", more information is provided, and if "full", a data frame with informations about the test and both variables is created. Default: "simple".

Returns

A data frame. Its dimensions will vary, according to the output argument.

Details

This test is used to compare two variables, usually a proposed method, and a standard variable.This test is popular among forestry engineers, specially because, since it considers all data in it's analysis, it's usually more precise than a standard mean t-test. If the data has outliers, the mean may not represent the data correctly, so Graybill F test is specially useful for heterogeneous data.

A simple model regression is applied, and it's significance is evaluated by applying Graybill F test for the parameters estimate, according to the methodology described by Graybill (1976).

Examples

library(forestmangr)
data("exfm11")
head(exfm11)

# The data frame exfm11 contains a height variable called "TH". This will be our
# standard value. We'll compare it to height estimated using different hypsometric equations.
# These are variables "TH_EST1" and "TH_EST2":
graybill_f( exfm11,"TH", "TH_EST1")

# TH_EST1 is statistically different from "TH".

# It's possible to alter the test's significance level using the signif argument:
graybill_f( exfm11,"TH", "TH_EST1", signif = 0.01)

# Different output options are available through the output argument:
graybill_f( exfm11,"TH", "TH_EST2", output="table")
graybill_f( exfm11,"TH", "TH_EST2", output="full")

References

Campos, J. C. C. and Leite, H. G. (2017) Mensuracao Florestal: Perguntas e Respostas. 5a. Vicosa: UFV.

Graybill, F. A. (1976) Theory and application of the linear model. Massachusets: Ouxburg 239 Press.

Leite, H. G. and Oliveira, F. H. T. (2006) Statistical procedure to test identity between analytical methods, Communications in Soil Science and Plant Analysis, 33(7–8), pp. 1105–1118.

Author(s)

Sollano Rabelo Braga sollanorb@gmail.com