This class of objects is returned by EnvStats functions that perform hypothesis tests based on censored data. Objects of class "htestCensored" are lists that contain information about the null and alternative hypotheses, the censoring side, the censoring levels, the percentage of observations that are censored, the estimated distribution parameters (if applicable), the test statistic, the p-value, and (optionally, if applicable) confidence intervals for distribution parameters.

Details

Objects of S3 class "htestCensored" are returned by the functions listed in the section Hypothesis Tests in the help file EnvStats Functions for Censored Data. Currently, the only function listed is twoSampleLinearRankTestCensored.

Methods

Generic functions that have methods for objects of class "htestCensored" include:
print.

Value

Required Components
The following components must be included in a legitimate list of class "htestCensored".

statistic

numeric scalar containing the value of the test statistic, with a names attribute indicating the null distribution.

parameters

numeric vector containing the parameter(s) associated with the null distribution of the test statistic. This vector has a names attribute describing its element(s).

p.value

numeric scalar containing the p-value for the test under the null hypothesis.

null.value

numeric vector containing the value(s) of the population parameter(s) specified by the null hypothesis. This vector has a names attribute describing its elements.

alternative

character string indicating the alternative hypothesis (the value of the input argument alternative). Possible values are "greater", "less", or "two-sided".

method

character string giving the name of the test used.

sample.size

numeric scalar containing the number of non-missing observations in the sample used for the hypothesis test.

data.name

character string containing the actual name(s) of the input data.

bad.obs

the number of missing (NA), undefined (NaN) and/or infinite (Inf, -Inf) values that were removed from the data object prior to performing the hypothesis test.

censoring.side

character string indicating whether the data are left- or right-censored.

censoring.name

character string indicating the name of the data object used to identify which values are censored.

censoring.levels

numeric scalar or vector indicating the censoring level(s).

percent.censored

numeric scalar indicating the percent of non-missing observations that are censored.

Optional Components
The following component may optionally be included in an object of of class "htestCensored":

estimate

numeric vector containing the value(s) of the estimated population parameter(s) involved in the null hypothesis. This vector has a names attribute describing its element(s).

estimation.method

character string containing the method used to compute the estimated distribution parameter(s). The value of this component will depend on the available estimation methods (see Distribution.df).

interval

a list containing information about a confidence, prediction, or tolerance interval.

Note

Since objects of class "htestCensored" are lists, you may extract their components with the $ and [[ operators.

Author

Steven P. Millard (EnvStats@ProbStatInfo.com)

Examples

  # Create an object of class "htestCensored", then print it out. 
  #--------------------------------------------------------------

  htestCensored.obj <-   with(EPA.09.Ex.16.5.PCE.df, 
    twoSampleLinearRankTestCensored(
      x = PCE.ppb[Well.type == "Compliance"], 
      x.censored = Censored[Well.type == "Compliance"], 
      y = PCE.ppb[Well.type == "Background"], 
      y.censored = Censored[Well.type == "Background"], 
      test = "tarone-ware", alternative = "greater"))

  mode(htestCensored.obj) 
#> [1] "list"
  #[1] "list" 

  class(htestCensored.obj) 
#> [1] "htestCensored"
  #[1] "htest" 

  names(htestCensored.obj) 
#>  [1] "statistic"         "parameters"        "p.value"          
#>  [4] "estimate"          "null.value"        "alternative"      
#>  [7] "method"            "estimation.method" "sample.size"      
#> [10] "data.name"         "bad.obs"           "censoring.side"   
#> [13] "censoring.name"    "censoring.levels"  "percent.censored" 
  # [1] "statistic"         "parameters"        "p.value"          
  # [4] "estimate"          "null.value"        "alternative"      
  # [7] "method"            "estimation.method" "sample.size"      
  #[10] "data.name"         "bad.obs"           "censoring.side"   
  #[13] "censoring.name"    "censoring.levels"  "percent.censored" 
 
  htestCensored.obj 
#> 
#> Results of Hypothesis Test
#> Based on Censored Data
#> --------------------------
#> 
#> Null Hypothesis:                 Fy(t) = Fx(t)
#> 
#> Alternative Hypothesis:          Fy(t) > Fx(t) for at least one t
#> 
#> Test Name:                       Two-Sample Linear Rank Test:
#>                                  Tarone-Ware Test
#>                                  with Hypergeometric Variance
#> 
#> Censoring Side:                  left
#> 
#> Censoring Level(s):              x = 5 
#>                                  y = 2 4 5 
#> 
#> Data:                            x = PCE.ppb[Well.type == "Compliance"]
#>                                  y = PCE.ppb[Well.type == "Background"]
#> 
#> Censoring Variable:              x = Censored[Well.type == "Compliance"]
#>                                  y = Censored[Well.type == "Background"]
#> 
#> Sample Sizes:                    nx = 8
#>                                  ny = 6
#> 
#> Percent Censored:                x = 12.5%
#>                                  y = 50.0%
#> 
#> Test Statistics:                 nu     =  8.458912
#>                                  var.nu = 20.912407
#>                                  z      =  1.849748
#> 
#> P-value:                         0.03217495
#> 
  
  #Results of Hypothesis Test
  #Based on Censored Data
  #--------------------------
  #
  #Null Hypothesis:                 Fy(t) = Fx(t)
  #
  #Alternative Hypothesis:          Fy(t) > Fx(t) for at least one t
  #
  #Test Name:                       Two-Sample Linear Rank Test:
  #                                 Tarone-Ware Test
  #                                 with Hypergeometric Variance
  #
  #Censoring Side:                  left
  #
  #Data:                            x = PCE.ppb[Well.type == "Compliance"]
  #                                 y = PCE.ppb[Well.type == "Background"]
  #
  #Censoring Variable:              x = Censored[Well.type == "Compliance"]
  #                                 y = Censored[Well.type == "Background"]
  #
  #Sample Sizes:                    nx = 8
  #                                 ny = 6
  #
  #Percent Censored:                x = 12.5%
  #                                 y = 50.0%
  #
  #Test Statistics:                 nu     =  8.458912
  #                                 var.nu = 20.912407
  #                                 z      =  1.849748
  #
  #P-value:                         0.03217495

  #==========

  # Extract the test statistics
  #----------------------------

  htestCensored.obj$statistic
#>        nu    var.nu         z 
#>  8.458912 20.912407  1.849748 
  #       nu    var.nu         z 
  # 8.458912 20.912407  1.849748

  #==========

  # Clean up
  #---------
  rm(htestCensored.obj)