allDifferences.data.frame {asremlPlus}R Documentation

Using supplied predictions and standard errors of pairwise differences or the variance matrix of predictions, forms all pairwise differences between the set of predictions, and p-values for the differences.

Description

Uses supplied predictions and standard errors of pairwise differences, or the variance matrix of predictions to form, in an alldiffs.object, for those components not already present, (i) a table of all pairwise differences of the predictions, (ii) the p-value of each pairwise difference, and (iii) the minimum, mean and maximum LSD values. Predictions that are aliased (or nonestimable) are removed from the predictions component of the alldiffs.object and standard errors of differences involving them are removed from the sed component.

If necessary, the order of the columns of the variables in the predictions component are changed to be the initial columns of the predictions.frame and to match their order in the classify. Also, the rows of predictions component are ordered so that they are in standard order for the variables in the classify. That is, the values of the last variable change with every row, those of the second-last variable only change after all the values of the last variable have been traversed; in general, the values of a variable are the same for all the combinations of the values to the variables to its right in the classify. The sortFactor or sortOrder arguments can be used to order of the values for the classify variables, which is achieved using sort.alldiffs.

Each p-value is computed as the probability of a t-statistic as large as or larger than the absolute value of the observed difference divided by its standard error. The p-values are stored in the p.differences component. The degrees of freedom of the t-distribution is the degrees of freedom stored in the tdf attribute of the alldiffs.object. This t-distibrution is also used in calculating the LSD statistics stored in the alldiffs.object.

Usage

## S3 method for class 'data.frame'
allDifferences(predictions, classify, vcov = NULL, 
               differences = NULL, p.differences = NULL, sed = NULL, 
               LSD = NULL, meanLSD.type = "overall", LSDby = NULL, 
               backtransforms = NULL, 
               response = NULL, response.title = NULL, 
               term = NULL, tdf = NULL,  
               x.num = NULL, x.fac = NULL,  
               level.length = NA, 
               pairwise = TRUE, alpha = 0.05,
               transform.power = 1, offset = 0, scale = 1, 
               inestimable.rm = TRUE,
               sortFactor = NULL, sortWithinVals = NULL, 
               sortOrder = NULL, decreasing = FALSE, ...)

Arguments

predictions

A predictions.frame, or a data.frame, beginning with the variables classifying the predictions and also containing columns named predicted.value, standard.error and est.status; each row contains a single predicted value. It may also contain columns for the lower and upper limits of error intervals for the predictions. Note that the names standard.error and est.status have been changed to std.error and status in the pvals component produced by asreml-R4; if the new names are in the data.frame supplied to predictions, they will be returned to the previous names.

classify

A character string giving the variables that define the margins of the multiway table that has been predicted. Multiway tables are specified by forming an interaction type term from the classifying variables, that is, separating the variable names with the : operator.

vcov

A matrix containing the variance matrix of the predictions; it is used in computing the variance of linear transformations of the predictions.

differences

A matrix containing all pairwise differences between the predictions; it should have the same number of rows and columns as there are rows in predictions.

p.differences

A matrix containing p-values for all pairwise differences between the predictions; each p-value is computed as the probability of a t-statistic as large as or larger than the observed difference divided by its standard error. The degrees of freedom of the t distribution for computing it are computed as the denominator degrees of freedom of the F value for the fixed term, if available; otherwise, the degrees of freedom stored in the attribute tdf are used; the matrix should be of the same size as that for differences.

sed

A matrix containing the standard errors of all pairwise differences between the predictions; they are used in computing the p-values.

LSD

A data.frame containing the mean, minimum and maximum LSD for determining the significance of pairwise differences, the mean LSD being calculated using the square root of the mean of the variances of pairwise differences. If factor.combination was specified for meanLSD.type when the LSDs were being calculated, then LSD contains an LSD for each factor.combination of the factors specified by LSDby. Each LSD is calculated from the square root of the mean of the variances for all pairwise differences for each factor combination, unless there is only one predicted value for each factor.combination, when it is based on the standard error of the prediction multiplied by the square root of two. If LSD is not NULL then the overall mean LSD will be added as an attribute named meanLSD of the alldiffs.object, as will the values of meanLSD.type and LSDby. The LSD for a single prediction assumes that any predictions to be compared are independent; this is not the case if residual errors are correlated.

meanLSD.type

A character string determining whether the mean LSD stored is (i) the overall mean, based on the square root of the mean of the variances of all pairwise variances, (ii) the mean for each factor.combination of the factors specified by LSDby, which is based on the square root of the mean of the variances for all pairwise differences for each factor combination, unless there is only one prediction for a factor.combination, when notional LSDs are calculated that are based on the standard error of the prediction multiplied by the square root of two, or (iii) the per.prediction mean, based, for each prediction, on the square root of the mean of the variances for all pairwise differences involving that prediction. If LSD is not NULL then meanLSD.type will be added as an attribute of the alldiffs.object.

LSDby

A character (vector) of variables names, being the names of the factors or numerics in the classify for each combination of which a mean LSD, minLSD and max LSD is stored in the LSD component of the alldiffs.object when meanLSD.type is factor.combinatons.

backtransforms

A data.frame containing the backtransformed values of the predicted values that is consistent with the predictions component, except that the column named predicted.value is replaced by one called backtransformed.predictions. Any error.interval values will also be the backtransformed values. Each row contains a single predicted value.

response

A character specifying the response variable for the predictions. It is stored as an attribute to the alldiffs.object.

response.title

A character specifying the title for the response variable for the predictions. It is stored as an attribute to the alldiffs.object.

term

A character string giving the variables that define the term that was fitted using asreml and that corresponds to classify. It is often the same as classify. It is stored as an attribute to the alldiffs.object.

tdf

an integer specifying the degrees of freedom of the standard error. It is used as the degrees of freedom for the t-distribution on which p-values and confidence intervals are based. It is stored as an attribute to the alldiffs.object.

x.num

A character string giving the name of the numeric covariate that corresponds to x.fac, is potentially included in terms in the fitted model and which corresponds to the x-axis variable. It should have the same number of unique values as the number of levels in x.fac.

x.fac

A character string giving the name of the factor that corresponds to x.num, is potentially included in terms in the fitted model and which corresponds to the x-axis variable. It should have the same number of levels as the number of unique values in x.num. The levels of x.fac must be in the order in which they are to be plotted - if they are dates, then they should be in the form yyyymmdd, which can be achieved using as.Date. However, the levels can be non-numeric in nature, provided that x.num is also set.

level.length

The maximum number of characters from the the levels of factors to use in the row and column labels of the tables of pairwise differences and their p-values and standard errors.

pairwise

A logical indicating whether all pairwise differences of the predictions and their standard errors and p-values are to be computed and stored. If FALSE, the components differences and p.differences will be NULL in the returned alldiffs.object.

alpha

The significance level for an LSD to compare a pair of predictions.

transform.power

A numeric specifying the power of a transformation, if one has been applied to the response variable. Unless it is equal to 1, the default, back-transforms of the predictions will be obtained and presented in tables or graphs as appropriate. The back-transformation raises the predictions to the power equal to the reciprocal of transform.power, unless it equals 0 in which case the exponential of the predictions is taken.

offset

A numeric that has been added to each value of the response after any scaling and before applying any power transformation.

scale

A numeric by which each value of the response has been multiplied before adding any offset and applying any power transformation.

inestimable.rm

A logical indicating whether rows for predictions that are not estimable are to be removed from the components of the alldiffs.object.

sortFactor

A character containing the name of the factor that indexes the set of predicted values that determines the sorting of each component of the the alldiffs.object by sort.alldiffs. If NULL then sorting is not carried out. If there is more than one variable in the classify term then sortFactor is sorted for the predicted values within each combination of the values of the sortWithin variables: the classify variables, excluding the sortFactor. There should be only one predicted value for each unique value of sortFactor within each set defined by a combination of the values of the sortWithin variables.

sortWithinVals

A list with a component named for each factor and numeric that is a classify variable for the predictions, excluding sortFactor. Each component should contain a single value that is a value of the variable. The combination of this set of values will be used to define a subset of the predicted values whose order will define the order of sortFactor to be used for all combinations of the sortWithinVals variables. If sortWithinVals is NULL then the first value of each sortWithin variable in predictions component is used to define sortWithinVals. If there is only one variable in the classify then sortWithinVals is ignored.

sortOrder

A character vector whose length is the same as the number of levels for sortFactor in the predictions component of the alldiffs.object. It specifies the desired order of the levels in the reordered components of the alldiffs.object. The argument sortWithinVals is ignored.

The following creates a sortOrder vector levs for factor f based on the values in x:
levs <- levels(f)[order(x)].

decreasing

A logical passed to order that detemines whether the order for sorting the components of the alldiffs.object is for increasing or decreasing magnitude of the predicted values.

...

provision for passsing arguments to functions called internally - not used at present.

Value

An alldiffs.object with components predictions, vcov, differences, p.differences sed, and LSD.

The name of the response, the response.title, the term, the classify, tdf, sortFactor and the sortOrder will be set as attributes to the object. Note that the classify in an alldiffs.object is based on the variables indexing the predictions, which may differ from the classify used to obtain the original predictions (for example, when the alldiffs.objects stores a linear transformation of predictions.

Also, see predictPlus.asreml for more information.

Author(s)

Chris Brien

See Also

asremlPlus-package, as.alldiffs, as.predictions.frame, sort.alldiffs, subset.alldiffs,
print.alldiffs, renewClassify.alldiffs, redoErrorIntervals.alldiffs,
recalcLSD.alldiffs, plotPredictions.data.frame, predictPlus.asreml,
predictPresent.asreml

Examples

  data(Oats.dat)
  
  ## Use asreml to get predictions and associated statistics

  ## Not run: 
  m1.asr <- asreml(Yield ~ Nitrogen*Variety, 
                   random=~Blocks/Wplots,
                   data=Oats.dat)
  current.asrt <- as.asrtests(m1.asr)
  Var.pred <- asreml::predict.asreml(m1.asr, classify="Nitrogen:Variety", 
                                      sed=TRUE)
  if (getASRemlVersionLoaded(nchar = 1) == "3")
    Var.pred <- Var.pred$predictions
  Var.preds <- Var.pred$pvals
  Var.sed <- Var.pred$sed
  Var.vcov <- NULL
  wald.tab <-  current.asrt$wald.tab
  den.df <- wald.tab[match("Variety", rownames(wald.tab)), "denDF"]
  
## End(Not run)

  ## Use lmerTest and emmmeans to get predictions and associated statistics
  if (requireNamespace("lmerTest", quietly = TRUE) & 
      requireNamespace("emmeans", quietly = TRUE))
  {
    m1.lmer <- lmerTest::lmer(Yield ~ Nitrogen*Variety + (1|Blocks/Wplots),
                              data=Oats.dat)
    Var.emm <- emmeans::emmeans(m1.lmer, specs = ~ Nitrogen:Variety)
    Var.preds <- summary(Var.emm)
    den.df <- min(Var.preds$df)
    ## Modify Var.preds to be compatible with a predictions.frame
    Var.preds <- as.predictions.frame(Var.preds, predictions = "emmean", 
                                      se = "SE", interval.type = "CI", 
                                      interval.names = c("lower.CL", "upper.CL"))
    Var.vcov <- vcov(Var.emm)
    Var.sed <- NULL
  }

  ## Use the predictions obtained with either asreml or lmerTest
  if (exists("Var.preds"))
  {
    ## Order the Varieties in decreasing order for the predictions values in the 
    ## first N level 
    Var.diffs <- allDifferences(predictions = Var.preds, 
                                classify = "Nitrogen:Variety", 
                                sed = Var.sed, vcov = Var.vcov, tdf = den.df,
                                sortFactor = "Variety", decreasing = TRUE)
    print.alldiffs(Var.diffs, which="differences")
  
    ## Change the order of the factors in the alldiffs object and reorder components
    Var.reord.diffs <- allDifferences(predictions = Var.preds,
                                classify = "Variety:Nitrogen", 
                                sed = Var.sed, vcov = Var.vcov, tdf = den.df)
    print.alldiffs(Var.reord.diffs, which="predictions")
  }

[Package asremlPlus version 4.2-32 Index]