metRology (version 0.9-17)

ilab-class: The 'ilab' class.

Description

The ‘ilab’ class and its constructor function.

Usage

construct.ilab(org, item, measurand, x, u, df, k, U, U.lower, U.upper, distrib=NULL, distrib.pars=NULL, study=NA, title=NA, p=0.95, ...)

Arguments

org

Character vector or factor of organisation names.

item

vector or factor of identifiers for test items. Coerced to factor on storage.

measurand

vector or factor identifying the measurand(s) involved in the study.

x

numeric vector of reported values.

u

numeric vector of reported standard uncertainties or standard errors associated with x.

df

optional numeric vector of degrees of freedom associated with each reported uncertainty.

k

numeric vector of coverage factors. The coverage factor is the factor multiplying u to obtain U.

U

numeric or character vector of expanded uncertainties or confidence interval half-widths. Coerced to numeric but may include a character representation of interval limits; see Details.

U.lower, U.upper

numeric vectors of lower and upper limits for the confidence interval around x, allowing asymmetric intervals. Defaults to U or to the limits specified by U. See Details.

distrib

A character vector of length length(x) or a named list of names of distribution functions associated with u. If a character vector, distrib is recycled to length length(x).

distrib.pars

A named list of lists of parameters describing the distributions associated with u to be passed to the relevant distribution function. If distrib is present but distrib.pars is not, an attempt is made to set defaults based on other parameters; see Details.

study

A character value or vector or a factor identifying different studies or study populations within the data set. Typically used, for example, for identifying participants in global and regional components of a combined study. Recycled to length length(x) if necessary.

title

An optional title for the study. May be a character vector, in which case each element is displayed on a separate line when printed.

p

Confidence level assumed to apply to k. Used only to set a default value for df when distrib indicates a t-distribution and df is unspecified.

...

Other named factors or character vectors used to group observations.

Value

An object of class ‘ilab’ consisting of:

title

A character value or vector describing the study

subset

A character string describing any subset operation used to form the object.

data

A data frame with columns:

org

Factor of organisations submitting results in the study

item

Factor of test item identifiers.

measurand

Factor of measurands determined for each item

x

numeric vector of reported values.

u

numeric vector of reported standard uncertainties or standard errors associated with x.

df

numeric vector of degrees of freedom associated with each reported uncertainty. Set to NA if not provided.

k

numeric vector of coverage factors. The coverage factor is the factor multiplying u to obtain U.

U

numeric or character vector of expanded uncertainties or confidence interval half-widths. U is coerced to numeric but may include a character representation of interval limits; see Details.

U.lower, U.upper

numeric vectors of lower and upper limits for the confidence interval around x.

study

Identifier for study groups (see Arguments above).

...

Other grouping factors (supplied in ‘...’ in construct.ilab) which can be used for sub-categorisation.

distrib

An unnamed list of distribution names.

distrib.pars

An unnamed list of lists of parameters describing the distributions associated with u.

Details

If U is a character vector, it may contain character representations of range. Two forms are permitted:

"a-b"

Interpreted as limits of a range from a to b. U.lower and U.upper are calculated from these limits and x

"+a[/]-b" (or "-a[/]+b")

U.upper is set to a in "+a", and U.lower is set to b in "-b".

If distrib.pars is missing, an attempt is made to deduce appropriate distribution parameters from x, u, df and distrib. In doing so, the following assumptions and values apply for the respective distributions:

norm

mean=x$name, sd=u$name.

unif

min=x-sqrt(3)*u, max=x+sqrt(3)*u.

tri

min=x-sqrt(6)*u, max=x+sqrt(6)*u, mode=x.

t, t.scaled

df=df, mean=x, sd=u.

In addition, if distrib contains "t" or "t.scaled", and df is NA, the corresponding degrees of freedom are chosen based on k and p.

References

None, yet.

See Also

print.ilab, subset.ilab, plot.ilab

Examples

data(Pb)
construct.ilab(org=Pb$lab, x=Pb$value, measurand="Pb", item="none", 
                u=Pb$u, k=Pb$k, U=Pb$U, title=c("CCQM K30", "Lead in wine"), 
                method=Pb$method)

#Illustrate default for U and automatic distrib.pars
construct.ilab(org=Pb$lab, x=Pb$value, measurand="Pb", item="none", 
                u=Pb$u, k=Pb$k, distrib="norm")

construct.ilab(org=Pb$lab, x=Pb$value, measurand="Pb", item="none", 
                u=Pb$u, k=Pb$k, distrib="t.scaled")