NCStats

About NCStats
Package R docs
News
Download/Files
Check results
Installation
Other

News/Changelog

0.3.4 ongoing
o print.anova(): Added (actually resurrected with modifications to work properly).

0.3.3 21Dec12
o Replaced ImportFrom for TeachingDemos with relax, as that is the original slider().
o General: added call.=FALSE to several stop()s.
o General: replaced paste() inside of several cat()s.

o ciSim(): Modified in a variety of ways.  First, I changed to using gslider()
    rather than slider().  Second, I streamlined the main plotting function
    (creating an internal function called ciSimPlot()).  Third, I removed the y-axis
    from the plot, changed the "summary statistics" printed at the top so that they
    fit better, and darkened the parameter line.  Fourth, I used match.arg() to
    catch problems with method=.
o cltSim(): Modified.  I changed to using gslider() rather than slider() and streamlined
    the main plotting function (creating an internal function called cltSimPlot()).
o compIntercepts(): Modified.  Moved the handling of digits to the main function
    rather than the print method.
o compSlopes(): Modified.  Moved the handling of digits to the main function
    rather than the print method.
o distribSim family of functions:  Modified.  I changed to using gslider() rather
    than slider().
o meanMedian(): Modified in a variety of ways.  First, I changed to using gslider()
    rather than slider().  Second, I changed the plots so that the histogram is
    stacked on top of the scatterplot which now contains all of the color coding
    for examining both the mean and the median.  This change makes the plotting
    window much smaller and easier to manipulate.  Third, I add the outlier= argument
    so that outliers can be easily modeled.  Fourth, I made several aesthertic
    changes (shading of histogram, color of points relative to the median, color
    of the residuals relative to the mean, default values for sliders, maximum
    values for alpha and beta sliders).
o NCStatsSims(): Deleted.  Only works in non-Rstudio versions of Windows.
o powerSim(): Modified by I changed to using gslider() rather than slider() and
    streamlined the code (creating an internal function called powerSimPlot()).
o print.confint.glht(): Added.
o print.summary.glht(): Added.
o print.summary.glm(): Added.
o print.htest(): Added.
o print.lm(): Added.
o print.summary.lm(): Added.
o sdCalc(): Modified by replacing numdigs= with digits=.
o sumTable(): Modified by adding the digits= argument.
o transChooser(): Modified in a variety of ways.  First, I changed to using gslider()
    rather than slider().  Third, streamlined the code so that the one-way and
    two-way share more code and the SLR and IVR do as well.  Fourth, I eliminated
    the fitPlot() and the possible call to ncvTest() for the regression code.
    Fifth, I made show.stats=TRUE the default.  Sixth, changed the starty= and
    startx= argument names to shifty= and shiftx=, respectively.  Seventh, I
    removed the dontrun{}s from the examples.  Eight, I added an argument to control
    the color of the bars in the histogram and made the default a light gray.
    Ninth, I updated the help file.
o transChooser_ANOVA() and transChooser_REGRESS(): added these internal files.
    See transChooser() above.
o transChooser.lm(), transChooser.ONEWAY(), transChooser.TWOWAY(), transChooser.SLR(),
    transChooser.IVR(): deleted these files.  See transChooser() above.

0.3.2 1Dec12
o Changed R dependency to >2.14.0

o .onLoad(): slight modification to avoid possible warning on RCMD Check.
o addMargins(): moved to FSA.
o addSigLetters(): moved to FSA.
o bootCase methods: moved to FSA.
o hist.formula(): moved to FSA.
o print.anova(): deleted.
o view(): deleted, moved to FSA.

0.3.1 25Nov12
o Switched to using Project development in RStudio.
o Switched to using semantic versioning for version numbers (which means that the
    hyphen before the last number in the version is replaced with a period).
o Deleted all references to deprecated functions (these are now very old).
o Added depends on FSA (so functions will be available to NCStats users).
o Deleted depends on sciplot.
o Deleted importsFrom gdata and gplots.
o Moved car from importsFrom to depends (so functions will be available to NCStats users).
o Moved plotrix from depends to importsFrom (only need thigmophobe() in highlight()).

o attached(): deleted.
o bootCoef(): deleted.  No longer needed as bootCase() from car was modified to
    meet the same needs.
o chooseColors(): deleted, moved to FSA.
o compIntercepts(): changed class name to "compIntercepts" from "CompInts".
o compSlopes(): changed class name to "compSlopes" from "compSlopes".
o confint.bootCase(), htest.bootCase(), hist.bootCase(), plot.bootCase():  Added,
    but were all originally *.bootCoef.
o compIntercepts: changed a cat()-based warning to warning().  Set the digits for
    printing the common covariate to be the same as digits=.
o detachAll(): deleted.
o Ecoli: deleted, moved to FSA.
o fitPlot(): deleted, moved to FSA.
o hoCoef(): deleted, moved to FSA.
o htest(): deleted, moved to FSA.
o legendHelp(): deleted, moved to FSA.
o Mirex: deleted, moved to FSAdata.
o plotBinResp()(): deleted, moved to FSA.
o residPlot(): deleted, moved to FSA.
o sdCalc(): modified class name to "sdCalc" from "sdcalc".
o Subset(): deleted, moved to FSA.
o Summarize(): slightly modified two warnings about factors on the RHS.  Then
    deleted and moved to FSA.
o sumTable(): modified to use tapply() in place of aggregate.table().  This reduces
    a need to importFrom gdata.
o transChooser(): removed require(tcltk) from code (seems to work without it).

0.3-0 8-Nov-12
o Moved several functions to FSA that are used quite often for fisheries
    analyses.  Ultimately, I want to remove the dependency on NCStats.

o binCI(): moved to FSA.
o cltSim(): changed color of bars in both the population and sampling distribution.
o confint.nlsBoot(): moved to FSA.
o fact2num(): moved to FSA.
o gofCI(): removed dependence on binCI() (which is now in FSA) by using binconf()
    from Hmisc directly.  Added a type="goodman" that will perform the method
    from Goodman (1965) (thanks to Paul Rabie for the discussion and code!).
o htest.nlsBoot(): moved to FSA.
o hist.formula(): modified.  Slightly changed mgp() in par().
o hyperCI(): moved to FSA.
o ks2d(): moved to FSA.
o ks2dp(): moved to FSA.
o ksTest(): moved to FSA.
o lagratio(): moved to FSA.
o lsmean(), and related internals: moved to FSA.
o plotH(): moved to FSA.
o poiCI(): moved to FSA.
o popSizesPlot(): moved to FSA.
o pos2adj(): moved to FSA.
o rcumsum(): moved to FSA.

0.2-8 21Jun12
o Switched to testing under R version 2.15.0 (32-bit).
o Changed license specification from "GPL version 2 or newer" to "GPL (>= 2)" to
    avoid warning on check.

o bootCoef(): Added.  Basically a wrapper to bootCase() in the car package with
    the ability to call some S3 methods (see below).
o confint.bootCoef(): Added.  This is essentially the (very) old ci.bc().
o faqNC(): Added.
o fitPlot.glm(): Added.
o fitPlot.logreg(): Added.
o hist.bootCoef(): Added.
o htest.bootCoef(): Added.  This is essentially the (very) old ho.bc().
o levenesTest(): Added.  A convenience wrapper to leveneTest().
o logregPlot(): Removed.  Functionality is now in fitPlot.logreg().
o plot.bootCoef(): Added.
o plotBinResp(): Added.  Used in fitPlot.logreg() and as stand-alone.
o poiCI.rd: Updated details and references in the help page based on information
    provided by Jerry Lewis about the historical use of the method described in
    Ulm.

0.2-7 2Mar12
o added an importFrom for multcomp for glhtSig() below.

o .onLoad(): Modified.  Moved the startup message into packageStartupMessage()
    in hopes of eliminating the warning when checking the package.
o addMargins(): Modified.  Corrected error when dealing with 1-dimenstional
    tables and handling call from percTable() correctly with no margin=.
o addSigLetters(): Modified.  Added code to deal with a bug that appeared if
    the x-axis variable when adding to a two-way ANOVA interaction plot was an
    ordered, rather than regular, factor.
o chisqPostHoc(): Added.
o distrib(): Modified.  Added a digits= argument for printed answers.
o glhtSig(): Added.
o percTable(): Modified.  Added addMargins= argument.  Corrected problem with
    2-dimensional table and call to addMargins with margin=NULL argument.
o sumTable(): Added.
o transChooser(): Modified.  Added alpha= argument and set outlier test p-values
    that are greater than 1 to 1.
o typeoflm(): Modified.  Fixed a call to the deprecated lgrep().

0.2-6 1Oct11
o Switched to compiling under R version 2.13.1 (32-bit).
o Removed utils from importFrom because it is a base package and due to the
    removal of functions below.
o Removed tools from importFrom due to the removal of functions below.
o Removed suggested packages from description.

o findUpdates(): Removed.  Moved to miscOgle package.
o ksTest(): Added.
o swvCode(): Removed.  Moved to miscOgle package.
o swvCounts(): Removed.  Moved to miscOgle package.
o swvPvalue(): Removed.  Moved to miscOgle package.
o swvANOVA(): Removed.  Moved to miscOgle package.
o swvGLHT(): Removed.  Moved to miscOgle package.
o swvHtest(): Removed.  Moved to miscOgle package.
o swvREG(): Removed.  Moved to miscOgle package.
o swv2ad(): Removed.  Moved to miscOgle package.
o swv2PDF(): Removed.  Moved to miscOgle package.
o updateNCStats(): Removed.

0.2-5 19Aug11
o Modified description file to show my e-mail address.
o Added an ImportFrom from tools for texi2dvi used in swv2pdf().

o ABCens90: Added this data file.
o binCI(): Removed reference to an unused type in the .Rd file.
o c.region(): Modified.  Changed all arguments (except cex.ans=) so that they
    did not have defaults.  Also added an argument, lbl.col, to allow the axis
    label to be a different color then the shaded area.  Changed the show.axis=
    argument to lbl.axis=.  Re-ordered the arguments.
o distrib(): Modified.  Changed label for normal distribution so that the sd=
    value is rounded to three decimal places.  This makes for a "cleaner" plot
    when the provided sd value is really a computed standard error.  Added a
    shade.trans= argument to allow the shaded area to have a transparency value;
    this will allow the plots to be a bit softer to the eye.  Added a lbl.col=
    argument for the axis label so that it is not printed in the transparent
    color if transparency is used.  Modified the call to c.region() per the
    changes above.
o expandTable(): Modified.  Added methodology to deal with a "contingency table"
    with only one row or one column.  Also fixed a bug that implied that three
    dimensional tables could be used.
o geomean(): Added.
o geosd(): Added.
o gofCI(): Added.
o meanMedian(): Modified.  Added an argument that allows one to plot user-derived
    data rather than just the data simulated from a beta distribution.  Also
    streamlined the code by adding two internal functions -- meanMedianPlots()
    and meanMedianBeta().
o percTable(): Added.
o plot.htest(): Modified.  Modified the call to c.region() per the changes above.
o plotH(): Modified.  Modified to plot variable names on axes by default.  Fixed
    error in how x- and y-axes were labeled if a formula was not used.
o powerSim(): Modified.  Modified the call to c.region() per the changes above.
o rSquared(): Added.
o swvCode():  Modified.  Modified to allow file= argument to include the .rnw
    extension and still output a file with a .r extension.
o swv2PDF(): Added.

0.2-4 15Jun11
o Switched to compiling under R version 2.13.0.

o accuracyPrecision():  Modified.  Added several arguments that allowed more control
    over the plottings, changed the default number of points plotted, changed the
    default pch used for plotting the mean, renamed the internal functions (to
    camel code), removed the make.labels() internal function (incorporated into
    the makeTargets() internal function), and changed the default coloring for the
    rings on the targets.  Functionality is unchanged.  Hopefully it is a bit more
    general.
o ABDLakes: Added this data file.
o Frogs: Added this data file.
o hist.formula(): Modified.  Added the ylab= argument with a default to "Frequency".
    Added the right= argument with a default to FALSE; this is opposite of the
    default used in base hist().  Added warn.unused=FALSE arguments to the two
    calls to hist() with plot=FALSE so that the warnings would be suppressed.
o lsmean(): Added.  This is the exact lsmean.default() and lsmean.lme() functions
    from Yandell's pda package, including three dependent functions which were
    kept as internal functions in NCStats.  The pda package does not seem to work
    with R >2.13.0.  To make this function available to those that use it I put
    it in NCStats.
o SquareLakePopn: Added this data file (from FSAdata).
o srsdf(): Added.
o updateNCStats(): Modified.  Added the tools package to the packages that should
    not be detached when updating.  Also made sure that the package.dependencies
    function is from the tools package.  This eliminated a warning when checking
    the packages.

0.2-3 18Apr11
o Added importFrom utils for installed.packages() and available.packages().

o compIntercepts(): Modified.  Changed catches at front of function to note if
    a model with an interaction was sent and, if so, that the interaction term
    would be ignored.  Also added an if that will allow the RHS of the model
    formula to have the covariate and factor in either order.
o findUpdates(): Added.
o lgrep(): Modified.  Changed to be a pass-through to grepl() and added a message
    telling the user to use grepl() instead.  Basically I just found grepl() and
    no longer need lgrep().
o pos2adj(): Added.  Allows ability to place point labels according to directional
    coordinates (e.g., "NE").
o residPlot(): Modified.  Corrected an error related to labeling points if they
    were an outlier.  Basically needed to realize that the studentized residuals
    are a named vactor.
o Subset(): Modified.  Added a catch to send a warning message if the resultant
    data frame has no rows in it.
o swvCode(): Modified.  Added the moreItems= argument to allow the user to more
    easily add more items to be removed shen Stangling.  Basically the items in
    moreItems get appended to the items in itemsToRemove.  In addition, changed
    the ItemsToRemove= argument to itemsToRemove (note capitalization).
o swvHtest(): Added.
o updateNCStats(): Added.
o z.test():  Added.  Basically same as z.test() from TeachingDemos package.
    Modified for my use at Northland.


0.2-2 3Mar11
o moved to compiling under 2.12.1 (32 bit)
o changed dependency to >2.11.1
o added a dependency to tcltk package (see transChoooser below).

o addMargins(): Added.  Largely the same as addmargins() except that I changed
    the margin= argument such that margin=1 gives a marginal value at the end of
    the rows and margin=2 gives a marginal value at the end of the columns.  This
    change makes addMargins() fit more closely (in my opinion) with the use of
    margin= in prop.table().
o fact2num(): Added.  Changes a factor to numeric values.  Useful for example
    if age or year is a factor but you need it to be numeric.
o fitPlot.nls(): Modified.  Deleted the col= argument and added col.pt= and
    col.mdl= arguments to allow for different colors for the points and the
    fitted line.
o plotH(): Modified.  Added catches for determining if Y variable is not quantitative
    and if more than one X (or RHS) variable is sent (i.e., function only works
    for one explanatory variable).  Created separate internal functions for whether
    the X variable is quantitative (the old plotH()) or categorical (simply a
    pass through to barplot).
o predictPlot(): Added.  Simply a pass-through to (i.e., does the exact same thing
    as) predictionPlot().  Simply for convenience.
o residualPlot(): Added.  Simply a pass through to (i.e., does the exact same thing
    as) residPlot().  Sipmly for convenience.
o Summarize(): Modified.  Removed dimnames attribute on the level labels in the
    resulting dataframe when a quantitative variable is summarized by a categorical
    variable.  Mostly for aesthetic reasons.  Added a digits= argument to be more
    compatible with other R functions.  Retained the numdigs= argument for
    backward compatability.
o swvCode(): Modified.  Added two more items to ItemsToRemove= argument to handle
    changes to Stangle() output.  Added a blanks= argument that can be used to
    remove all or just "extra" blank lines in the Stangle() output.  Also added
    the annotate=FALSE argument to set as default for sending to Stangle().  This
    forces Stangle() to not include code block comments.
o transChooser(): Modified.  Added require(tcltk) to attempt to ensure that this
    package is loaded.  Corrected an error which caused the Levene's test
    results not to be printed above the residual plot when show.stats=TRUE for
    the ONEWAY and TWOWAY functions.  Corrected an error which caused overprinting
    of the outlier test p-value if outlierTest() returned more than one value
    (only the minimum will be now be printed).  Created internal functions for
    placing the p-value labels above each graph (reduced redundancy).  Removed
    several calls to round() inside of a formatC() call.


0.2-1 31Jan11
o removed dependency on nortest package (see adTest() below).

o adTest(): brought, with name change, ad.test() from nortest package into NCStats.
    I changed to 'camel-case' naming convention and then removed dependency on
    the nortest package.
o Summarize(): moved a bunch of code into internal functions.  Modified the code
    for a quantitative variable by a factor variable so that the result prints
    as a data.frame (rather than a matrix) that includes one or two columns that
    contain the levels of the factor variables.  This provides a nicer looking
    output and an output that could be subsetted if needed.  Also modified the
    code so that multiple "values" of a factor variable can be excluded.  The
    help code examples were modified to illustrate the changes.
o transChooser(): updated ad.test() references to adTest().


0.2-0 1-Dec-10
o added a work-around .Rd files for the three functions imported from the car
    package namespace and exported as part of the NCStats namespace.  This seems
    to have corrected the warning produced during Rcmd check.

o distrib(): modified.  Added the se= argument in the hopes that this might be
    somewhat easier for students that are learning to distinguish between
    sampling and population distributions.
o hist.formula(): modified.  Added the ability to create a histogram of a
    variable not separated by groups with code such as ~x, x~1, or x~0.  This
    allows one to use the data= argument for a single histogram.  Added a
    main= argument for the situation where a single histogram is plotted and set
    the default to "" to avoid the annoying main plot label.  Added a byrow=
    argument so that the user can choose the order in which the plots are populated.
    Changed ylmts= to ymax= argument so as not to confuse ylmts= with ylim=.
    Changed main.pre= to pre.main= to avoid confusion with partial matching with
    main= argument.  Changed cols= argument to ncol= argument (and rows= to nrow=)
    to avoid confusion with partial matching of col= that may be sent in the
    "dots" argument (and to make consistent with notation in matrix()).  Modified
    code such that if pre.main=NULL then no main title will be printed on the
    histograms.  Removed the breaks= argument (will work as part of dots argument).
    Removed the code that "caught" if the formula= argument was not a formula --
    this was redundant because this is an S3 generic for formula types.
o .onLoad(): modified.  Changed to include version number of loaded version.
o plotH(): added.
o residPlot(): modified.  Added code to plot multiple points returned by
    outlierTest().  Add the use of thigmophobe.labels to place outlier text marks
    with a minimum of overlapping.  Added three internal functions --
    addOutlierTestResults(), addLoessLine(), and getMainTitle() to reduce and
    streamline code among the different generics.
o spois(): modified.  Fixed small bug (extra comma).
o summarize.formula(): modified.  Added the ability to create summaries of a
    variable not separated by groups with code such as ~x, x~1, or x~0.  This
    allows one to use the data= argument for a single variable.  Modified default
    exclude= value to better handle these situations.

All prior NEWS in ONEWS file.