NEWS.md

In misty: Miscellaneous Functions 'T. Yanagida'

misty 0.5.4 (2023-11-14)

New features

• New function shift() to compute lagged or leading values of a vector.

Bug fix

• Fixed a bug in the function libraries(), version of the packages were not correctly displayed.

misty 0.5.3 (2023-09-17)

Bug fix

• Fixed a bug in the function test.welch(), to remove errors for r-devel from a recent change in r-devel.

misty 0.5.2 (2023-08-24)

Minor features and improvements

• Added the argument group.ind to the function result.lca() to specify latent class indicators as grouping variable in the bar charts.

misty 0.5.1 (2023-08-22)

Minor features and improvements

• Function mplus.lca() can be used to conduct latent class analysis with count, unordered categorical, and ordered categorical indicator variables.
• Function result.lca() can be used to save bar charts with error bars for confidence intervals for each of the latent class solutions.

Bug fix

• Fixed a bug in the function dominance.manual(), function provided the wrong rank orderning.

User-visible changes

• Renamed the functions mplus.lpa() and results.lpa() to mplus.lca() and results.lca().

misty 0.5.0 (2023-08-08)

New features

• New function item.invar() for evaluating configural, metric, scalar, and strict between-group or longitudinal (partial) measurement invariance.
• New function robust.coef() for computing heteroscedasticity-consistent standard errors and significance values for linear models estimated by using the lm() function and generalized linear models estimated by using the glm() function.
• New function dominance() for linear models estimated by using the lm() function and dominance.manual() to conduct dominance analysis based on a (model-implied) correlation matrix of the manifest or latent variables.
• New function check.resid() for performing residual diagnostics to detect nonlinearity (partial residual or component-plus-residual plots), nonconstant error variance (predicted values vs. residuals plot), and non-normality of residuals (Q-Q plot and histogram with density plot).
• New function mplus.lpa() for writing Mplus input files for conducting latent profile analysis based on six different variance-covariance structures.
• New function result.lpa() for creating a summary result table for latent profile analysis from multiple Mplus output files within subfolders.

Minor features and improvements

• Added the argument order to the function multilevel.cor() to order variables in the output table so that variables specified in the argument between are shown first.
• Added modification indices for parameter constraints to the function multilevel.cfa() and multilevel.invar().
• Added residual correlation matrix to the function item.cfa(), multilevel.cfa(), and multilevel.invar().
• Function write.result() can also write results based on the return object of the std.coef function.

User-visible changes

• Renamed the argument min.value in the function item.cfa(), multilevel.cfa(), and multilevel.invar() to mod.minval and changed the default setting to 6.63.

misty 0.4.12 (2023-07-08)

User-visible changes

• Removed the R package r2mlm from the Imports field in the DESCRIPTION due to dependencies issues.

Minor features and improvements

• Function multilevel.descript() can also deal with between-cluster variables by reporting means and standard deviations at the cluster level.
• Added the argument print to the function multilevel.descript() to request standard deviation of the variance components.

misty 0.4.11 (2023-06-10)

New features

• New function multilevel.fit() for computing simultaneous and level-specific model fit information for a fitted multilevel model containing no cross-level constraints from the R package lavaan.
• New function multilevel.cfa() for conducting multilevel confirmatory factor analysis using the R package lavaan to investigate four types of constructs, i.e., within-cluster, shared, configural, and simultaneous shared and configural cluster constructs.
• New function multilevel.invar() for evaluating configural, metric, and scalar cross-level measurement invariance using multilevel confirmatory factor analysis.
• New function multilevel.omega() for computing point estimate and Monte Carlo confidence interval for the multilevel composite reliability defined by Lai (2021) for a within-cluster construct, shared cluster-level construct, and configural cluster construct.

Minor features and improvements

• Added convergence checks to the function multilevel.cor(), e.g., warning message is printed when absolute correlations are greater than 1.
• Argument cluster in the function multilevel.cor(), multilevel.descript(), and multilevel.icc() can also be specified using the variable name of the cluster variable in x.

User-visible changes

• Revised the output of the item.cfa() function, e.g., loglikelihood and information criteria are shown above chi-square test of model fit and label Ad Hoc changed to Scaled.

misty 0.4.10 (2023-05-18)

Bug fix

• Fixed a bug in the function multilevel.cor(), which caused an error message (thanks to Richard Janzen).

misty 0.4.9 (2023-05-02)

New features

• New function libraries() to load and attach multiple add-on packages at once.
• New function check.outlier() computes statistical measures for leverage, distance, and influence for linear models estimated by using the lm() function

Minor features and improvements

• When using function write.result(), result tables are in line with the arguments print, tri, digits, p.digits, and icc.digits specified in the object x (thanks to Stefan Kulakow).
• Function crosstab() displays marginal row-wise, column-wise, and total percentages in the output (thanks to Joachim Fritz Punter and Lisa Bucher). Note that the function now also returns the crosstable in the list element result$crosstab of the return object .

User-visible changes

• Revised the Value sections in the documentation of the functions.
• Changed the default setting of the argument weighted in the test.t and the na.auxiliary function to FALSE in line with the recommendation by Delacre et al. (2021).
• Renamed the function collin.diag() to check.collin().

Bug fix

• Fixed a bug in the function read.mplus(), an error message was printed if comments in the Mplus input file contains special characters (e.g., ä, ü, ö).
• Fixed a bug in the function std.coef(), the function was not applicable to predictors specified as character vector or factor.

misty 0.4.8 (2023-03-10)

New features

• New functions script.close(), script.new(), script.open(), and script.save() to close, open, and save R scripts in RStudio.
• New function setsource() to set the working directory to the source file location in RStudio equivalent to using the menu item Session - Set Working Directory - To Source File Location.
• New function restart() to restart the RStudio session equivalent to using the menu item Session - Restart R.
• New function multilevel.r2.manual() to compute R-squared measures by Rights and Sterba (2019) for multilevel and linear mixed effects models by manually inputting parameter estimates.

Minor features and improvements

• Functions center(), cluster.scores(), rec(), and item.reverse() can be applied to more than one variable at once.

misty 0.4.7 (2023-01-06)

New features

• New function aov.w() for performing repeated measures analysis of variance (within-subject ANOVA) including paired-samples t-tests for multiple comparison, descriptive statistics, effect size measures, and a plot showing error bars for within-subject confidence intervals.
• New function ci.mean.w() for computing difference-adjusted Cousineau-Morey within-subject confidence intervals.

Minor features and improvements

• Function ci.mean.diff() computes the confidence interval for the difference for an arithmetic mean in a one-sample design.
• Functions aov.b(), test.t(), test.welch(), and test.z() plot difference-adjusted confidence intervals in two-sample design by default.
• Added the argument jitter.height to the functions aov.b(), test.levene(), test.t(), aov.welch(), and test.z().
• Added the argument adjust to the function ci.mean(), to apply difference-adjustment for the confidence interval.

User-visible changes

• Function test.t() displays the confidence interval for the mean difference in the one-sample t-test.

Bug fix

• Fixed a bug in the function test.t(), result table provided by the function did not display the confidence interval correctly.

misty 0.4.6 (2022-06-08)

New features

• New function aov.b() for performing between-subject analysis of variance including Tukey HSD post hoc test for multiple comparison.

Minor features and improvements

• Function as.na() is also applicable to arrays
• Added the argument plot and arguments for various graphical parameters for plotting results to the functions test.levene(), test.t(), test.welch(), and test.z().
• Added the argument write for writing results into an Excel file to the functions cor.matrix(), crosstab(), descript(), freq(), item.alpha(), item.cfa(), item.omega(), multilevel.cor(), multilevel.descript(), na.coverage(), na.descript(), and na.pattern()
• Added the argument posthoc for conducting Games-Howell post hoc test for multiple comparison to the functions test.welch().

misty 0.4.5 (2022-04-29)

New features

• New function item.cfa() for conducting confirmatory factor analysis using the R package lavaan.

Minor features and improvements

• Function write.result() can also write results based on the return object of the item.cfa() function.
• Argument exclude of the function freq() can also be set to FALSE.

User-visible changes

• Revised the output of the function multilevel.cor() to make it consistent with the output of the function item.cfa().
• Changed the argument na.omit in the functions multilevel.cor() to missing to make it consistent with the arguments of the function item.cfa().
• Changed the default setting of the argument estimator in the functions multilevel.cor() to ML, so that full information maximum likelihood method is used for dealing with missing data.

Bug fix

• Fixed a bug in the function multilevel.cor(), function did not use Huber-White robust standard errors, but conventional standard errors when specifying estimator = "MLR".

misty 0.4.4 (2022-02-24)

New features

• New function multilevel.r2() for computing R-squared measures for multilevel and linear mixed effects models.
• New function write.xlsx() for writing Excel files (.xlsx).
• New function write.result() for writing results of a misty object into an Excel file.

Minor features and improvements

• Added mean and variance components to the output of the function multilevel.descript().
• Added the argument round to the function freq() for rounding numeric variables.

User-visible changes

• Added a warning message in the na.test() function when running into numerical problems.
• Changed the default setting of the argument sig in the functions cor.matrix() and multilevel.cor() to FALSE.

Bug fix

misty 0.4.3 (2021-09-30)

User-visible changes

• Examples added to the documentation of the collin.diag() function.

Bug fix

• Fixed a bug in the function print.misty.object(), function did not print the result object of the the function crosstab() correctly when requesting percentages.

misty 0.4.2 (2021-08-19)

New features

• New function multilevel.cor() for computing the within-group and between-group correlation matrix using the lavaan package.
• New function na.test() for performing Little's missing completely at random (MCAR) test.
• New function indirect() for computing confidence intervals for the indirect effect using the asymptotic normal method, the distribution of the product method, and the Monte Carlo method.
• New function multilevel.indirect() for computing confidence intervals for the indirect effect in a 1-1-1 multilevel mediation model using the Monte Carlo method.

Minor features and improvements

• Function cor.matrix() highlights statistically significant correlation coefficients in boldface.
• Function cor.matrix() shows the results in a table when computing a correlation coefficient for two variables.
• Added test statistic (stat) and degrees of freedom (df) to the argument print in the function cor.matrix().
• Added the argument continuity for continuity correction to the function cor.matrix() for testing Spearman's rank-order correlation coefficient and Kendall's Tau-b correlation.
• Substantial speed improvement for the function cor.matrix() when computing Spearman's rank-order correlation coefficient or Kendall's Tau-b correlation.

User-visible changes

• Changed the argument group in the functions center(), group.scores(), multilevel.descript(), multilevel.icc(), and rwg.lindell() to cluster.
• Renamed the function group.scores() to cluster.scores().

Bug fix

• Fixed a bug in the function cor.matrix(), function did not print sample sizes when specifying a grouping variable and using listwise deletion.

misty 0.4.1 (2021-06-07)

Minor features and improvements

• Function write.mplus() writes a Mplus input template with variables names specified in the DATA command along with the tab-delimited data file by default.

User-visible changes

• Removed the argument print() in the write.mplus() function.
• Changed the default setting of the argument weighted in the test.welch() function into FALSE following the recommendation by Delacre et al. (2021).

Bug fix

• Fixed a bug in the function cohens.d(), function printed warning messages of the pt() function.
• Fixed a bug in the function cohens.d(), function could not deal with more than one variable in a one-sample design.

misty 0.4.0 (2021-05-13)

New features

• New function test.t() for performing one-sample, two-sample, and paired-sample t-tests including Cohen's d effect size measure.
• New function test.welch() for performing Welch's t-test including Cohen's d effect size measure and Welch's ANOVA including $\eta^2$ and $\omega^2$ effect size measures.

Minor features and improvements

• Added standard error of the mean to the argument print in the function descript().
• Added the arguments format, label, labels, missing to the function read.sav() to remove variable formats, variable labels, value labels, value labels for user-defined missings, and widths from attributes of the variable.
• Function item.reverse() can also be applied to to items with non-integer values.
• Return object of the function cor.matrix() when specifying a grouping variable comprises the combined results of both groups in the matrices.
• Function read.mplus() can also deal with consecutive variables (e.g., x1-x5).
• Added group and split arguments to the function cohens.d().
• Added Cohen's d effect size measure to the output of the test.z function.
• Function cohens.d() computes various kinds of Cohen's d, Hedges' d, and Glass's $\Delta$ including confidence intervals, e.g., weighted and unweighted pooled standard deviation in a two-sample design, with and without controlling for the correlation between the two sets of measurement in a paired-sample design, or with and without the small-sample correction factor.

User-visible changes

• Renamed following functions: alpha.coef() to item.alpha(), cont.coef() to cor.cont(), cramers.v() to cor.cramer(), levenes.test() to test.levene(), mgsub() to chr.gsub(), omega.coef() to item.omega(), reverse.item() to item.reverse(), phi.coef() to cor.phi(), poly.cor() to cor.poly(), scores() to item.scores(), stromit() to chr.omit(), trim() to chr.trim(), z.test() to test.z(),
• Changed the argument use in the cor.matrix() function into na.omit.
• Changed the default setting of the argument method in the functions multilevel.descript() and multilevel.icc() to "lme4"; if the lme4 package is not installed, "aov" will be used.
• Changed the output of the functions ci.mean.diff() and ci.mean.prop() when computing confidence intervals in two-sample designs, i.e., results are divided in two rows according to the grouping variable.
• Changed the output of the functions ci.mean.diff() and ci.mean.prop() when computing confidence intervals in paired-sample designs, i.e., output reports the number of missing data pairs (nNA), instead of number of missing values for each variable separately (nNA1 and nNA2).
• Changed the output of the functions descript() when specifying the argument levenes.test(), i.e., duplicated labels in the column group or variable are not shown.
• Changed the functions cohens.d() into a generic function with the methods cohens.d.default() and cohens.d.formula().
• Added arguments hypo and descript to the functions test.levene() and test.z().
• Added titles to the output of the freq, descript, and crosstab function.
• Changed the argument as.na in the as.na() function into na.

Bug fix

• Fixed a bug in the function center() which caused an error message in case of groups with only one observation when trying to apply group mean centering.
• Fixed a bug in the function center() which caused an error message when trying to apply grand mean centering of a Level 1 predictor.
• Fixed a bug in the function cohens.d(), an error message was printed in the between subject design whenever specifying a grouping variable with missing values.
• Fixed a bug in the function cor.matrix(), which caused an error when using listwise deletion for missing data while specifying a grouping variable.
• Fixed a bug in the function descript(), which caused an error message when selection only one or two argument statistical measures using the argument print.
• Fixed a bug in the function freq(), where the argument split was broken.
• Fixed a bug in the function test.zz(), where the alternative hypothesis was displayed wrong when specifying alternative = "greater" or alternative = "less".

misty 0.3.2 (2020-06-08)

New features

• New function collin.diag() for collinearity diagnostics including tolerance, (generalized) standard error inflation factor, (generalized) variance inflation factor, eigenvalues, conditional indices, and variance proportions for linear, generalized linear, and mixed-effects models.
• New function std.coef() for computing standardized coefficients (StdX, StdY, and StdYX) for linear models estimated by using the lm() function.
• New function mgsub() for multiple pattern matching and replacements, i.e., gsub() function for matching and replacing a vector of character strings.
• New functions df.duplicated() and df.unique() extracting duplicated or unique rows of a matrix or data frame.

Bug fix

• Fixed a bug in the function read.xlsx(), default setting of the argument progress was wrong.

User-visible changes

• Merged all print functions to a single print function called print.misty.object().

misty 0.3.1 (2020-04-25)

New features

• New function z.test() for performing one sample, two sample, and paired sample z-test.

Bug fix

• Function omega.coef() does not access internal slots of a fitted lavaan object anymore (requested by Yves Rosseel).

User-visible changes

• Added descriptive statistics and confidence intervals to the function levenes.test().
• Changed the output of the functions size.mean(), size.prop(), and size.cor() to include greek letters.
• Changed the argument theta in the size.mean() function into delta.

misty 0.3.0 (2020-04-06)

New features

• New functions ci.mean(), ci.mean.diff(), ci.median(), ci.prop(), ci.prop.diff(), ci.sd(), ci.var() for computing confidence interval for the arithmetic mean, the difference in arithmetic means, the median, the proportion, the difference in proportions, the variance, and the standard deviation.
• New function levenes.test() for conducting Levene's test for homogeneity of variance.
• New function omega.coef() for computing coefficient omega (McDonald, 1978), hierarchical omega (Kelley & Pornprasertmanit, 2016), and categorical omega (Green & Yang, 2009).
• New function read.xlsx() for reading Excel files (.xlsx).

Minor features and improvements

• Added ordinal coefficient alpha to the function coef.alpha().
• Added Kendall-Stuart's Tau-c correlation coefficient to the function cor.matrix().
• Function as.na() can also replace user-specified values with missing values in lists.

User-visible changes

• Changed the argument use in the alpha.coef() function into a logical argument na.omit.
• Changed the argument pval.digits in the cor.matrix() function into p.digits.
• Merged print functions print.cont.coef(), print.cramers.v(), print.na.auxiliary(), print.na.coverage(), print.phi.coef(), and print.poly.cor() into print.square.matrix()

Bug fix

• Fixed a bug in several function, where is.vector() function was used to test if an object is a vector. Instead is.atomic() function is used to test if an object is a vector.
• Fixed a bug in the function as.na(), function converted strings in data frames to factors.

misty 0.2.2 (2020-02-26)

New features

• New function trim() for removing whitespace from start and/or end of a string. Note that this function is equivalent to the function trimws() in the base package. However, the trimws() function fails to remove whitespace in some instances.

Bug fix

• Fixed a bug in the function cohens.d(), function returned NA for Cohen's d in within-subject design in the presence of missing values
• Fixed a bug in the function alpha.coef(), function did not provide any item statistics irrespective of the argument print
• Fixed a bug in the function as.na(), function always generated a warning message irrespective of the argument as.na

misty documentation built on Nov. 15, 2023, 1:06 a.m.