RcmdrPlugin.IPSUR by G. Jay Kerns and G. Andy Chang, Youngstown State University

Features of the RcmdrPlugin.IPSUR Package:

This document describes selected features of the RcmdrPlugin.IPSUR package. There are three (3) sections.

  1. THE R COMMANDER: describes extra features the IPSUR plugin offers.
  2. DATA: describes certain datasets provided by the package.
  3. DOCUMENTS: describes documents provided by the package.
    1. Statistics Menu
      • Summaries
        1. Allows numerical variables that are not necessarily factors to be tabularized, with a corresponding modification to the Chi-square goodness-of-fit test.
        2. Sample skewness and kurtosis were added to the Numerical Summaries.
        3. Adds Fisher's exact test with a simulate p-value option.
        4. Calculates relative frequencies instead of percentages.
      • Contingency Tables
        1. Option "Add Marginal Distributions" is available which adds column and row sums to table.
        2. Hypothesis Tests has additional "Simulate p-value" option with customizable number of Iterations.
      • Proportions
        1. Enter table... of summarized data for a one-sample proportion test.
        2. Enter table... of summarized data of independent samples for a test of equality of several proportions.
      • Power Menu: based on functions written by Peter Dalgaard
        1. Power calculation dialog for one sample, two sample, and paired t tests.
        2. Power calculation dialog for two sample proportion test.
        3. Power calculation dialog for balanced ANOVA.
    2. Graphs Menu
      • A Title field was added to the following plots:
        Index Plot, Histogram, Boxplot, Quantile-comparison plot, Scatterplot, Scatterplot matrix, Line Plot, Bar Graph, XY Conditioning plot.
      • Boxplot has additional options “Horizontal”, "Notches" and "Variable box width", and Horizontal, Variable box width are the default settings.
      • It is possible to plot Bar Graphs by groups. Further, if only summarized data are available then you can "Enter a table" for a bar graph.
      • Pareto Diagrams were added.
      • Strip charts were added.
    3. Distributions Menu
      • Birthday Problem dialog menu available corresponding to the stats functions pbirthday() and qbirthday(). These two functions were modified and named pbirthday.ipsur() and qbirthday.ipsur() to give the precise answer (instead of asymptotics) in the case of exactly two coincidences.
      • All parameter fields for each distribution are modified so that mathematical expressions may be entered, instead of the default restriction that only numbers may be entered into fields. For this reason, we may enter p=1/3 in the binomial menu or perhaps sd=7/sqrt(31) in a standard deviation field.
      • Each distribution has a Simulate variates... option. Simulations are generated "transposed" from the standard Commander, with rows as observations and columns as samples. The name of each generated column by default takes the form "model.sim#", for example, binom.sim1, binom.sim2, etc. The columns may be stored in the Active Data Set or in a new dataset (of customizable length) by default named "Simset". All parameters for the respective distributions are available. To generate multiple columns, a "Number of samples " field is available.
      • The discrete uniform distribution is available for simulation.
      • Noncentrality parameter fields added to Beta, Chisquared, F, and Student's t distribution.
      • A Sampling Distributions menu is added. This is similar to (and indeed was wholly inspired by) the Distributions menu in R Commander, with cosmetic differences to the interface. Rows are samples and columns are observations. Noncentrality parameters are available. A difference is that one may calculate up to three (3) sample statistics of choice from the simulated values, which can be later analyzed with the R Commander. A "discard original observations" option is available in case the number of simulated values is inconveniently large.
      • Discrete CDFs are plotted as right-continuous step functions.
      • Plots for Quantile functions are available, and in the discrete case are left-continuous step functions.
  2. DATA
    1. RcmdrTestDrive: These are simulated data specifically designed to allow the inexperienced user to browse the capabilities of the R Commander. The R Commander has extensive functionality, but many options are unavailable unless the correct types of data are loaded in the Active Data Set. This data set was randomly generated so that, when loaded, essentially all R Commander options would be available for the student to investigate. These data are entirely fictional.
    2. FeedingTimes: Data were collected concerning the feeding times of a recent newborn, Anna Lu Kerns. Variables include age in days, clock time, type of food, the length of feeding time (min), and amount of food (oz). (Only beginning of dataset so far.)
    3. BloodPressure: Data were collected from 2004 through 2006 from Taoying Bian concerning blood pressure and heart rate. Variables include date, time, systolic and diastolic blood pressure, and heart rate.
    1. "Murder Madness in Toon Town". An amusing tale written to relate the fictional variables in the RcmdrTestDrive dataset. An original story contributed by Jeff Cornfield in Summer 2006.

R logo © the R Foundation, http://www.r-project.org