The following is a summary of all the missing data supplemental functions provided in the Real Statistics Resource Pack in conjunction with Multiple Imputation (MI) and Full Information Maximum Likelihood (FIML).
The following are array functions (with the exception of CountPatterns, which is not an array function) where R1 is a range where each column represents sample data for a random variable. Blank or non-numeric elements are assumed to represent missing data.
MissingFreq(R1, head) – generates a summary with the frequency of non-missing data in R1
MissingPatterns(R1, head, s) – generates a summary with missing patterns of data in R1. The argument s is used to fill up any extra rows in the output which do not contain data. If s is not specified then it defaults to the error value #N/A.
CountPatterns(R1) = the number of different missing data patterns there are for the data in R1
ImputeSimple(R1, head, R2, iter) – generates a range with all the missing data in R1 filled in using the simple imputation approach
ImputeReg(R1, j, head, R2, iter) – generates a range where all the missing data in column j is filled in using one step of the FCS algorithm
ImputeFCS(R1, head, iterf, R2, iter) – generates one imputation of the missing data in R1 using FCS; iterf is the number of iterations of the FCS algorithm (default = 20)
If head is TRUE (the default) then it is assumed that the data range R1 as well as the output contain column headings, while if head = FALSE then the R1 should not contain column headings and the output will not contain column headings either.
R2 is a range containing constraints (if R2 is omitted then no constraints are used) and iter is the maximum number of iterations used to obtain a value within the min/max constraints (default = 25). ImputeReg can contain only one constraint, namely for the variable corresponding to column j.
For the following array functions R1 and head are as in Group 1, except that the default for head is FALSE. If lab = TRUE then an extra column is inserted in the output which contains labels (default = FALSE).
MISummary(R1, head) – generates a compact summary of the regression model for R1 where the last column in R1 is assumed to be the data for the y variable and the other columns are assumed to contain the data for the x variables.
ImputedData(R0, R1, head) – generates a range with all the elements in R1 where there is missing data in range R0.
DescStats(R1, lab, head) – generates a mini descriptive statistics report for the data in range R1.
DescStats(R1, lab, head, R0) – generates a mini descriptive statistics report for the data in range R1 which correspond to the missing data in range R0.
MissingPairwise(R1, head) – generates a summary of the percentage of non-missing data for each pair of variables in R1. If head = TRUE then the output includes row headings (as well as column headings).
The following is an array function:
MICombine(R1, nimp, ncols, head, raw) – generates a combined compact regression summary derived from the compact summaries of nimp imputations if raw = FALSE or derived directly from the nimp imputations if raw = TRUE (default).
If raw = FALSE then R1 is the range containing the first of the nimp compact regression summaries, while if raw = TRUE then R1 is the range containing the first of the nimp imputations. The nimp imputations or compact regression summaries are separated by ncols blank columns (default = 1). head is as in Group 2.
For the following function, R1 is a 1 × k row vector, R2 is a k × k covariance matrix and R3 is a 1 × k mean vector.
LLReg(R1, R2, R3) = the value of -2LLi for the row represented by R1, possibly containing missing data.
For the following array functions, R1 is a 2m × n range where the first m rows represent the values of m imputed population parameters and the second m rows represent the corresponding standard errors for these parameters. The columns represent separate imputations. The argument size is the number of elements in the original sample (including missing data) and head and lab are as in Group 2.
ImputeVar(R1, size, lab, head) – outputs a a range summarizing the combination rules for variance.
ImputeParam(R1, size, lab, head, alpha) – outputs a range based on the combination rules and the usual t test using the stated value of alpha (default = .05).