MAGMA.R (version 1.0.3)

MAGMA: MAGMA

Description

This function conducts many group matching for 2 to 4 groups. It augments the original data set by the relevant 'MAGMA.R' variables. For details, see below.

Usage

MAGMA(Data, group, dist, cores = 1, verbose = TRUE)

Value

Your input data frame augmented with matching relevant variables, namely weight, step, distance, and ID. In case of missing values on the distance or group variable, MAGMA excludes them for the matching process. The returned data set does not contain those excluded cases. For more information, see Details.

Arguments

Data

A data frame or tibble containing at least your grouping and distance variable. Data needs to be specified in your environment.

group

A character specifying the name of your grouping variable in the data. Note that MAGMA can only match your data for a maximum of 4 groups. Matching over two grouping variables (e.g., 2x2 Design) is possible by specifying group as a character vector with a length of two. In this case, each or the 2 grouping variables can only have two levels.

dist

A character specifying the name of your distance variable in data.

cores

An integer defining the number of cores used for parallel computation.

verbose

TRUE or FALSE indicating whether matching information should be printed to the console.

Author

Julian Urban

Details

This function conducts nearest neighbor many group matching. It is applicable for two to four groups or a 2x2 design. As output, this function augments your original data by the variables weight, step, distance, and ID. Weight indicates whether a case was matched. Step specifies the iteration in which a case was matched. It also shows which cases were matched together. Distance indicates the mean difference within a match. Since matches with a lower distance are matched in an earlier iteration, step and distance are strongly correlated. This function has some CPU and RAM load. In most four-group applications and three-group applications with large sample size, RAM may be not sufficient. Therefore MAGMA switches to random quasi-systematic matching. If this is the case, MAGMA informs you. The output of the function does not change, but balance might be slightly affected.

Examples

# Running this code will take a while
# Two-group exact matching using the data set 'MAGMA_sim_data'
# Matching variable 'gifted_support' (received giftedness support yes or no)
# 'MAGMA_sim_data_gifted' contains the result of the matching
MAGMA_sim_data_gifted <- MAGMA(Data = MAGMA_sim_data,
                                group = "gifted_support",
                                dist = "ps_gifted",
                                cores = 1)
head(MAGMA_sim_data_gifted)

# \donttest{
# Two-group exact matching using the data set 'MAGMA_sim_data'
# Matching variable 'teacher_ability_rating' (ability rated from teacher as
# below average, average, or above average)
# MAGMA_sim_data_tar' contains the result of the matching
# Cores per default = 1
MAGMA_sim_data_tar <- MAGMA(Data = MAGMA_sim_data,
                            group = "teacher_ability_rating",
                            dist = "ps_tar")
head(MAGMA_sim_data_tar)

# 2x2 matching using the data set 'MAGMA_sim_data'
# Matching variables are 'gifted_support' (received giftedness support yes
# or no) and 'enrichment' (participated in enrichment or not)
# 'MAGMA_sim_data_gift_enrich' contains the result of the matching
# 2x2 matching is equal to four-group matching
MAGMA_sim_data_gift_enrich <- MAGMA(Data = MAGMA_sim_data,
                                   group = c("gifted_support", "enrichment"),
                                   dist = "ps_2x2",
                                   cores = 2)
head(MAGMA_sim_data_gift_enrich)
# }