Subclassification
In matchit, setting method = "subclass" performs subclassification on the distance measure (i.e., propensity score). Treatment and control units are placed into subclasses based on quantiles of the propensity score in the treated group, in the control group, or overall, depending on the desired estimand. Weights are computed based on the proportion of treated units in each subclass. Subclassification implemented here does not rely on any other package.
This page details the allowable arguments with method = "subclass". See matchit for an explanation of what each argument means in a general context and how it can be specified.
Below is how matchit is used for subclassification:
matchit(formula, data = NULL, method = "subclass",
distance = "glm", link = "logit",
distance.options = list(), estimand = "ATT",
discard = "none", reestimate = FALSE,
s.weights = NULL, verbose = FALSE, ...)formula |
a two-sided |
data |
a data frame containing the variables named in |
method |
set here to |
distance |
the distance measure to be used. See |
link |
when |
distance.options |
a named list containing additional arguments supplied to the function that estimates the distance measure as determined by the argument to |
estimand |
the target |
discard |
a string containing a method for discarding units outside a region of common support. |
reestimate |
if |
s.weights |
the variable containing sampling weights to be incorporated into propensity score models and balance statistics. |
verbose |
|
... |
additional arguments that control the subclassification:
|
The arguments exact, mahvars, replace, m.order, caliper (and related arguments), and ratio are ignored with a warning.
After subclassification, effect estimates can be computed separately in the subclasses and combined, or a single marginal effect can be estimated by using the weights in the full sample. When using the weights, the method is sometimes referred to as marginal mean weighting through stratification (MMWS; Hong, 2010) or fine stratification weighting (Desai et al., 2017). The weights can be interpreted just like inverse probability weights.
Changing min.n can change the quality of the weights. Generally, a low min.w will yield better balance because subclasses only contain units with relatively similar distance values, but may yield higher variance because extreme weights can occur due to there being few members of a treatment group in some subclasses.
Note that subclassification weights can also be estimated using WeightIt, which provides some additional methods for estimating propensity scores. Where propensity score-estimation methods overlap, both packages will yield the same weights.
All outputs described in matchit are returned with method = "subclass" except that match.matrix is excluded and two additional component, q.cut and qn, are included, containing a vector of the distance measure cutpoints used to define the subclasses and a matrix of the subclass sample sizes, respectively. Note that when min.n > 0, the subclass assignments may not strictly obey the quantiles listed in q.cut.
In a manuscript, you don't need to cite another package when using method = "subclass" because the subclassification is performed completely within MatchIt. For example, a sentence might read:
Propensity score subclassification was performed using the MatchIt package (Ho, Imai, King, & Stuart, 2011) in R.
It may be a good idea to cite Hong (2010) or Desai et al. (2017) if the treatment effect is estimated using the subclassification weights.
Desai, R. J., Rothman, K. J., Bateman, B. . T., Hernandez-Diaz, S., & Huybrechts, K. F. (2017). A Propensity-score-based Fine Stratification Approach for Confounding Adjustment When Exposure Is Infrequent: Epidemiology, 28(2), 249–257. doi: 10.1097/EDE.0000000000000595
Hong, G. (2010). Marginal mean weighting through stratification: Adjustment for selection bias in multilevel data. Journal of Educational and Behavioral Statistics, 35(5), 499–531. doi: 10.3102/1076998609359785
matchit for a detailed explanation of the inputs and outputs of a call to matchit.
method_full for optimal full matching, which is similar to subclassification except that the number of subclasses and subclass membership are chosen to optimize the within-subclass distance.
data("lalonde")
# PS subclassification for the ATT with 7 subclasses
s.out1 <- matchit(treat ~ age + educ + race + nodegree +
married + re74 + re75, data = lalonde,
method = "subclass", subclass = 7)
s.out1
summary(s.out1, subclass = TRUE)
# PS subclassification for the ATE with 10 subclasses
# and at least 2 units in each group per subclass
s.out2 <- matchit(treat ~ age + educ + race + nodegree +
married + re74 + re75, data = lalonde,
method = "subclass", subclass = 10,
estimand = "ATE", min.n = 2)
s.out2
summary(s.out2)Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.