Become an expert in R — Interactive courses, Cheat Sheets, certificates and more!
Get Started for Free

baggr

Bayesian aggregate treatment effects model


Description

Bayesian inference on parameters of an average treatment effects model that's appropriate to the supplied individual- or group-level data, using Hamiltonian Monte Carlo in Stan. (For overall package help file see baggr-package)

Usage

baggr(
  data,
  model = NULL,
  pooling = "partial",
  effect = NULL,
  covariates = c(),
  prior_hypermean = NULL,
  prior_hypersd = NULL,
  prior_hypercor = NULL,
  prior_beta = NULL,
  prior = NULL,
  ppd = FALSE,
  test_data = NULL,
  quantiles = seq(0.05, 0.95, 0.1),
  outcome = "outcome",
  group = "group",
  treatment = "treatment",
  silent = FALSE,
  warn = TRUE,
  ...
)

Arguments

data

data frame with summary or individual level data to meta-analyse

model

if NULL, detected automatically from input data otherwise choose from "rubin", "mutau", "individual", "quantiles" (see Details).

pooling

Type of pooling; choose from "none", "partial" (default) and "full". If you are not familiar with the terms, consult the vignette; "partial" can be understood as random effects and "full" as fixed effects

effect

Label for effect. Will default to "mean" in most cases, "log OR" in logistic model, quantiles in quantiles model etc. These labels are used in various print and plot outputs. Comparable models (e.g. in baggr_compare) should have same effect.

covariates

Character vector with column names in data. The corresponding columns are used as covariates (fixed effects) in the meta-regression model (in case of aggregate data). In the case of individual level data the model does not differentiate between group-level variables (same values of the covariate for all rows related to a given group) and individual-level covariates.

prior_hypermean

prior distribution for hypermean; you can use "plain text" notation like prior_hypermean=normal(0,100) or uniform(-10, 10). See Details:Priors below for more possible specifications. If unspecified, the priors will be derived automatically based on data (and printed out in the console).

prior_hypersd

prior for hyper-standard deviation, used by Rubin and "mutau"`` models; same rules apply as for _hypermean';

prior_hypercor

prior for hypercorrelation matrix, used by the "mutau" model

prior_beta

prior for regression coefficients if covariates are specified; will default to experimental normal(0, 10^2) distribution

prior

alternative way to specify all priors as a named list with hypermean, hypersd, hypercor, beta, analogous to prior_ arguments above, e.g. prior = list(hypermean = normal(0,10), beta = uniform(-50, 50))

ppd

logical; use prior predictive distribution? (p.p.d.) Default is no. If ppd=TRUE, Stan model will sample from the prior distributions and ignore data in inference. However, data argument might still be used to infer the correct model and to set the default priors.

test_data

data for cross-validation; NULL for no validation, otherwise a data frame with the same columns as data argument

quantiles

if model = "quantiles", a vector indicating which quantiles of data to use (with values between 0 and 1)

outcome

character; column name in (individual-level) data with outcome variable values

group

character; column name in data with grouping factor; it's necessary for individual-level data, for summarised data it will be used as labels for groups when displaying results

treatment

character; column name in (individual-level) data with treatment factor;

silent

Whether to silence messages about prior settings and about other automatic behaviour.

warn

print an additional warning if Rhat exceeds 1.05

...

extra options passed to Stan function, e.g. control = list(adapt_delta = 0.99), number of iterations etc.

Details

Running baggr requires 1/ data preparation, 2/ choice of model, 3/ choice of priors. All three are discussed in depth in the package vignette (vignette("baggr")).

Data. For aggregate data models you need a data frame with columns tau and se or tau, mu, se.tau, se.mu. An additional column can be used to provide labels for each group (by default column group is used if available, but this can be customised – see the example below). For individual level data three columns are needed: outcome, treatment, group. These are identified by using the outcome, treatment and group arguments.

Many data preparation steps can be done through a helper function prepare_ma. It can convert individual to summary-level data, calculate odds/risk ratios (with/without corrections) in binary data, standardise variables and more. Using it will automatically format data inputs to work with baggr().

Models. Available models are:

  • for the continuous variable means: "rubin" model for average treatment effect, "mutau" version which takes into account means of control groups, "full", which works with individual-level data

  • for continuous variable quantiles: '"quantiles"“ model (see Meager, 2019 in references)

  • for binary data: "logit" model can be used on individual-level data; you can also analyse continuous statistics such as log odds ratios and logs risk ratios using the models listed above; see vignette("baggr_binary") for tutorial with examples

If no model is specified, the function tries to infer the appropriate model automatically. Additionally, the user must specify type of pooling. The default is always partial pooling.

Covariates. Both aggregate and individual-level data can include extra columns, given by covariates argument (specified as a character vector of column names) to be used in regression models. We also refer to impact of these covariates as fixed effects.

Two types of covariates may be present in your data:

  • In "rubin" and "mutau" models, covariates that change according to group unit. In that case, the model accounting for the group covariates is a meta-regression model. It can be modelled on summary-level data.

  • In "logit" and "full" models, covariates that change according to individual unit. Then, the model can be called a mixed model . It has to be fitted to individual-level data. Note that the first case can also be accounted for by using a mixed model.

Priors. It is optional to specify priors yourself, as the package will try propose an appropriate prior for the input data if you do not pass a prior argument. To set the priors yourself, use prior_ arguments. For specifying many priors at once (or re-using between models), a single prior = list(...) argument can be used instead. Appropriate examples are given in vignette("baggr").

Outputs. By default, some outputs are printed. There is also a plot method for baggr objects which you can access via baggr_plot (or simply plot()). Other standard functions for working with baggr object are

Value

baggr class structure: a list including Stan model fit alongside input data, pooling metrics, various model properties. If test data is used, mean value of -2*lpd is reported as mean_lpd

Author(s)

Witold Wiecek, Rachael Meager

Examples

df_pooled <- data.frame("tau" = c(1, -1, .5, -.5, .7, -.7, 1.3, -1.3),
                        "se" = rep(1, 8),
                        "state" = datasets::state.name[1:8])
baggr(df_pooled) #baggr automatically detects the input data
# same model, but with correct labels,
# different pooling & passing some options to Stan
baggr(df_pooled, group = "state", pooling = "full", iter = 500)
# model with different (very informative) priors
baggr(df_pooled, prior_hypersd = normal(0, 2))


# "mu & tau" model, using a built-in dataset
# prepare_ma() can summarise individual-level data
ms <- microcredit_simplified
ms$outcome <- microcredit_simplified$consumerdurables + 1
microcredit_summary_data <- prepare_ma(ms)
baggr(microcredit_summary_data, model = "mutau",
      pooling = "partial", prior_hypercor = lkj(1),
      prior_hypersd = normal(0,10),
      prior_hypermean = multinormal(c(0,0),matrix(c(10,3,3,10),2,2)))

baggr

Bayesian Aggregate Treatment Effects

v0.4.0
GPL (>= 3)
Authors
Witold Wiecek [cre, aut], Rachael Meager [aut], Brice Green [ctb] (predict(), loo_compare, many visuals), Trustees of Columbia University [cph] (tools/make_cc.R)
Initial release

We don't support your browser anymore

Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.