| mxPowerSearch {OpenMx} | R Documentation |
Power curve
Description
mxPower is used to evaluate the power to distinguish between a hypothesized effect (trueModel)
and a model where this effect is set to the value of the null hypothesis (falseModel).
mxPowerSearch evaluates power across a range of sample sizes or effect sizes,
choosing intelligent values of, for example, sample size to explore.
Both functions are flexible, with multiple options to control n, sig.level, method, and
other factors. Several parameters can take a vector as input. These options are described in detail below.
Evaluation can either use the non-centrality parameter (which can be very quick) or conduct an empirical evaluation.
note: During longer evaluations, the functions printout helpful progress consisting of a note about what task is being conducted, e.g. "Search n:power relationship for 'A11'" and, beneath that, an update on the run, the model being evaluated, and the current candidate value being considered, e.g. "R 15 1p N 79"
Usage
mxPowerSearch(trueModel, falseModel, n=NULL, sig.level=0.05, ...,
probes=300L, previousRun=NULL,
gdFun=mxGenerateData,
method=c('empirical', 'ncp'),
grid=NULL,
statistic=c('LRT','AIC','BIC'),
OK=mxOption(trueModel, "Status OK"),
checkHess=FALSE,
silent=!interactive())
mxPower(trueModel, falseModel, n=NULL, sig.level=0.05, power=0.8, ...,
probes=300L, gdFun=mxGenerateData,
method=c('empirical', 'ncp'),
statistic=c('LRT','AIC','BIC'),
OK=mxOption(trueModel, "Status OK"), checkHess=FALSE,
silent=!interactive())
Arguments
trueModel |
The true generating model for the data. |
falseModel |
Model representing the null hypothesis that we wish to reject. |
n |
Total rows summing across all submodels proportioned as given in the trueModel. Default = NULL. If provided, result (power or alpha) solved-for at the given total sample size. |
sig.level |
A single value for the p-value (aka false positive or type-1 error rate). Default = .05. |
power |
One or values for power (a.k.a. 1 - type 2 error) to evaluate. Default = .80 |
... |
Not used. |
probes |
The number of probes to use when method = ‘empirical’. |
previousRun |
Output from a prior run of ‘mxPowerSearch’ to build on. |
gdFun |
The function invoked to generate new data for each Monte Carlo probe. Default = |
method |
To estimate power: ‘empirical’ (Monte Carlo method) or ‘ncp’ (average non-centrality method). |
grid |
A vector of locations at which to evaluate the power. If not provided, a reasonable default will be chosen. |
statistic |
Which test to use to compare models (Default = 'LRT'). |
OK |
The set of status codes that are considered successful. |
checkHess |
Whether to approximate the Hessian in each replication. |
silent |
Whether to show a progress indicator. |
Details
Power is the chance of obtaining a significant difference when there
is a true difference, i.e., (1 - false negative rate).
The likelihood ratio test is used by default. There are two
methods available to produce a power curve. The default,
method = empirical, works for any model where the likelihood
ratio test works. For example, definition variables and missing data
are fine, but parameters estimated at upper or lower bounds will cause
problems. The method = empirical can require a lot of time
because the models need to be fit 100s of times. An alternate
approach, method = ncp, is much more efficient and takes
advantage of the fact that the non-null distribution of likelihood
ratio test statistic is often \chi^2(df_1 - df_0, N \lambda).
That is, the non-centrality parameter, \lambda (lambda),
can be assumed, on average, to contribute equally per row. This permits essentially
instant power curves without the burden of tedious simulation. However, definition
variables, missing data, or unconventional models (e.g., mixture distributions) can
violate this assumption. Therefore, we recommend verifying that the output from
method = empirical roughly matches method = ncp on the model of interest
before relying on method = ncp.
note: Unlike method = empirical, method = ncp does not use the
gdFun argument and can be used with models that have summary statistics rather than raw data.
When method = ncp, parameters of both ‘trueModel’ and
‘falseModel’ are assumed to be converged to their desired values.
In contrast, when method = empirical, ‘trueModel’
need not be run or even contain data. On each replication,
data are generated from ‘trueModel’ at the given parameter
vector. Then both ‘trueModel’ and ‘falseModel’ are
fit against these data.
When statistic = 'LRT' then the models must be nested and
sig.level is used to determine whether the test is rejected.
For ‘AIC’ and ‘BIC’, the test is regarded as rejected
when the ‘trueModel’ obtains a lower score than the
‘falseModel’. In contrast to statistic='LRT', there is
no nesting requirement. For example, ‘AIC’ can be used to
compare a ‘trueModel’ against its corresponding saturated
model.
mxPower operates in many modes.
When power is passed as NULL then power is calculated and returned.
When power (as a scalar or vector) is given then sample or effect size is (are) returned.
If you pass a list of models for ‘falseModel’, each model
will be checked in turn and the results returned as a vector.
If you pass a vector of sample sizes, each sample size will
be checked in turn and the results returns as a vector.
mxPowerSearch
In contrast to mxPower, mxPowerSearch attempts to model
the whole relationship between sample size or effect size and power.
A naive Monte Carlo estimate of power examines a single candidate
sample size at a time. To obtain the whole curve, and simultaneously,
to reduce the number of simulation probes, mxPowerSearch
employs a binomial family generalized linear model with a logit link
to predict the power curve across the sample sizes of interest
(similar to Schoemann et al, 2014).
The mxPowerSearch algorithm works in 3
stages. Without loss of generality, our description will use the
sample size to power relationship, but a similar process is used when evaluating
the relationship of parameter value to power. In the first stage, a crude
binary search is used to find the range reasonable values for N. This
stage is complete once we have at least two rejections and at least two
non-rejections. At this point, the binomial intercept and slope model
is fit to these data. If the p-value for the slope is less than
0.25 then we jump to stage 3. Otherwise, we fit an intercept only
binomial model. Our goal is to nail down the intercept (where
power=0.5) because this is the easiest point to find and is a
necessary prerequisite to estimate the variance (a.k.a. slope).
Therefore, we probe at the median of previous probes stepping by 10%
in the direction of the model's predicted intercept. Eventually, after enough
probes, we reach stage 3 where the p-value for the slope is
less than 0.25. At stage 3, our goal is to nail down the interesting
part of the power curve. Therefore, we cycle serially through probes
at 0, 1, and 2 logits from the intercept. This process is continued
for the permitted number of probes.
Occasionally, the p-value for the slope in the stage 3 model
grows larger than 0.25. In this case, we switch back to stage 2
(intercept only) until the stage 3 model start working again.
There are no other convergence criteria. Accuracy continues to improves until the probe
limit is reached.
note: After mxPowerSearch returns, you might find you wanted to run
additional probes (i.e., bounds are wider than you'd like). You can run more probes
without starting from scratch by passing the previous result back into the function using the
previousRun argument.
When ‘n’ is fixed then mxPowerSearch helps answer the
question, “how small of a true effect is likely to be detected
at a particular sample size?” Only one parameter can be considered at
a time. The way the simulation works is that a candidate value for
the parameter of interest is loaded into the trueModel, data
are generated, then both the true and false model are fit to the data to
obtain the difference in fit. The candidate parameter is initially set
to halfway between the trueModel and falseModel. The
power curve will reflect the smallest distance, relative to the false
model, required to have a good chance to reject the false model
according to the chosen statistic.
Note that the default grid is chosen to show the interesting
part of the power curve (from 0.25 to 0.97).
Especially for method=ncp, this curve is practically
identical for any pair of models (but located at a different range of
sample sizes). However, if you wish to align power curves from more than one
power analysis, you should select your own grid points (perhaps pass in the
power array from the first to subsequent using grid = ).
Note: CI is not given for method=ncp: The estimates are exact (to the precision of the true and null/false model solutions provided by the user).
Value
mxPower returns a vector of sample sizes, powers, or effect sizes.
mxPowerSearch returns a data.frame with one row for each ‘grid’ point. The
first column is either the sample size ‘N’ (given as the
proportion of rows provided in trueModel) or the parameter
label. The second column is the power. If method=empirical, lower
and upper provide a +/-2 SE confidence interval (CI95)
for the power (as estimated by the binomial logit linear model).
When method=empirical then the ‘probes’ attribute
contains a data.frame record of the power estimation process.
References
Schoemann, A. M., Miller, P., Pornprasertmanit, S. & Wu, W. (2014). Using Monte Carlo simulations to determine power and sample size for planned missing designs. International Journal of Behavioral Development, 38, 471-479.
See Also
Examples
library(OpenMx)
# Make a 1-factor model of 5 correlated variables
data(demoOneFactor)
manifests <- names(demoOneFactor)
latents <- c("G")
factorModel <- mxModel("One Factor", type="RAM",
manifestVars = manifests,
latentVars = latents,
mxPath(from= latents, to= manifests, values= 0.8),
mxPath(from= manifests, arrows= 2,values= 1),
mxPath(from= latents, arrows= 2, free= FALSE, values= 1),
mxPath(from= "one", to= manifests),
mxData(demoOneFactor, type= "raw")
)
factorModelFit <- mxRun(factorModel)
# The loading of x1 on G is estimated ~ 0.39
# Let's test fixing it to .3
indModel <- factorModelFit
indModel$A$values['x1','G'] <- 0.3
indModel$A$free['x1','G'] <- FALSE
indModel <- mxRun(indModel)
# What power do we have at different sample sizes
# to detect that the G to x1 factor loading is
# really 0.3 instead of 0.39?
mxPowerSearch(factorModelFit, indModel, method='ncp')
# If we want to conduct a study with 80% power to
# find that the G to x1 factor loading is
# really 0.3 instead of 0.39, what sample size
# should we use?
mxPower(factorModelFit, indModel, method='ncp')