1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
|
#
# Copyright 2007-2019 by the individuals mentioned in the source code history
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
#------------------------------------------------------------------------------
# Author: Michael D. Hunter
# Date: 2014.06.10
# Filename: SummaryCheck.R
# Purpose: Actually test that pieces of MxSummary() do what they're supposed
# to do.
# Procedure: If you're going to muck around with some part of summary,
# first make sure this file tests that part. If this file does not test the
# part of summary that you want to change, then you must add to this test file
# before modifying summary. Once the test has been added, muck around with
# summary to your heart's content.
# TODO add multigroup case to make sure this works
# TODO add model with constraint to make sure it adds back DoF
#------------------------------------------------------------------------------
#------------------------------------------------------------------------------
# Load package and data
require(OpenMx)
data(demoTwoFactor)
#------------------------------------------------------------------------------
# Specify raw data model
vpf <- 5
varnames <- names(demoTwoFactor)
latnames <- c("F1", "F2")
fl <- mxPath(from=rep(latnames, each=vpf), to=varnames, values=.8, labels=paste("load_", varnames, sep=""), free=TRUE)
fc <- mxPath(latnames, connect="unique.pairs", arrows=2, free=c(FALSE, TRUE, FALSE), values=c(1, .5, 1), labels=c("varF1", "covF1F2", "varF2"))
mm <- mxPath(from="one", to=varnames, free=TRUE, values=colMeans(demoTwoFactor), labels=paste("mean_", varnames, sep=""))
rv <- mxPath(from=varnames, arrows=2, free=TRUE, values=.2, labels=paste("resid_", varnames, sep=""))
raw <- mxModel("Raw Test Model to Check MxSummary",
type="RAM", manifestVars=varnames, latentVars=latnames,
fl, fc, mm, rv,
mxData(demoTwoFactor, "raw")
)
raw.fit <- mxRun(raw)
#------------------------------------------------------------------------------
# Specify Saturated raw data model
sat.fit <- mxRefModels(raw.fit, run=TRUE)
cmp <- mxCompare(sat.fit[['Saturated']], raw.fit)
omxCheckCloseEnough(cmp$diffLL[2], 49.764, 1e-2)
omxCheckCloseEnough(cmp$diffdf[2], 34)
omxCheckCloseEnough(cmp$p[2], 0.03961044, 1e-3)
cmp <- mxCompare(raw.fit, raw.fit)
omxCheckTrue(is.na(cmp$p[2]))
cmp <- omxCheckWarning(mxCompare(raw.fit, sat.fit[['Saturated']]),
"Model 'Raw Test Model to Check MxSummary' has more degrees of freedom than Saturated Raw Test Model to Check MxSummary which means that the models need to be compared in the opposite order")
omxCheckTrue(is.na(cmp$p[2]))
omxCheckError(mxCompare(raw.fit, raw), "Model 'Raw Test Model to Check MxSummary' has not been run. Please, mxRun() your model before trying to compare it.")
#------------------------------------------------------------------------------
# Specify a multiple group model
data(demoOneFactor)
manifests <- names(demoOneFactor)
latents <- c("G")
factorModel <- mxModel("OneFactor",
type="RAM",
manifestVars = manifests,
latentVars = latents,
mxPath(from=latents, to=manifests),
mxPath(from=manifests, arrows=2),
mxPath(from=latents, arrows=2, free=FALSE, values=1.0),
mxPath(from = 'one', to = manifests),
mxData(demoOneFactor, type="raw"))
manifests <- manifests[-1]
factorModelLess <- mxModel("OneVarFewer",
type="RAM",
manifestVars = manifests,
latentVars = latents,
mxPath(from=latents, to=manifests),
mxPath(from=manifests, arrows=2),
mxPath(from=latents, arrows=2, free=FALSE, values=1.0),
mxPath(from = 'one', to = manifests),
mxData(demoOneFactor, type="raw"))
mg <- mxModel("bob", factorModel, factorModelLess, mxFitFunctionMultigroup(c("OneFactor", "OneVarFewer")))
mg.fit <- mxRun(mg)
mg.sat <- mxRefModels(mg.fit, run=TRUE)
omxCheckEquals(dim(mxEval(satCov, mg.sat[[1]]$`Saturated OneFactor`))[1], 5)
omxCheckEquals(dim(mxEval(satCov, mg.sat[[1]]$`Saturated OneVarFewer`))[1], 4)
# the saturated model for OneVarFewer should be only on the variables used x2:x5, not all the variables x1:x5.
refStats <- lapply(mg.sat, function(model) {
list(model$output$Minus2LogLikelihood, summary(model)$degreesOfFreedom)
})
refStats$Independence[[1]] <- mg.fit$output$fit - 100
ign <- omxCheckWarning(summary(mg.fit, refModels=refStats),
"Your model may be mis-specified (and fit worse than an independence model), or you may be using the wrong independence model, see ?mxRefModels")
#------------------------------------------------------------------------------
# Specify a multiple group cov model
data(demoOneFactor)
latents = c("G")
manifests = names(demoOneFactor)
m1 <- mxModel("model1", type = "RAM",
manifestVars = manifests, latentVars = latents,
mxPath(from = latents, to = manifests),
mxPath(from = manifests, arrows = 2),
mxPath(from = latents, arrows = 2, free = F, values = 1.0),
mxData(cov(demoOneFactor), type = "cov", numObs = 500)
)
m2 <- mxModel("model2", type = "RAM",
manifestVars = manifests, latentVars = latents,
mxPath(from = latents, to = manifests),
mxPath(from = manifests, arrows = 2),
mxPath(from = latents, arrows = 2, free = F, values = 1.0),
mxData(cov(demoOneFactor), type = "cov", numObs = 500)
)
m3 = mxModel("bob", m1, m2, mxFitFunctionMultigroup(c("model1","model2")))
m3r = mxRun(m3)
ref = mxRefModels(m3r, run=TRUE)
expecteError = "List of length two (illegal argument) passed to 'SaturatedLikelihood' argument of summary function. You probably meant to use the refModels argument instead."
omxCheckError(summary(m3r, SaturatedLikelihood=ref), expecteError)
# TODO add checking of fit stats for this model
#------------------------------------------------------------------------------
# Specify cov data models
cov <- mxModel("Covariance Test Model to Check MxSummary", type="RAM",
manifestVars=varnames, latentVars=latnames,
fl, fc, rv,
mxData(cov(demoTwoFactor), "cov", numObs=nrow(demoTwoFactor))
)
cov.fit = mxRun(cov)
cov.sat = mxRefModels(cov.fit, run=TRUE)
covm = mxModel("Covariance and Means Test Model to Check MxSummary", type="RAM",
manifestVars=varnames, latentVars=latnames,
fl, fc, rv, mm,
mxData(cov(demoTwoFactor), "cov", numObs=nrow(demoTwoFactor), means=colMeans(demoTwoFactor))
)
covm.fit = mxRun(covm)
#------------------------------------------------------------------------------
# Get summaries of models
sat.sum <- summary(sat.fit[[1]])
raw.sum <- summary(raw.fit)
raws.sum <- summary(raw.fit, SaturatedLikelihood=sat.fit[[1]])
rawr.sum <- summary(raw.fit, refModels=sat.fit)
cov.sum <- summary(cov.fit)
covs.sum <- summary(covm.fit)
mg.sum <- summary(mg.fit, refModels=mg.sat)
#------------------------------------------------------------------------------
# Compare Fit statistics to what they should be.
# Things to check for raw w/o sat, raw w/ sat, cov, cov & means, multi-group
# number of observed statistics
# degrees of freedom
# -2LL, sat -2LL, numObs
# chi-square, chi dof, chi p
# AIC df AIV, param, BIC df, BIC param, BIC sample size
# CFI, TLI, RMSEA
# number of observed statistics
omxCheckEquals(raw.sum$observedStatistics, 5000)
omxCheckEquals(raws.sum$observedStatistics, 5000)
omxCheckEquals(cov.sum$observedStatistics, 55)
omxCheckEquals(covs.sum$observedStatistics, 65)
omxCheckEquals(mg.sum$observedStatistics, 5*500 + 4*500)
# degrees of freedom
# TODO add model with constraint to make sure it adds back DoF
ep <- 31
omxCheckEquals(raw.sum$degreesOfFreedom, 5000-ep)
omxCheckEquals(raws.sum$degreesOfFreedom, 5000-ep)
omxCheckEquals(sat.sum$degreesOfFreedom, 5000-sat.sum$estimatedParameters)
omxCheckEquals(raws.sum$satDoF, 5000-sat.sum$estimatedParameters)
omxCheckEquals(cov.sum$degreesOfFreedom, 55-(ep-10)) #no means estimated
omxCheckEquals(covs.sum$degreesOfFreedom, 65-ep)
omxCheckEquals(mg.sum$degreesOfFreedom, 5*500+4*500-27)
# -2LL, sat -2LL, numObs
omxCheckWithinPercentError(raw.sum$Minus2LogLikelihood, 9236.675, percent=1e-4)
omxCheckTrue(is.na(raw.sum$SaturatedLikelihood))
omxCheckWithinPercentError(raws.sum$SaturatedLikelihood, 9186.911, percent=1e-4)
omxCheckWithinPercentError(cov.sum$SaturatedLikelihood, -2.464397, percent=1e-4)
omxCheckWithinPercentError(covs.sum$SaturatedLikelihood, -2.464397, percent=1e-4)
omxCheckEquals(raw.sum$numObs, 500)
omxCheckEquals(raws.sum$numObs, 500)
omxCheckEquals(cov.sum$numObs, 500)
omxCheckEquals(covs.sum$numObs, 500)
omxCheckEquals(mg.sum$numObs, 1000)
# chi-square, chi dof, chi p
omxCheckTrue(is.na(raw.sum$Chi))
omxCheckTrue(!is.na(raws.sum$Chi))
omxCheckCloseEnough(cov.sum$Chi, 49.75399, epsilon=1e-3)
omxCheckCloseEnough(covs.sum$Chi, 49.75399, epsilon=1e-3)
# ChiDoF = df - sat.df, i.e. df = obsStat-ep, sat.df = obsStat-sat.ep, so ChiDoF = sat.ep - ep
chi.df = 34
# line asserting ChiDoF is na is inconsistent with the next, that it = chi.df...
# omxCheckTrue(is.na(raw.sum$ChiDoF)) # is still populated, but it reported as NA when satLik is NA
omxCheckEquals(raws.sum$ChiDoF, chi.df)
omxCheckEquals(cov.sum$ChiDoF, chi.df)
omxCheckEquals(covs.sum$ChiDoF, chi.df)
omxCheckEquals(mg.sum$ChiDoF, 7)
omxCheckCloseEnough(mg.sum$Chi, 13.1811, 0.001)
omxCheckTrue(is.na(raw.sum$p))
omxCheckCloseEnough(raws.sum$p, pchisq(raws.sum$Chi, raws.sum$ChiDoF, lower.tail=FALSE), 1e-8)
omxCheckCloseEnough(cov.sum$p, pchisq(cov.sum$Chi, cov.sum$ChiDoF, lower.tail=FALSE), 1e-8)
omxCheckCloseEnough(covs.sum$p, pchisq(covs.sum$Chi, covs.sum$ChiDoF, lower.tail=FALSE), 1e-8)
# AIC df, AIC param, BIC df, BIC param, BIC sample size
# CFI, TLI, RMSEA
# RMSEA
omxCheckTrue(is.na(raw.sum$RMSEA))
omxCheckTrue(!is.na(raws.sum$RMSEA))
omxCheckCloseEnough(raws.sum$RMSEA, 0.03045151, .001)
omxCheckCloseEnough(raws.sum$RMSEACI, c(0, 0.05064688), .001)
omxCheckCloseEnough(cov.sum$RMSEA, 0.03035523, .001)
omxCheckCloseEnough(cov.sum$RMSEACI, c(0, 0.05056999), .001)
omxCheckCloseEnough(mg.sum$RMSEA, 0.02971556, .001)
omxCheckCloseEnough(mg.sum$RMSEACI, c(0, 0.0582353), .001)
# CFI
omxCheckTrue(is.na(raw.sum$CFI))
omxCheckCloseEnough(cov.sum$CFI, 0.9978234, .001)
omxCheckCloseEnough(rawr.sum$CFI, 0.997814, .001)
omxCheckCloseEnough(mg.sum$CFI, 0.999077, .001)
# TLI
omxCheckTrue(is.na(raw.sum$TLI))
omxCheckCloseEnough(cov.sum$TLI, 0.9971192, .001)
omxCheckCloseEnough(rawr.sum$TLI, 0.9971067, .001)
omxCheckCloseEnough(mg.sum$TLI, 0.9978904, .001)
# Information Criteria
omxCheckCloseEnough(raw.sum$informationCriteria['AIC:','par'], 9298.67, .01)
omxCheckCloseEnough(raw.sum$informationCriteria['BIC:','par'], 9429.32, .01)
omxCheckCloseEnough(sat.sum$informationCriteria['AIC:','par'], 9316.91, .01)
omxCheckCloseEnough(sat.sum$informationCriteria['BIC:','par'], 9590.86, .01)
omxCheckCloseEnough(raws.sum$informationCriteria['AIC:','par'], 9298.67, .01)
omxCheckCloseEnough(raws.sum$informationCriteria['BIC:','par'], 9429.32, .01)
omxCheckCloseEnough(cov.sum$informationCriteria['AIC:','par'], 91.75, .01)
omxCheckCloseEnough(cov.sum$informationCriteria['BIC:','par'], 180.26, .01)
omxCheckCloseEnough(covs.sum$informationCriteria['AIC:','par'], 111.75, .01)
omxCheckCloseEnough(covs.sum$informationCriteria['BIC:','par'], 242.40, .01)
omxCheckCloseEnough(mg.sum$informationCriteria[c(1:4,6)], c(-6936.958, -28889.347, 2063.042, 2195.552, 2109.798), .01)
# -----------------------------------
library(OpenMx)
data(demoOneFactor)
manifests <- names(demoOneFactor)
latents <- "G1"
fit1 <- mxRun(mxModel(model="One Factor", type="RAM",
manifestVars = manifests, latentVars = latents,
mxPath(from = latents, to=manifests),
mxPath(from = manifests, arrows = 2),
mxPath(from = latents, arrows = 2, free = FALSE, values = 1.0),
mxPath(from = 'one', manifests, free=FALSE),
mxData(cov(demoOneFactor), colMeans(demoOneFactor), type = "cov", numObs = 500)
))
fit2 <- mxRun(mxModel(model="One Factor", type="RAM",
manifestVars = manifests, latentVars = latents,
mxPath(from = latents, to=manifests, values=0, free=c(F,T,T,T,T)),
mxPath(from = manifests, arrows = 2),
mxPath(from = latents, arrows = 2, free = FALSE, values = 1.0),
mxPath(from = 'one', manifests, free=FALSE),
mxData(cov(demoOneFactor), colMeans(demoOneFactor), type = "cov", numObs = 500)
))
got <- mxCompare(fit1, fit2, boot=T, replications = 10)
omxCheckCloseEnough(got[2,'p'], 0, .01)
|
