bug since new version?: se=robust.cluster

[Stephan Junker]

There seems to be a bug in the new version of lavaan (0.6-6). I have done the multiple group regression below in lavaan 0.6-3 without any problems but now get an error message concerning the standard erros.

You can finde some example data, code an error message below.

Thanks for the great package to all the contributers!

example <-

structure(list(cluster = c(1216551L, 1212750L, 1218932L, 1211959L, 
1218821L, 1211202L, 1217646L, 1210088L, 1217206L, 1215094L, 1214164L, 
1213861L, 1210688L, 1210194L, 1214785L, 1211827L, 1215240L, 1217153L, 
1218138L, 1217631L, 1215094L, 1210718L, 1217417L, 1215406L, 1213593L, 
1218782L, 1212019L, 1212914L, 1219324L, 1215725L, 1217357L, 1218680L, 
1217249L, 1219149L, 1219774L, 1215314L, 1210718L, 1217249L, 1212091L, 
1213969L, 1216195L, 1215900L, 1210842L, 1211176L, 1212966L, 1216241L, 
1217423L, 1214284L, 1213582L, 1218932L, 1214221L, 1212750L, 1210097L, 
1211808L, 1215959L, 1219554L, 1217357L, 1216163L, 1213582L, 1217631L, 
1210642L, 1210165L, 1215966L, 1212232L, 1212600L, 1214754L, 1214785L, 
1216374L, 1215240L, 1210642L, 1218145L, 1212232L, 1218138L, 1210562L, 
1211204L, 1213969L, 1212643L, 1210755L, 1219112L, 1210974L, 1215094L, 
1212966L, 1212643L, 1213969L, 1212582L, 1215959L, 1217901L, 1214341L, 
1211176L, 1215972L, 1212232L, 1216575L, 1215156L, 1212116L, 1210874L, 
1211942L, 1219560L, 1217144L, 1215240L, 1216308L, 1214463L, 1216387L, 
1217430L, 1215082L, 1210131L, 1216661L, 1215156L, 1214754L, 1212056L, 
1211204L, 1217144L, 1219605L, 1215406L, 1211210L, 1210088L, 1214341L, 
1215725L, 1213517L, 1211204L, 1211500L, 1218782L, 1213969L, 1210165L, 
1211210L, 1211202L, 1215314L, 1216009L, 1217430L, 1217631L, 1212600L, 
1211500L, 1212709L, 1215094L, 1216575L, 1214727L, 1214284L, 1213582L, 
1215576L, 1212232L, 1216465L, 1215156L, 1210165L, 1211827L, 1213517L, 
1217430L, 1216176L, 1217249L, 1216465L, 1215634L, 1211202L, 1216551L, 
1215634L, 1215634L, 1214785L, 1216387L, 1216163L, 1216163L, 1216241L, 
1213582L, 1219605L, 1215871L, 1217452L, 1216551L, 1216163L, 1210729L, 
1212143L, 1219905L, 1214058L, 1216003L, 1210842L, 1217249L, 1215406L, 
1215634L, 1211500L, 1217423L, 1215966L, 1212339L, 1217500L, 1217153L, 
1213969L, 1214086L, 1216619L, 1210974L, 1216223L, 1217423L, 1219982L, 
1213759L, 1212431L, 1214754L, 1211232L, 1217249L, 1215634L, 1217153L, 
1218418L, 1217572L, 1216465L, 1218932L, 1212582L, 1212667L, 1215679L, 
1212091L, 1212056L, 1212232L, 1219516L, 1215725L, 1213582L, 1214284L, 
1219774L, 1218373L, 1213582L, 1219268L, 1215679L, 1214221L, 1216176L, 
1218932L, 1215576L, 1219560L, 1213720L, 1214086L, 1218773L, 1217417L, 
1218890L, 1216004L, 1215156L, 1219112L, 1219324L, 1212914L, 1218863L, 
1210194L, 1217894L, 1215406L, 1213646L, 1213698L, 1219516L, 1212667L, 
1216575L, 1219216L, 1216223L, 1211833L, 1219516L, 1210097L, 1210061L, 
1210097L, 1212966L, 1216661L, 1219560L, 1212143L, 1212582L, 1215082L, 
1218391L, 1216465L, 1218391L, 1216223L, 1219560L, 1217144L, 1212750L, 
1219982L, 1216575L, 1212431L, 1219560L, 1217572L, 1219324L, 1216465L, 
1217631L, 1213698L, 1218773L, 1216777L, 1211176L, 1210522L, 1215240L, 
1217901L, 1213759L, 1215634L, 1210031L, 1216843L, 1219605L, 1210522L, 
1211959L, 1218782L, 1217206L, 1215314L, 1211403L, 1215583L, 1213759L, 
1211827L, 1211216L, 1216387L, 1210562L, 1210088L, 1213517L, 1218373L, 
1215900L, 1215715L, 1219216L, 1216003L, 1212232L, 1210165L, 1219216L, 
1215093L, 1210194L), group = structure(c(4L, 1L, 1L, 1L, 1L, 
6L, 1L, 5L, 5L, 4L, 3L, 3L, 4L, 4L, 5L, 4L, 5L, 2L, 6L, 4L, 2L, 
2L, 4L, 3L, 4L, 1L, 1L, 1L, 1L, 3L, 4L, 4L, 1L, 6L, 4L, 4L, 2L, 
5L, 5L, 5L, 5L, 1L, 4L, 6L, 3L, 2L, 2L, 5L, 2L, 6L, 4L, 6L, 6L, 
1L, 3L, 2L, 1L, 3L, 6L, 4L, 1L, 3L, 2L, 1L, 1L, 5L, 1L, 1L, 3L, 
4L, 1L, 1L, 5L, 6L, 3L, 5L, 1L, 2L, 1L, 2L, 4L, 1L, 6L, 4L, 3L, 
5L, 3L, 4L, 3L, 6L, 4L, 4L, 3L, 6L, 3L, 2L, 3L, 1L, 2L, 2L, 3L, 
1L, 6L, 6L, 1L, 3L, 3L, 4L, 1L, 4L, 4L, 3L, 4L, 1L, 6L, 5L, 6L, 
1L, 4L, 6L, 1L, 4L, 2L, 5L, 1L, 1L, 5L, 5L, 6L, 4L, 5L, 1L, 3L, 
2L, 2L, 6L, 3L, 4L, 6L, 3L, 4L, 2L, 2L, 6L, 3L, 5L, 2L, 3L, 3L, 
3L, 2L, 3L, 4L, 1L, 1L, 4L, 6L, 6L, 6L, 3L, 1L, 5L, 2L, 5L, 1L, 
2L, 6L, 6L, 6L, 3L, 1L, 1L, 6L, 2L, 5L, 5L, 5L, 5L, 1L, 2L, 1L, 
4L, 3L, 3L, 3L, 5L, 2L, 2L, 6L, 4L, 4L, 6L, 4L, 3L, 1L, 3L, 3L, 
1L, 1L, 1L, 3L, 3L, 1L, 3L, 4L, 6L, 1L, 6L, 6L, 3L, 4L, 4L, 2L, 
1L, 2L, 5L, 4L, 3L, 2L, 1L, 3L, 3L, 4L, 2L, 1L, 5L, 3L, 1L, 2L, 
1L, 4L, 1L, 3L, 2L, 1L, 6L, 1L, 1L, 5L, 3L, 4L, 4L, 6L, 6L, 6L, 
3L, 1L, 4L, 6L, 1L, 4L, 1L, 2L, 6L, 1L, 3L, 3L, 1L, 1L, 6L, 5L, 
6L, 1L, 3L, 4L, 4L, 1L, 2L, 5L, 1L, 3L, 3L, 3L, 4L, 4L, 3L, 6L, 
3L, 4L, 3L, 2L, 2L, 6L, 2L, 4L, 5L, 4L, 5L, 2L, 1L, 1L, 6L, 6L, 
3L, 6L, 4L, 3L, 1L, 4L, 5L), .Label = c("KiGGS baseline.boys", 
"KiGGS wave 1.boys", "KiGGS wave 2.boys", "KiGGS baseline.girls", 
"KiGGS wave 1.girls", "KiGGS wave 2.girls"), class = "factor"), 
    weight = c(0.748, 1.643, 1.738, 2.256, 1.39, 1.039049, 1.363, 
    0.71, 0.278, 0.479, 1.75635, 0.424747, 0.784, 0.885, 0.401, 
    2.761, 2.272, 0.145, 1.376533, 0.846, 1.279, 0.721, 0.353, 
    0.350872, 0.872, 0.941, 1.252, 1.906, 2.781, 0.579801, 1.05, 
    0.884, 0.726, 0.98626, 0.879, 0.944, 0.838, 2.025, 0.512, 
    1.675, 0.521, 0.298, 0.858, 0.610177, 0.479534, 2.716, 1.076, 
    0.34, 0.746, 4.495547, 2.175, 0.523466, 0.54173, 0.57, 0.934128, 
    1.429, 1.777, 0.975983, 0.271093, 0.84, 0.828, 0.972643, 
    0.184, 0.898, 0.932, 0.816, 0.691, 0.773, 1.717032, 0.35, 
    0.272, 1.084, 0.886, 2.416281, 0.478514, 4.046, 2.451, 0.221, 
    1.734, 0.448, 0.495, 0.621, 2.523893, 1.692, 1.296136, 0.688, 
    1.204867, 0.752, 1.406717, 0.594652, 1.126, 1.163, 0.488388, 
    0.622952, 0.490857, 1.313, 0.176891, 1.141, 0.542, 0.895, 
    1.885996, 0.952, 0.288519, 0.998647, 0.211, 0.410095, 0.720388, 
    1.997, 0.154, 1.679, 1.046, 0.3649, 0.371, 0.175, 3.336931, 
    0.751, 0.333203, 0.97, 3.115, 0.796146, 0.695, 1.065, 0.749, 
    0.135, 1.604, 1.818, 2.201, 0.147, 0.538155, 1.095, 0.672, 
    2.578, 0.845262, 1.375, 1.159, 0.493564, 0.282503, 0.957, 
    2.360684, 0.454013, 0.984, 0.767, 0.506, 0.836162, 0.417105, 
    1.626, 0.334, 0.560284, 0.519435, 1.415994, 1.914, 0.394791, 
    1.142, 0.756, 1.123, 0.787, 1.733964, 1.201784, 0.354825, 
    0.35148, 2.403, 1.546, 0.431, 1.016, 1.047, 0.662, 2.229074, 
    1.196944, 0.183968, 0.921822, 1.417, 1.315, 0.662693, 0.619, 
    0.79, 0.133, 0.444, 2.51, 0.427, 1.295, 0.709, 1.817, 1.306946, 
    0.841076, 1.80179, 0.608, 0.479, 0.227, 3.958422, 1.453, 
    1.771, 0.565018, 0.567, 1.145934, 0.382, 0.799837, 1.975776, 
    0.909, 0.276, 1.753, 0.186561, 0.137862, 0.986, 1.054853, 
    1.03, 0.240388, 1.213, 0.191534, 0.98522, 0.350949, 1.059, 
    1.061, 0.837, 0.991, 0.866, 1.515, 0.437, 3.893398, 0.189, 
    1.064, 0.31965, 0.657363, 0.729, 0.753, 1.264, 4.549, 0.891621, 
    2.225, 0.829, 1.526, 0.345, 1.189, 0.158029, 0.683, 0.747, 
    0.425448, 1.522, 1.006, 0.165, 2.143665, 0.674, 1.946, 0.436076, 
    0.265575, 0.449879, 0.418279, 0.819, 1.16, 0.821852, 0.713, 
    0.858, 1.051, 0.99, 0.307935, 1.352, 1.568886, 1.231436, 
    0.739, 0.286, 0.325997, 0.281, 1.501395, 1.422, 0.86128, 
    0.302, 0.886, 0.848, 1.561, 1.576, 1.14, 1.498008, 0.460689, 
    0.431643, 0.258, 1.658, 0.668756, 1.457878, 5.933396, 0.786, 
    0.795328, 1.972, 0.178, 0.592963, 1.849, 2.312, 0.126, 1.11, 
    1.246, 2.453, 0.444, 0.859, 0.505447, 1.017738, 0.782035, 
    0.12149, 1.849, 3.061805, 1.236, 1.182, 2.462), Y = c(17, 
    9, 7, 7, 12, 5, 9, 6, 4, 1, 5, 16, 12, 5, 6, 12, 23, 1, 16, 
    10, 5, 14, 10, 16, 7, 8, 9, 3, 10, 12, 7, 4, 8, 19, 9, 11, 
    12, 30, 17, 9, 10, 9, 12, 8, 3, 20, 9, 7, 8, 9, 4, 16, 10, 
    9, 11, 20, 7, 4, 8, 16, 7, 6, 12, 3, 10, 13, 11, 10, 10, 
    14, 5, 11, 6, 14, 8, 15, 5, 11, 5, 8, 13, 17, 23, 13, 13, 
    15, 8, 15, 10, 6, 17, 10, 10, 8, 6, 8, 9, 5, 4, 1, 13, 8, 
    11, 9, 2, 16, 9, 13, 6, 11, 3, 12, 12, 13, 6, 8, 15, 11, 
    12, 14, 4, 15, 13, 4, 9, 8, 13, 8, 7, 14, 0, 12, 11, 5, 6, 
    15, 16, 13, 5, 5, 11, 5, 7, 9, 10, 7, 13, 18, 11, 5, 8, 15, 
    8, 14, 8, 13, 7, 3, 8, 8, 2, 3, 8, 6, 6, 15, 17, 17, 17, 
    10, 8, 9, 7, 10, 11, 10, 0, 13, 7, 8, 9, 6, 4, 9, 9, 7, 12, 
    4, 13, 10, 10, 14, 7, 9, 9, 9, 6, 9, 6, 16, 9, 13, 6, 6, 
    10, 16, 7, 8, 11, 2, 5, 8, 9, 11, 4, 7, 8, 8, 4, 9, 8, 9, 
    12, 9, 8, 3, 19, 10, 4, 16, 20, 14, 10, 5, 14, 7, 5, 10, 
    1, 14, 7, 13, 4, 5, 13, 8, 7, 16, 4, 11, 7, 13, 7, 7, 9, 
    3, 11, 7, 3, 6, 13, 17, 12, 7, 11, 10, 6, 6, 8, 3, 8, 8, 
    10, 11, 19, 12, 1, 3, 8, 8, 4, 10, 7, 9, 13, 3, 9, 11, 4, 
    11, 2, 19, 3, 7, 21, 19, 8, 5, 11, 4), X = c(4, 3, 0, 4, 
    6, 2, 1, 0, 0, 2, 5, 0, 1, 1, 0, 5, 2, 6, 0, 2, 5, 0, 2, 
    3, 3, 6, 3, 6, 4, 0, 2, 6, 1, 5, 4, 6, 5, 6, 1, 5, 0, 5, 
    3, 2, 4, 0, 3, 2, 4, 5, 4, 5, 5, 2, 4, 6, 5, 2, 4, 5, 6, 
    1, 5, 0, 1, 3, 6, 0, 1, 2, 2, 4, 3, 6, 2, 2, 1, 1, 6, 3, 
    1, 2, 1, 2, 0, 2, 0, 2, 2, 2, 5, 5, 3, 4, 6, 6, 3, 3, 2, 
    3, 1, 0, 6, 5, 1, 3, 6, 0, 1, 1, 5, 1, 1, 0, 4, 4, 1, 6, 
    3, 3, 4, 2, 0, 5, 6, 2, 3, 3, 0, 5, 5, 1, 4, 5, 5, 3, 6, 
    6, 2, 1, 6, 3, 0, 6, 3, 4, 2, 5, 0, 0, 5, 0, 1, 4, 1, 3, 
    5, 2, 1, 6, 2, 4, 2, 5, 0, 4, 5, 6, 4, 4, 2, 4, 1, 1, 6, 
    4, 0, 3, 3, 1, 5, 0, 2, 2, 6, 4, 2, 0, 3, 3, 1, 3, 4, 5, 
    1, 6, 5, 4, 1, 1, 3, 1, 4, 0, 3, 4, 5, 3, 4, 6, 4, 2, 0, 
    1, 4, 1, 4, 1, 6, 5, 4, 1, 6, 1, 1, 0, 3, 5, 4, 3, 0, 6, 
    3, 3, 5, 3, 6, 4, 1, 5, 5, 1, 5, 0, 1, 6, 0, 5, 3, 3, 2, 
    6, 5, 4, 1, 0, 4, 6, 2, 3, 6, 6, 5, 5, 0, 2, 5, 5, 0, 4, 
    4, 0, 2, 2, 5, 4, 5, 6, 0, 5, 6, 6, 0, 4, 0, 6, 3, 3, 4, 
    3, 5, 3, 6, 4, 3, 1, 6, 4, 3, 1)), class = "data.frame", row.names = c(NA, 
-300L))

model <- '# Pfade
              Y ~ c(a.2w, a.2m, a.1w, a.1m, a.0w, a.0m)*X
             
               #Intercepts
              Y ~ cc(i.2w, i.2m, i.1w, i.1m, i.0w, i.0m)*1
              '



sdq_fit <- sem(data=example, model=model, group="group", cluster = "cluster", sampling.weights = "weight", estimator="GLS")

Error in lav_options_set(opt) : lavaan ERROR: invalid value for `se' argument when estimator is GLS: robust.cluster

6.	stop("lavaan ERROR: invalid value for `se' argument when estimator is GLS: ", opt$se, "\n")

5.	lav_options_set(opt)

4.	lavaan::lavaan(model = model, data = example, sampling.weights = "weight", group = "group", cluster = "cluster", estimator = "GLS", model.type = "sem", int.ov.free = TRUE, int.lv.free = FALSE, auto.fix.first = TRUE, auto.fix.single = TRUE, auto.var = TRUE, ...

3.	eval(mc, parent.frame())

2.	eval(mc, parent.frame())

1.	sem(data = example, model = model, group = "group", cluster = "cluster", sampling.weights = "weight", estimator = "GLS")

[Yves Rosseel]

On 6/8/20 1:06 PM, 'Stephan Junker' via lavaan wrote:
> sdq_fit <-sem(data=example,model=model,group="group",cluster
> ="cluster",sampling.weights ="weight",estimator="GLS")

When sampling.weights are used, the only available 'estimator' is MLR
(which is just ML, but with cluster-robust standard errors). lavaan
0.6-6 correctly detects that estimator = "GLS" is not compatible with se
= "robust.cluster". It would seem that lavaan failed to detect this in
earlier versions.

The solution is simple: remove estimator = "GLS", and lavaan will pick
the MLR estimator automatically.

Yves.