thr3ads.net - R help - [R] CFA and Factor scores from ordered Likert scale items using SEM and Lavaan packages [Aug 2013]

If this information is useful, please help other people find it:
Share via:

Jānis Mucenieks

2013-Aug-07 09:36 UTC

[R] CFA and Factor scores from ordered Likert scale items using SEM and Lavaan packages

Hello!
I'm trying to do some confirmatory factor analysis with for ordered Likert
scale items using sem and lavaan packages. So far I have read a lot and
searched a lot but working example has so far eluded me. Can someone please
take a look at the data and tell me if what I'm doing is correct (lavaan
case) and where is the error (sem case)?

Original data (ordered factors):
EI.data <- structure(list(Q21a_makeAnythingEntrep_likert = structure(c(3L,
1L, 7L, 1L, 1L, 6L, 1L, 1L, 5L, 1L, 1L, 1L, 3L, 4L, 1L, 7L, 2L,
1L, 7L, 7L, 1L, 3L, 1L, 2L, 5L, 1L, 5L, 2L, 4L, 1L, 1L, 2L), .Label
c("1",
"2", "3", "4", "5", "6",
"7"), class = c("ordered", "factor")),
    Q21b_goalBecomingEntrep_likert = structure(c(3L, 1L, 7L,
    1L, 1L, 6L, 1L, 1L, 7L, 1L, 1L, 7L, 3L, 2L, 2L, 7L, 2L, 3L,
    7L, 7L, 1L, 4L, 1L, 2L, 4L, 2L, 6L, 3L, 5L, 1L, 1L, 1L), .Label
c("1",
    "2", "3", "4", "5", "6",
"7"), class = c("ordered", "factor"
    )), Q21c_effortFirm_likert = structure(c(3L, 3L, 7L, 1L,
    1L, 6L, 1L, 1L, 5L, 1L, 1L, 5L, 3L, 3L, 1L, 7L, 2L, 2L, 7L,
    7L, 1L, 5L, 1L, 2L, 4L, 2L, 7L, 2L, 5L, 1L, 1L, 1L), .Label =
c("1",
    "2", "3", "4", "5", "6",
"7"), class = c("ordered", "factor"
    )), Q21d_createFirm_likert = structure(c(3L, 3L, 7L, 1L,
    1L, 6L, 1L, 2L, 7L, 1L, 1L, 7L, 3L, 6L, 3L, 7L, 2L, 2L, 7L,
    5L, 1L, 6L, 1L, 2L, 4L, 5L, 7L, 5L, 4L, 1L, 5L, 6L), .Label =
c("1",
    "2", "3", "4", "5", "6",
"7"), class = c("ordered", "factor"
    )), Q21e_.thoughtFirm_likert = structure(c(3L, 6L, 7L, 1L,
    1L, 6L, 2L, 1L, 7L, 1L, 4L, 7L, 5L, 4L, 2L, 7L, 2L, 2L, 7L,
    3L, 1L, 5L, 1L, 2L, 6L, 2L, 7L, 4L, 5L, 2L, 6L, 2L), .Label =
c("1",
    "2", "3", "4", "5", "6",
"7"), class = c("ordered", "factor"
    )), Q21f_intentionFirm_likert = structure(c(4L, 4L, 7L, 1L,
    1L, 6L, 1L, 2L, 7L, 1L, 1L, 7L, 5L, 6L, 3L, 7L, 2L, 2L, 7L,
    2L, 1L, 7L, 1L, 2L, 6L, 4L, 7L, 7L, 4L, 2L, 5L, 5L), .Label =
c("1",
    "2", "3", "4", "5", "6",
"7"), class = c("ordered", "factor"
    ))), .Names = c("Q21a_makeAnythingEntrep_likert",
"Q21b_goalBecomingEntrep_likert",
"Q21c_effortFirm_likert", "Q21d_createFirm_likert",
"Q21e_.thoughtFirm_likert",
"Q21f_intentionFirm_likert"), row.names = c(NA, 32L), class =
"data.frame")

Original data (continuous numbers):
EI.dataN <- structure(list(Q21a_makeAnythingEntrep_likert = c(3L, 1L, 7L,
1L, 1L, 6L, 1L, 1L, 5L, 1L, 1L, 1L, 3L, 4L, 1L, 7L, 2L, 1L, 7L,
7L, 1L, 3L, 1L, 2L, 5L, 1L, 5L, 2L, 4L, 1L, 1L, 2L),
Q21b_goalBecomingEntrep_likert = c(3L,
1L, 7L, 1L, 1L, 6L, 1L, 1L, 7L, 1L, 1L, 7L, 3L, 2L, 2L, 7L, 2L,
3L, 7L, 7L, 1L, 4L, 1L, 2L, 4L, 2L, 6L, 3L, 5L, 1L, 1L, 1L),
    Q21c_effortFirm_likert = c(3L, 3L, 7L, 1L, 1L, 6L, 1L, 1L,
    5L, 1L, 1L, 5L, 3L, 3L, 1L, 7L, 2L, 2L, 7L, 7L, 1L, 5L, 1L,
    2L, 4L, 2L, 7L, 2L, 5L, 1L, 1L, 1L), Q21d_createFirm_likert = c(3L,
    3L, 7L, 1L, 1L, 6L, 1L, 2L, 7L, 1L, 1L, 7L, 3L, 6L, 3L, 7L,
    2L, 2L, 7L, 5L, 1L, 6L, 1L, 2L, 4L, 5L, 7L, 5L, 4L, 1L, 5L,
    6L), Q21e_.thoughtFirm_likert = c(3L, 6L, 7L, 1L, 1L, 6L,
    2L, 1L, 7L, 1L, 4L, 7L, 5L, 4L, 2L, 7L, 2L, 2L, 7L, 3L, 1L,
    5L, 1L, 2L, 6L, 2L, 7L, 4L, 5L, 2L, 6L, 2L), Q21f_intentionFirm_likert
= c(4L,
    4L, 7L, 1L, 1L, 6L, 1L, 2L, 7L, 1L, 1L, 7L, 5L, 6L, 3L, 7L,
    2L, 2L, 7L, 2L, 1L, 7L, 1L, 2L, 6L, 4L, 7L, 7L, 4L, 2L, 5L,
    5L)), .Names = c("Q21a_makeAnythingEntrep_likert",
"Q21b_goalBecomingEntrep_likert",
"Q21c_effortFirm_likert", "Q21d_createFirm_likert",
"Q21e_.thoughtFirm_likert",
"Q21f_intentionFirm_likert"), class = "data.frame",
row.names = c(NA,
-32L))

Lavaan case:
EI.model.lavaan <- ' EI  =~ Q21a_makeAnythingEntrep_likert +
Q21b_goalBecomingEntrep_likert + Q21c_effortFirm_likert +
Q21d_createFirm_likert + Q21e_.thoughtFirm_likert +
Q21f_intentionFirm_likert '
EI.fit <- cfa(EI.model.lavaan, data = EI.data)
summary(EI.fit, fit.measures = TRUE)
predict(EI.fit)

Sem continuous items case:

EI.dataN.cov <- cov(EI.dataN)
model.EI.dataN <- specifyModel()
EI -> Q21a_makeAnythingEntrep_likert, NA, 1
EI -> Q21b_goalBecomingEntrep_likert, ei1
EI -> Q21c_effortFirm_likert, ei2
EI -> Q21d_createFirm_likert, ei3
EI -> Q21e_.thoughtFirm_likert, ei4
EI -> Q21f_intentionFirm_likert, ei5
EI <-> EI, var1
Q21a_makeAnythingEntrep_likert <-> Q21a_makeAnythingEntrep_likert, error1
Q21b_goalBecomingEntrep_likert <-> Q21b_goalBecomingEntrep_likert, error2
Q21c_effortFirm_likert <-> Q21c_effortFirm_likert, error3
Q21d_createFirm_likert <-> Q21d_createFirm_likert, error4
Q21e_.thoughtFirm_likert <-> Q21e_.thoughtFirm_likert, error5
Q21f_intentionFirm_likert <-> Q21f_intentionFirm_likert, error6

EI.dataN.cfa <- sem( model.EI.dataN, EI.dataN.cov, nrow(EI.dataN))
summary(EI.dataN.cfa)
fscores(EI.dataN.cfa, EI.dataN)

Sem ordered items case:
hcor <- function(data) hetcor(data, std.err=FALSE)$correlations
R.EI.data <- hcor(EI.data)
EI.model.sem.o <- specifyModel()

EI -> Q21a_makeAnythingEntrep_likert, NA, 1
EI -> Q21b_goalBecomingEntrep_likert, ei1
EI -> Q21c_effortFirm_likert, ei2
EI -> Q21d_createFirm_likert, ei3
EI -> Q21e_.thoughtFirm_likert, ei4
EI -> Q21f_intentionFirm_likert, ei5
EI <-> EI, var1
Q21a_makeAnythingEntrep_likert <-> Q21a_makeAnythingEntrep_likert, error1
Q21b_goalBecomingEntrep_likert <-> Q21b_goalBecomingEntrep_likert, error2
Q21c_effortFirm_likert <-> Q21c_effortFirm_likert, error3
Q21d_createFirm_likert <-> Q21d_createFirm_likert, error4
Q21e_.thoughtFirm_likert <-> Q21e_.thoughtFirm_likert, error5
Q21f_intentionFirm_likert <-> Q21f_intentionFirm_likert, error6

cfaEI.sem.o <- sem(EI.model.sem.o, R.EI.data, nrow(EI.data))
summary(cfaEI.sem.o)
fscores(cfaEI.sem.o, EI.data)

Error (can't get fscrores):
Error in colMeans(x, na.rm = TRUE) : 'x' must be numeric

Thank You!
Janis

	[[alternative HTML version deleted]]

John Fox

2013-Aug-08 03:35 UTC

head link

[R] CFA and Factor scores from ordered Likert scale items using SEM and Lavaan packages

Dear J?nis,

On Wed, 7 Aug 2013 12:36:11 +0300
 J?nis Mucenieks <mucenieks at gmail.com> wrote:> Hello!
> I'm trying to do some confirmatory factor analysis with for ordered
Likert
> scale items using sem and lavaan packages. So far I have read a lot and
> searched a lot but working example has so far eluded me. Can someone please
> take a look at the data and tell me if what I'm doing is correct
(lavaan
> case) and where is the error (sem case)?
> 
. . .
> 
> cfaEI.sem.o <- sem(EI.model.sem.o, R.EI.data, nrow(EI.data))
> summary(cfaEI.sem.o)
> fscores(cfaEI.sem.o, EI.data)
> 
> Error (can't get fscrores):
> Error in colMeans(x, na.rm = TRUE) : 'x' must be numeric
fscores() can't handle ordinal data.

Also, though it's not relevant to this error, you could greatly simplify the
model specification by using cfa() instead of specifyModel(). More generally,
specifyEquations() is usually more convenient than specifyModel().

Best,
 John
> 
> Thank You!
> Janis
> 
> 	[[alternative HTML version deleted]]
> 
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
------------------------------------------------
John Fox
Sen. William McMaster Prof. of Social Statistics
Department of Sociology
McMaster University
Hamilton, Ontario, Canada
http://socserv.mcmaster.ca/jfox/

R help - Aug 2013 - CFA and Factor scores from ordered Likert scale items using SEM and Lavaan packages

[R] CFA and Factor scores from ordered Likert scale items using SEM and Lavaan packages

[R] CFA and Factor scores from ordered Likert scale items using SEM and Lavaan packages