3

i wanted to do a CFA (which worked out fine) and then I deleted the two items with the lowest R² to improve the fit of the model. The problem is that it just worked out for one of them - and of course the fit was better than before. But when I tried to delete the second item (in addition to the first one), the warning named above appeared. The first deleted item was "External7" and the second one was "Negative2". Any ideas or supports?

install.packages("sem")
library(sem)
cov.matrix<-cov(na.omit(verbesserungfit))
View(cov.matrix)

cfa.model11 <- specifyModel() #definition welche items zu welchen latenten variablen gehören #und benennung der pfade von latenter variable zu item
EXTERNAL -> External1, external1
EXTERNAL -> External2, external2
EXTERNAL -> External3, external3
EXTERNAL -> External4, external4
EXTERNAL -> External5, external5
EXTERNAL -> External6, external6
Here, I deleted the row EXTERNAL -> External7, external7
SELF -> Self1, self1
SELF -> Self2, self2
SELF -> Self3, self3
SELF -> Self4, self4
POSITIVE -> Positive1, positive1
POSITIVE -> Positive2, positive2
POSITIVE -> Positive3, positive3
POSITIVE -> Positive4, positive4
NEGATIVE -> Negative1, negative1
Here, i deleted the row NEGATIVE -> Negative2, negative2
NEGATIVE -> Negative3, negative3
NEGATIVE -> Negative4, negative4
PROMOTION -> Promotion1, promotion1
PROMOTION -> Promotion2, promotion2
PROMOTION -> Promotion3, promotion3
PROMOTION -> Promotion4, promotion4
PROMOTION -> Promotion5, promotion5
GENERAL -> General1, general1
GENERAL -> General2, general2
GENERAL -> General3, general3
GENERAL -> General4, general4
GENERAL -> General5, general5
CAREER -> Career1, career1
CAREER -> Career2, career2
CAREER -> Career3, career3
CAREER -> Career4, career4
CAREER -> Career5, career5
FINANCIAL -> Financial1, financial1
FINANCIAL -> Financial2, financial2
FINANCIAL -> Financial3, financial3
FINANCIAL -> Financial4, financial4
FINANCIAL -> Financial5, financial5
ENTERTAINMENT -> Entertainment1, entertainment1
ENTERTAINMENT -> Entertainment2, entertainment2
ENTERTAINMENT -> Entertainment3, entertainment3
ENTERTAINMENT -> Entertainment4, entertainment4
ENTERTAINMENT -> Entertainment5, entertainment5
EXTERNAL<->EXTERNAL,NA,1 #Varianz der latenten Variablen definieren # NA und 1 für idendification des modells (standardisierte Variablen) # fixierung der varianzen #varExternal
SELF<->SELF, NA, 1 #varSelf
POSITIVE<->POSITIVE, NA, 1 #varPositive
NEGATIVE<->NEGATIVE, NA, 1 #varNegative
PROMOTION<->PROMOTION, NA, 1 #varPromotion
GENERAL<->GENERAL, NA, 1 #varGeneral
CAREER<->CAREER, NA, 1 #varCarrer
FINANCIAL<->FINANCIAL, NA, 1 #varFinancial
ENTERTAINMENT<->ENTERTAINMENT, NA, 1 #varEnterntainment
External1<->External1, error01 #hinzufuegen von stoerterm zu items, unerklärter teil pro item
External2<->External2, error02
External3<->External3, error03
External4<->External4, error04
External5<->External5, error05
External6<->External6, error06
Here, i deleted the row External7<->External7, error07
Self1<->Self1, error08
Self2<->Self2, error09
Self3<->Self3, error10
Self4<->Self4, error11
Positive1<->Positive1, error12
Positive2<->Positive2, error13
Positive3<->Positive3, error14
Positive4<->Positive4, error15
Negative1<->Negative1, error16
Here, i deleted the row Negative2<->Negative2, error17
Negative3<->Negative3, error18
Negative4<->Negative4, error19
Promotion1<->Promotion1, error20
Promotion2<->Promotion2, error21
Promotion3<->Promotion3, error22
Promotion4<->Promotion4, error23
Promotion5<->Promotion5, error24
General1<->General1, error25
General2<->General2, error26
General3<->General3, error27
General4<->General4, error28
General5<->General5, error29
Career1<->Career1, error30
Career2<->Career2, error31
Career3<->Career3, error32
Career4<->Career4, error33
Career5<->Career5, error34
Financial1<->Financial1, error35
Financial2<->Financial2, error36
Financial3<->Financial3, error37
Financial4<->Financial4, error38
Financial5<->Financial5, error39
Entertainment1<->Entertainment1, error40
Entertainment2<->Entertainment2, error41
Entertainment3<->Entertainment3, error42
Entertainment4<->Entertainment4, error43
Entertainment5<->Entertainment5, error44
EXTERNAL<->SELF, cov12 #covarianz zwischen den latenten variablen
EXTERNAL<->POSITIVE, cov13
EXTERNAL<->NEGATIVE, cov14
EXTERNAL<->PROMOTION, cov15
EXTERNAL<->GENERAL, cov16
EXTERNAL<->CAREER, cov17
EXTERNAL<->FINANCIAL, cov18
EXTERNAL<->ENTERTAINMENT, cov19
SELF<->POSITIVE, cov23
SELF<->NEGATIVE, cov24
SELF<->PROMOTION, cov25
SELF<->GENERAL, cov26
SELF<->CAREER, cov27
SELF<->FINANCIAL, cov28
SELF<->ENTERTAINMENT, cov29
POSITIVE<->NEGATIVE, cov34
POSITIVE<->PROMOTION, cov35
POSITIVE<->GENERAL, cov36
POSITIVE<->CAREER, cov37
POSITIVE<->FINANCIAL, cov38
POSITIVE<->ENTERTAINMENT, cov39
NEGATIVE<->PROMOTION, cov45
NEGATIVE<->GENERAL, cov46
NEGATIVE<->CAREER, cov47
NEGATIVE<->FINANCIAL, cov48
NEGATIVE<->ENTERTAINMENT, cov49
PROMOTION<->GENERAL, cov56
PROMOTION<->CAREER, cov57
PROMOTION<->FINANCIAL, cov58
PROMOTION<->ENTERTAINMENT, cov59
GENERAL<->CAREER, cov67
GENERAL<->FINANCIAL, cov68
GENERAL<->ENTERTAINMENT, cov69
CAREER<->FINANCIAL, cov78
CAREER<->ENTERTAINMENT, cov79
FINANCIAL<->ENTERTAINMENT, cov89

cfa11 <- sem( cfa.model11, cov.matrix, nrow(verbesserungfit)) # was muss alles an cfa gesendet werden #modellname #name für covarianzmatrix #wie viele participants in dataset mit n row function
summary(cfa11,fit.indices=c("GFI", "AGFI", "RMSEA", "NFI", "NNFI", "CFI", "RNI", "IFI", "SRMR", "AIC", "AICc", "BIC", "CAIC"))

Thanks for your support!

John Conde
  • 217,595
  • 99
  • 455
  • 496
Jakob
  • 31
  • 1

1 Answers1

0

The warning means the model couldn't converge, i.e., no solution could be found in the estimation procedure for the parameters you specified.

Two issues may be at work here. First, I find it easier to use my.mod <- cfa() when examining huge measurement models. Check the sem package help page for more on that. You update models in the same way as with specifyModel

Second, it will be more helpful if you post the model object for each model, it is possible you deleted a variance or covariance that you did not need to, or, alternatively forgot to delete.

Finally, it would be helpful if you could supply the sample size.

Rick Hass
  • 101
  • 3