Closed KaiyeeHe closed 1 year ago
This issue has been solved by assigning the readModels() result to a variable instead of running readModels() directly. I have attached the codes below and hope it helps someone who meets the same error.
Previous: readModels(here("ex11.8part2.out"))
Current codes that work well. a <- readModels(here("ex11.8part2.out")) a$summaries
Hey @KaiyeeHe that doesn't solve the error, the error is in the print function ;)
Can you try to run:
install.packages("remotes")
remotes::install_github("michaelhallquist/MplusAutomation@fix190")
And then run your original code that errored?
Hello both,
Thank you for your contributions to this matter. Despite trying the method suggested by @KaiyeeHe and @cjvanlissa, I am still unable to resolve my issue with extracting model fit information (such as CFI, TLI, etc.). The summaries within the object do not include the model chi2 and other indices. Could you advise on what steps I should take? Below is a snippet of my results. Thank you, everyone.
> readModels("./Mplus_Med/Execution/p1_acc", what = "summaries")
$p1_acc_mi_step1.out
p1_accuracy_mi_step1
Estimated using ML
Number of obs: 177, number of (free) parameters: 24
Model: Chi2(df = 66) = 619.203, p = 0
Baseline model: Chi2(df = 66) = 619.203, p = 0
Fit Indices:
CFI = 0, TLI = 0, SRMR = 0.302
RMSEA = 0.218, 90% CI [0.202, 0.233], p < .05 = 0
AIC = 3369.428, BIC = 3445.655
NULL
$p1_acc_mi_step2.out
p1_accuracy_fiml
Estimated using ML
Number of obs: 177, number of (free) parameters: 40
Fit Indices:
CFI = NA, TLI = NA, SRMR = NA
RMSEA = NA, 90% CI [NA, NA], p < .05 = NA
AIC = NA, BIC = NA
NULL
attr(,"class")
[1] "mplus.model.list" "list"
` ``
After consulting the Mplus support team. They informed me that as of version 8.10, Mplus only provides a single value for each fit index for continuous outcomes using maximum likelihood estimation, rather than detailed proportions and percentiles. Does this change affect how "MplusAutomation" retrieves and formats output from Mplus? Below is a snippet of my fit results for your reference, thank you guys!
Number of Free Parameters 40
Loglikelihood
H0 Value -2127.141
H1 Value -1747.253
* The loglikelihood cannot be used directly for chi-square testing with
imputed data.
Information Criteria
Akaike (AIC) 4334.282
Bayesian (BIC) 4454.707
Sample-Size Adjusted BIC 4328.115
(n* = (n + 2) / 24)
Chi-Square Test of Model Fit
Value 56.175
Degrees of Freedom 48
P-Value 0.1953
RMSEA (Root Mean Square Error Of Approximation)
Estimate 0.034
90 Percent C.I. 0.000 0.066
Probability RMSEA <= .05 0.768
CFI/TLI
CFI 0.890
TLI 0.849
Chi-Square Test of Model Fit for the Baseline Model
Value 140.614
Degrees of Freedom 66
P-Value 0.0000
SRMR (Standardized Root Mean Square Residual)
Value 0.588
While attempting to use readModels function to read models from Mplus output file, I received the following error message [Error in sprintf("Model: Chi2(df = %s) = %s, p = %s \n", ChiSqM_DF, ChiSqM_Value, : object 'ChiSqM_Value' not found].
Upon searching for a solution, I stumbled upon a similar problem posted on [https://github.com/michaelhallquist/MplusAutomation/issues/38]. I attempted to use the output file ex11.8part2.txt shared in the post, but still encountered the same error.
For your reference, I am using Mplus version 8.3 (Mac) and I have installed the MplusAutomation package in version 1.1.0 (R 4.2.1).
PS. Thank you for creating such a fantastic package!