request to have a "use lavaan code" for identifying factors

[Phillipkwood]

Hi all,
Sometimes there are structural models which need to mix identificaiton strategies. One example of this is the random intercept factor model in which two orthogonal factors are proposed. One factor is a congeneric factor with unit variance (i.e., all loadings freely estimated) while the other factor has factor loadings all set to unity and a freely estimated factor variance. Although it's possible to work around this for a single group model by specifying unit factor variances with one tau-equivalent factor and another congeneric factor, it's not possible to use this trick for weakly invariant multi-group factor models. Other packages such as JASP have the option of not forcing any identification strategy and just using the lavaan code. This would be very helpful.