Page 91 - ISMCON souvenir 2021

ISMSCON - 2021

OS68: Application of recursive and non-recursive Structural
Equation modelling for categorical indicators with estimation
methods

Sylvia J & Dr. Leo Alexander
Department of Statistics
Loyola College, Sterling Road, Nungambakkam, Chennai 600 034
Email: sylvia.jayakumar@gmail.com

Structural equation modelling (SEM) is a causal modelling approach widely used in the field of ecology,
sociology and econometrics. The performance of SEM can be verified using model specification,
identification, parameter estimation, model evaluation and model modification (Kline 2010; Hoyle 2011;
Byrne 2013). In the estimation of model parameters, two prominent issues are observed the first one
is maximum likelihood (ML) estimator was often a subject of criticism in the literature because of the
unrealistic assumptions of the continuous observable, e latent variables (e.g., multivariate normal
distributions), and the large sample sizes which were needed to meet the asymptotic properties of this
estimator and efficient testing. Maximum Likelihood Estimation (ML) is a default method for model fitting
programs. The second approach was categorization in observed variables when skewed lead to spurious
measurement error correlations and biased standardized coefficients. Similarly, model misspecification
and when the data did not follow normal distribution can be resolved using other estimation methods
that could be tested such as Partial least squares, two stage least squares (2SLS), generalized method
of moments (GMM); nonlinear 2SLS (N2SLS); nonlinear three-stage least squares (N3SLS); robust
weighted least squares; (WLSMV ) maximum likelihood (ML); robust maximum likelihood (MLR). All of
these does not have the assumption of normality. Here in this research, different estimation methods
are tested for datasets to best understand which of these techniques is best suitable when the data has
categorical variables. It was found that 2SLS performed better than ML estimation. Further results will
be presented.
Keywords: Structural Equation Modelling, recursive, non-recursive, estimation

OS69: IMMUNOGENICITY OF PLANT DERIVED VACCINES OF

ROTAVIRUS: A SYSTEMATIC REVIEW

Dr. Tanya Tanu, Dr. Dewesh Kumar, Dr. VidyaSagar, Dr. Vivek Kashyap
Affiliation: Rajendra Institute of Medical Sciences, Ranchi, Bariatu, Jharkhand, 834009
Email id: tanyatanu35@gmail.com

Keywords: Rotavirus, Plant Based Vaccines, Plant Derived Vaccines, Vaccine
INTRODUCTION: In a developing country like India where outbreaks of Acute Diarrheal Diseases among
children are common and so is the shortage of vaccines during such outbreaks, plant-derived rotaviral
vaccines can prove to be the turning point in the control of diarrheal diseases.

OBJECTIVES: To determine the immunogenicity of plant-based vaccines of Rotavirus.
MATERIALS AND METHODS: A comprehensive search of the electronic database Pubmed was
conducted. The keywords used were Rotavirus AND (Plant Based Vaccines OR Plant Derived Vaccines
or Plant Vaccines). The search was augmented using the related articles link and through Google Scholar
and web searches.
Inclusion Criteria:

CONFERENCE SOUVENIR 89