Estimasi Parameter Pada Moderating Structural Equation Modeling (MSEM) Menggunakan Principal Component Analysis

Aritonang, Sarah Permatasari (2025) Estimasi Parameter Pada Moderating Structural Equation Modeling (MSEM) Menggunakan Principal Component Analysis. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6003222016-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 April 2027. Download (2MB) | Request a copy

Abstract

Moderating Structural Equation Modeling (MSEM) merupakan suatu metode yang menggabungkan analisis moderasi dan SEM secara simultan untuk mengidentifikasi keberadaan variabel moderator pada hubungan jalur antar variabel laten. Metode estimasi dengan interaksi dalam MSEM sering digunakan, meskipun memiliki kelemahan berupa kompleksitas metodologis, terutama dalam analisis data yang melibatkan hubungan non-linear. Sebagai alternatif, subgroup analysis dapat digunakan untuk menyederhanakan analisis moderasi. Namun, metode ini masih jarang diterapkan dalam penelitian meskipun memiliki potensi untuk mengidentifikasi efek moderasi yang spesifik, memberikan hasil yang tetap valid dan lebih terfokus pada karakteristik subpopulasi tertentu. Oleh karena itu, penggunaan subgroup analysis menjadi penting untuk mengeksplorasi perbedaan efek moderasi antar grup yang tidak dapat ditangkap oleh metode interaksi tradisional. Tujuan penelitian ini ialah menjelaskan algoritma Principal Component Analysis (PCA) untuk estimasi parameter pada tahapan subgrup dan algoritma PLS untuk tahapan menganalisis peran moderasi. Selanjutnya diimplementasikan pada variabel atribut produk dan studi kasus mengenai Keputusan Beli Produk drekle’sskincare. Populasi pelanggan dan pengguna produk drekle’sskincare di Jabodetabek sebanyak 1169 orang kemudian diambil 240 orang (20,53%) sebagai responden menggunakan simple random sampling. Pernyataan responden menggunakan skala likert dengan 5 kategori. Hasil analisis menunjukkan bahwa atribut produk berperan sebagai variabel moderator hanya pada hubungan antara psikologi persepsi dan niat beli. Namun, atribut produk tidak memoderasi hubungan antara psikologi persepsi atau niat beli terhadap keputusan beli konsumen. Temuan ini memberikan implikasi bahwa strategi pemasaran dan optimalisasi produk dengan atribut unggulan perlu difokuskan dalam membangun niat beli, sementara tahap keputusan beli membutuhkan pendekatan yang berbeda.
==================================================================================================================================
Moderating Structural Equation Modeling (MSEM) is a method that simultaneously combines moderation analysis and SEM to identify the presence of a moderator variable in the pathway relationships between latent variables. Interaction-based estimation methods in MSEM are frequently used, although they have methodological complexities, especially when analyzing data involving non-linear relationships. As an alternative, subgroup analysis can be employed to simplify moderation analysis. However, this method is still rarely applied in research, despite its potential to identify specific moderation effects, providing valid results that are more focused on the characteristics of certain subpopulations. Therefore, the use of subgroup analysis is crucial for exploring differences in moderation effects between groups that cannot be captured by traditional interaction methods. The purpose of this study is to explain the Principal Component Analysis (PCA) algorithm for parameter estimation in the subgroup stage and the Partial Least Squares (PLS) algorithm for analyzing the role of moderation. This was then implemented on the product attribute variable and the case study of purchasing decisions for drekle’sskincare products. From a population
of 1,169 customers and users of drekle’sskincare products in Jabodetabek, 240 respondents (20.53%) were selected using simple random sampling. Respondents’ statements were measured using a 5-category likert scale. The analysis results showed that product attributes act as a moderator variable only in the relationship between psychological perception and purchase intention. However, product attributes do not moderate the relationship between psychological perception or purchase intention and consumers’ purchasing decisions. These findings imply that marketing strategies and product optimization with superior attributes need to focus on building purchase intention, while different approaches are required for the purchasing decision stage.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	drekle’sskincare, Moderating Structural Equation Modeling, Principal Component Analysis
Subjects:	H Social Sciences > HA Statistics > HA29 Theory and method of social science statistics H Social Sciences > HA Statistics > HA31.7 Estimation H Social Sciences > HD Industries. Land use. Labor > HD30.28 Planning. Business planning. Strategic planning.
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49101-(S2) Master Thesis
Depositing User:	SARAH PERMATASARI ARITONANG
Date Deposited:	06 Feb 2025 05:00
Last Modified:	06 Feb 2025 05:00
URI:	http://repository.its.ac.id/id/eprint/118291

Actions (login required)

View Item