tionnaire, which represents a 32.3%

tionnaire, which represents a 32.3% (64/198) response rate. This response rate was
considered very satisfactory, as potential participants were asked to provide sensitive
and confidential data about their performance.
We conducted an analysis of non-response bias based on the procedure described
by Armstrong & Overton (1977) and Lambert & Harrington (1990). We numbered
the responses sequentially in the order they were received and compared late responses
with early responses to all model variables using T-tests. We did not find
any noticeable pattern among the variables that could indicate the existence of a
non-response bias.
The theoretical model illustrated in Figure 1 was subjected to analysis using Structural
Equation Modelling (SEM). SEM is a very general linear statistical modeling
technique that encompasses factor analysis, regression, and many other estimation
methods as special cases. The proposed model had four latent variables or constructs:
internal integration in the Logistics-Production interface, internal integration
in the Logistics-Marketing interface, external integration, and company performance.
These constructs were not observed directly. Instead, they were measured
with error by instrumental variables. Typically, a model of structural equations
has two distinct parts which are analyzed simultaneously: the measurement
part and the construct part. The measurement part focuses on the relationship
between the observed measures and the latent constructs, while the construct part
focuses on the relationship between the latent variables.

Hasil (Bahasa Indonesia) 2:

tionnaire, yang merupakan tingkat tanggapan 32,3% (64/198). Tingkat respon ini
dianggap sangat memuaskan, sebagai calon peserta diminta untuk memberikan sensitif
data dan rahasia tentang kinerja mereka.
Kami melakukan analisis bias non-respon berdasarkan prosedur yang dijelaskan
oleh Armstrong & Overton (1977) dan Lambert & Harrington (1990 ). Kami bernomor
tanggapan berurutan dalam urutan mereka diterima dan dibandingkan tanggapan akhir
dengan tanggapan awal untuk semua variabel model menggunakan T-tes. Kami tidak menemukan
pola terlihat antara variabel-variabel yang bisa menunjukkan keberadaan
non-respon bias.
Model teoritis yang diilustrasikan pada Gambar 1 menjadi sasaran analisis menggunakan Structural
Equation Modelling (SEM). SEM adalah model yang sangat umum linear statistik
teknik yang mencakup analisis faktor, regresi, dan banyak estimasi lainnya
metode sebagai kasus khusus. Model yang diusulkan memiliki empat variabel laten atau konstruk:
integrasi internal di antarmuka Logistik-Produksi, integrasi internal
. Di antarmuka Logistik-Pemasaran, integrasi eksternal, dan kinerja perusahaan
konstruksi ini tidak diamati secara langsung. Sebaliknya, mereka diukur
dengan kesalahan oleh variabel instrumental. Biasanya, model persamaan struktural
memiliki dua bagian yang berbeda yang dianalisis secara simultan: pengukuran
bagian dan bagian konstruksi. Bagian pengukuran berfokus pada hubungan
antara tindakan diamati dan konstruksi laten, sementara bagian membangun
berfokus pada hubungan antara variabel laten.