PLS merupakan teknik analisis data untuk menganalisis hubungan di antara satu set blok variabel. Berdasarkan hasil model outer pada pengujian Covergent validity didapatkan nilai loading factor semua indikator > 0,5 maka indikator merefleksikan variabel latennya, pada pengujian composite reliability nilai dari semua variabel laten , kualitas pelayanan, kepercayaan konsumen, komitmen, kepuasan konsumen dan loyalitas konsumen memiliki nilai composite reliability atau cronbach’s alpha > 0,7 maka suatu model konstruk dikatakan memiliki validitas dan reliabilitas yang baik. pengujian validitas model konstruk diketahui model konstruk dari variabel laten memiliki validitas yang baik karena nilai AVE dari semua variabel laten > 0,5. interpretasi variabel dijelaskan dengan nilai koefisien determinasi (R-Squared) dimana nilai koefisien determinasi dari model 1 sebesar 0,888 atau sebesar 88,8%, artinya variabel laten eksogen yaitu kualitas pelayanan, kepercayaan konsumen dan komitmen mampu menjelaskan variabel laten endogen yaitu kepuasan konsumen sebesar 88,8%. model 2 nilai koefisien determinasi dari model 2 sebesar 0,906 atau sebesar 90,6%, artinya variabel laten eksogen yaitu kualitas pelayanan, kepercayaan konsumen dan komitmen serta kepuasan konsumen mampu menjelaskan variabel laten endogen yaitu loyalitas konsumen sebesar 90,6%. pengujian hipotesis didapatkan hanya variabel laten komitmen yang berpengaruh terhadap kepuasan konsumen dan loyalitas konsumen karena nilai t hitung > 1,96. (Kata kunci: PLS, AVE, composite reliability, Covergent validity, R-squared, pengujian hipotesis )

PLS is a data analysis technique for analyzing relationships between a set of blocks variabel.. Based on the test results of the model outer Covergent validity values obtained loading factor all indicators> 0.5, the indicator reflects the latent variables, the reliability of composite test scores of all latent variables, service quality, consumer confidence, commitment, customer satisfaction and customer loyalty has a composite value reliability or cronbach's alpha> 0.7 then construct a model is said to have good validity and reliability. testing the validity of the model constructs known constructs of latent variable models have good validity for value AVE of all latent variables> 0.5. interpretation of variables is described by the coefficient of determination (R-Squared) where the coefficient of determination of model 1 by 0.888, or by 88.8%, meaning that the exogenous latent variable quality of service, consumer confidence and commitment to be able to explain the endogenous latent variables namely customer satisfaction by 88 , 8%. model 2 coefficient of determination of model 2 amounted to 0,906, or by 90.6%, meaning that the exogenous latent variable quality of service, consumer confidence and commitment as well as customer satisfaction is able to explain the endogenous latent variables that consumer loyalty of 90.6%. hypothesis testing only latent variable obtained commitments that affects customer satisfaction and customer loyalty because t value> 1.96. (Keywords: PLS, AVE, composite reliability, validity Covergent, R-squared, hypothesis testing)

Baca Jurnal Baca Jurnal Baca Jurnal Baca Jurnal Baca Jurnal Baca Jurnal Baca Jurnal Baca Jurnal