Implementasi Metode PLS-SEM
Diyanti Pratiwi (24031554016), Kartika Nur Savira (240315554049), Ivan Andika Setyawan (24031554085)
2026-05-24
Analisis Faktor-Faktor yang
Memengaruhi Keputusan Pembelian Obat OTC di Indonesia: Peran Mediasi
Digital Marketing dengan Pendekatan PLS-SEM
Program
Studi S1 Sains Data
Universitas Negeri Surabaya
1 Pendahuluan
PLS-SEM atau Partial Least Squares Structural Equation Modeling adalah metode statistik yang digunakan untuk menganalisis hubungan antar beberapa variabel sekaligus dalam satu model. Metode dipilih karena tidak mengharuskan data berdistribusi normal dan bisa bekerja dengan baik meski ukuran sampel tidak terlalu besar.
Penelitian ini menggunakan metode PLS-SEM untuk melihat faktor-faktor apa saja yang memengaruhi keputusan pembelian obat OTC di Indonesia. PLS-SEM dipilih karena metode ini bisa menganalisis hubungan antar variabel sekaligus tanpa harus memenuhi asumsi distribusi normal. Data yang dipakai berasal dari Kaggle dengan total 643 responden dan terdiri dari empat variabel laten yaitu Reference Group (GE), Brand Experience (BE), Purchase Digital Marketing (PDM), dan Purchase Brand Decision (PBD), di mana PDM dijadikan sebagai variabel mediasi dalam model ini.
Dataset : https://www.kaggle.com/datasets/ikhwanyudakusuma/dataset-over-the-counter-otc-indonesia
2 Load Data dan EDA
## No. No..Sampel Age Sex Education.Level Internet.Access
## 1 1 3246 17-25 years Female Senior High School Poor
## 2 2 4485 17-25 years Female Senior High School Good
## 3 3 4568 46-55 years Female Elementary School Good
## 4 4 3119 26-35 years Female Junior High School Good
## 5 5 2736 17-25 years Female Senior High School Good
## Provinsi PDM1 PDM2 PDM3 PDM4 PDM5 GE1 GE2 GE3 GE4 GE5 BE1 BE2 BE3
## 1 Jawa Timur 4 5 4 4 5 5 4 3 5 4 3 5 5
## 2 Jawa Tengah 5 5 5 5 5 5 5 5 5 5 5 5 5
## 3 Riau 5 3 5 3 5 5 5 4 5 4 4 4 5
## 4 Maluku Utara 4 4 3 5 5 4 5 3 5 4 4 4 4
## 5 Kalimantan Selatan 3 2 3 2 1 2 2 3 3 1 3 2 3
## PBD1 PBD2 PBD3 PBD4 PBD5 PBD6 PBD7
## 1 3 4 5 4 4 3 5
## 2 5 4 5 4 5 5 4
## 3 4 5 5 4 3 4 5
## 4 3 3 4 3 4 4 3
## 5 2 5 5 2 3 2 3## 'data.frame': 643 obs. of 27 variables:
## $ No. : int 1 2 3 4 5 6 7 8 9 10 ...
## $ No..Sampel : int 3246 4485 4568 3119 2736 4423 262 3956 4294 5169 ...
## $ Age : chr "17-25 years" "17-25 years" "46-55 years" "26-35 years" ...
## $ Sex : chr "Female" "Female" "Female" "Female" ...
## $ Education.Level: chr "Senior High School" "Senior High School" "Elementary School" "Junior High School" ...
## $ Internet.Access: chr "Poor" "Good" "Good" "Good" ...
## $ Provinsi : chr "Jawa Timur" "Jawa Tengah" "Riau" "Maluku Utara" ...
## $ PDM1 : int 4 5 5 4 3 5 1 5 5 5 ...
## $ PDM2 : int 5 5 3 4 2 3 1 5 4 4 ...
## $ PDM3 : int 4 5 5 3 3 4 1 4 3 5 ...
## $ PDM4 : int 4 5 3 5 2 3 2 4 3 3 ...
## $ PDM5 : int 5 5 5 5 1 5 1 3 5 4 ...
## $ GE1 : int 5 5 5 4 2 4 1 3 4 5 ...
## $ GE2 : int 4 5 5 5 2 3 2 1 4 3 ...
## $ GE3 : int 3 5 4 3 3 2 1 2 5 5 ...
## $ GE4 : int 5 5 5 5 3 4 3 3 4 4 ...
## $ GE5 : int 4 5 4 4 1 2 2 1 4 4 ...
## $ BE1 : int 3 5 4 4 3 2 3 3 3 5 ...
## $ BE2 : int 5 5 4 4 2 5 2 3 5 4 ...
## $ BE3 : int 5 5 5 4 3 3 3 4 4 5 ...
## $ PBD1 : int 3 5 4 3 2 3 3 5 3 4 ...
## $ PBD2 : int 4 4 5 3 5 5 3 4 3 4 ...
## $ PBD3 : int 5 5 5 4 5 5 3 3 4 4 ...
## $ PBD4 : int 4 4 4 3 2 3 2 4 5 4 ...
## $ PBD5 : int 4 5 3 4 3 4 2 5 4 5 ...
## $ PBD6 : int 3 5 4 4 2 3 3 4 4 5 ...
## $ PBD7 : int 5 4 5 3 3 3 3 3 3 4 ...## No. No..Sampel Age Sex Education.Level
## 643 603 6 2 7
## Internet.Access Provinsi PDM1 PDM2 PDM3
## 2 35 5 5 5
## PDM4 PDM5 GE1 GE2 GE3
## 5 5 5 5 5
## GE4 GE5 BE1 BE2 BE3
## 5 5 5 5 5
## PBD1 PBD2 PBD3 PBD4 PBD5
## 5 4 4 5 5
## PBD6 PBD7
## 5 5## No. No..Sampel Age Sex Education.Level
## 0 0 0 0 0
## Internet.Access Provinsi PDM1 PDM2 PDM3
## 0 0 0 0 0
## PDM4 PDM5 GE1 GE2 GE3
## 0 0 0 0 0
## GE4 GE5 BE1 BE2 BE3
## 0 0 0 0 0
## PBD1 PBD2 PBD3 PBD4 PBD5
## 0 0 0 0 0
## PBD6 PBD7
## 0 02.1 Persiapan Data
2.2 EDA
data_konstruk %>%
pivot_longer(everything(), names_to = "item", values_to = "nilai") %>%
mutate(Laten = case_when(
grepl("^PDM", item) ~ "PDM",
grepl("^GE", item) ~ "GE",
grepl("^BE", item) ~ "BE",
grepl("^PBD", item) ~ "PBD"
)) %>%
ggplot(aes(x = factor(nilai), fill = Laten)) +
geom_bar() +
facet_wrap(~ item, ncol = 5) +
labs(
title = "Distribusi Jawaban per Item Variabel Laten",
x = "Nilai",
y = "Frekuensi"
) +
theme_minimal()
library(tidyverse)
# Matriks korelasi
cor_matrix <- cor(data_konstruk, use = "complete.obs")
# Ubah ke format long
cor_long <- as.data.frame(as.table(cor_matrix))
# Heatmap
ggplot(cor_long, aes(Var1, Var2, fill = Freq)) +
geom_tile(color = "white") +
geom_text(aes(label = round(Freq, 2)), size = 3) +
scale_fill_gradient2(
low = "red",
mid = "white",
high = "blue",
midpoint = 0
) +
labs(
title = "Heatmap Korelasi",
x = "",
y = ""
) +
theme_minimal() +
theme(
axis.text.x = element_text(angle = 45, hjust = 1)
)
3 Uji Asumsi
3.1 Uji Kaiser-Meyer-Olkin (KMO)
Uji Kaiser-Meyer-Olkin (KMO) merupakan metode yang digunakan untuk mengukur kecukupan sampel berdasarkan proporsi varians di antara variabel-variabel yang mungkin disebabkan oleh faktor laten yang mendasarinya. Kriteria nilai Measure of Sampling Adequacy (MSA) lebih dari 0.5 maka data dianggap layak untuk dianalisis lebih lanjut
## Kaiser-Meyer-Olkin factor adequacy
## Call: KMO(r = data_konstruk)
## Overall MSA = 0.88
## MSA for each item =
## PDM1 PDM2 PDM3 PDM4 PDM5 GE1 GE2 GE3 GE4 GE5 BE1 BE2 BE3 PBD1 PBD2 PBD3
## 0.90 0.85 0.88 0.86 0.90 0.89 0.87 0.86 0.91 0.84 0.88 0.90 0.87 0.88 0.88 0.86
## PBD4 PBD5 PBD6 PBD7
## 0.91 0.94 0.86 0.923.2 uji non multikolinearitas
Uji multikolinearitas dengan VIF (Variance Inflation Factor) digunakan untuk mengetahui apakah terjadi hubungan yang terlalu kuat antar variabel independen dalam model. Nilai VIF yang baik umumnya kurang dari 5, sehingga menunjukkan tidak terjadi multikolinearitas dan model layak digunakan.
data_konstruk$Overall <- rowMeans(data_konstruk)
model_vif <- lm(Overall ~ PDM1+PDM2+PDM3+PDM4+PDM5+
GE1+GE2+GE3+GE4+GE5+
BE1+BE2+BE3+
PBD1+PBD2+PBD3+PBD4+PBD5+PBD6+PBD7,
data = data_konstruk)
vif_values <- vif(model_vif)
print(vif_values)## PDM1 PDM2 PDM3 PDM4 PDM5 GE1 GE2 GE3
## 1.696356 1.835807 1.710319 1.858230 1.614708 1.564608 1.558965 1.670487
## GE4 GE5 BE1 BE2 BE3 PBD1 PBD2 PBD3
## 1.628587 1.600319 1.497133 1.422511 1.371026 1.823024 1.904068 1.932083
## PBD4 PBD5 PBD6 PBD7
## 1.820541 1.667004 1.926414 1.7251664 Pembentukan Model PLS_SEM
4.1 Measurement Model (Outer Model)
4.2 Structural Model (Inner Model)
4.3 Estimasi Model PLS
5 Evaluasi Outer Model
5.2 Outer Loading
## GE BE PDM PBD
## PDM1 0.000 0.000 0.769 0.000
## PDM2 0.000 0.000 0.744 0.000
## PDM3 0.000 0.000 0.750 0.000
## PDM4 0.000 0.000 0.774 0.000
## PDM5 0.000 0.000 0.731 0.000
## GE1 0.719 0.000 0.000 0.000
## GE2 0.700 0.000 0.000 0.000
## GE3 0.733 0.000 0.000 0.000
## GE4 0.734 0.000 0.000 0.000
## GE5 0.720 0.000 0.000 0.000
## BE1 0.000 0.771 0.000 0.000
## BE2 0.000 0.755 0.000 0.000
## BE3 0.000 0.712 0.000 0.000
## PBD1 0.000 0.000 0.000 0.702
## PBD2 0.000 0.000 0.000 0.740
## PBD3 0.000 0.000 0.000 0.732
## PBD4 0.000 0.000 0.000 0.737
## PBD5 0.000 0.000 0.000 0.726
## PBD6 0.000 0.000 0.000 0.725
## PBD7 0.000 0.000 0.000 0.7235.3 Uji Reliabilitas dan Validitas Konvergen
reliability_df <- data.frame(
Konstruk = rownames(sum_pls$reliability),
Cronbach_Alpha = round(sum_pls$reliability[, "alpha"], 3),
rho_A = round(sum_pls$reliability[, "rhoA"], 3),
Composite_Rel = round(sum_pls$reliability[, "rhoC"], 3),
AVE = round(sum_pls$reliability[, "AVE"], 3)
)
print(reliability_df, row.names = FALSE)## Konstruk Cronbach_Alpha rho_A Composite_Rel AVE
## GE 0.774 0.786 0.844 0.520
## BE 0.607 0.611 0.790 0.557
## PDM 0.810 0.812 0.868 0.568
## PBD 0.851 0.854 0.887 0.5285.4 Uji Discriminant Validity (HTMT)
## GE BE PDM PBD
## GE . . . .
## BE 0.663 . . .
## PDM 0.535 0.462 . .
## PBD 0.392 0.653 0.473 .5.5 Uji Fornell-Larcker Criterion
## GE BE PDM PBD
## GE 0.721 . . .
## BE 0.469 0.746 . .
## PDM 0.439 0.338 0.754 .
## PBD 0.341 0.478 0.399 0.726
##
## FL Criteria table reports square root of AVE on the diagonal and construct correlations on the lower triangle.6 Evaluasi Inner Model
6.1 Path Coefficient
## PDM PBD
## R^2 0.215 0.295
## AdjR^2 0.213 0.291
## GE 0.360 0.060
## BE 0.169 0.365
## PDM . 0.2506.2 Nilai R-Square
r2_df <- data.frame(
Konstruk = colnames(sum_pls$paths),
R_squared = round(as.numeric(sum_pls$paths["R^2", ]), 3),
R2_Adj = round(as.numeric(sum_pls$paths["AdjR^2", ]), 3)
)
r2_df <- r2_df[r2_df$R_squared > 0, ]
print(r2_df, row.names = FALSE)## Konstruk R_squared R2_Adj
## PDM 0.215 0.213
## PBD 0.295 0.2916.3 Hasil Bootstrapping Jalur Langsung
## Original Est. Bootstrap Mean Bootstrap SD T Stat. 2.5% CI 97.5% CI
## GE -> PDM 0.360 0.363 0.054 6.611 0.262 0.463
## GE -> PBD 0.060 0.066 0.047 1.278 -0.024 0.160
## BE -> PDM 0.169 0.168 0.047 3.583 0.079 0.258
## BE -> PBD 0.365 0.367 0.044 8.237 0.276 0.451
## PDM -> PBD 0.250 0.245 0.049 5.126 0.154 0.339
## Bootstrap P Val
## GE -> PDM 0.000
## GE -> PBD 0.156
## BE -> PDM 0.000
## BE -> PBD 0.000
## PDM -> PBD 0.0006.4 Pengaruh Tidak Langsung (Indirect Effect)
## Original Est. Bootstrap Mean Bootstrap SD T Stat. 2.5% CI 97.5% CI
## GE -> PBD 0.090 0.088 0.020 4.421 0.051 0.132
## BE -> PBD 0.042 0.042 0.016 2.588 0.015 0.076
## Bootstrap P Val
## GE -> PBD 0.000
## BE -> PBD 0.0007 Analisis Mediasi
7.1 Perhitungan Efek Langsung dan Tidak Langsung
GE_PDM <- sum_boot$bootstrapped_paths["GE -> PDM", "Original Est."]
BE_PDM <- sum_boot$bootstrapped_paths["BE -> PDM", "Original Est."]
PDM_PBD <- sum_boot$bootstrapped_paths["PDM -> PBD", "Original Est."]
GE_PBD <- sum_boot$bootstrapped_paths["GE -> PBD", "Original Est."]
BE_PBD <- sum_boot$bootstrapped_paths["BE -> PBD", "Original Est."]
indirect_GE <- GE_PDM * PDM_PBD
indirect_BE <- BE_PDM * PDM_PBD
total_GE <- GE_PBD + indirect_GE
total_BE <- BE_PBD + indirect_BE
result <- data.frame(
Jalur = c("GE → PDM → PBD (indirect)",
"BE → PDM → PBD (indirect)",
"GE → PBD (total)",
"BE → PBD (total)"),
Efek_Langsung = round(c(NA, NA, GE_PBD, BE_PBD), 3),
Efek_TidakLangsung = round(c(indirect_GE, indirect_BE,
indirect_GE, indirect_BE), 3),
Efek_Total = round(c(NA, NA, total_GE, total_BE), 3)
)
print(result, row.names = FALSE, na.print = "-")## Jalur Efek_Langsung Efek_TidakLangsung Efek_Total
## GE → PDM → PBD (indirect) NA 0.090 NA
## BE → PDM → PBD (indirect) NA 0.042 NA
## GE → PBD (total) 0.060 0.090 0.150
## BE → PBD (total) 0.365 0.042 0.407GE_PDM <- 0.360
GE_PBD <- 0.060 # efek langsung GE ke PBD
BE_PDM <- 0.169
BE_PBD <- 0.365 # efek langsung BE ke PBD
PDM_PBD <- 0.250
# Hitung indirect & total
indirect_GE <- GE_PDM * PDM_PBD # 0.360 × 0.250
indirect_BE <- BE_PDM * PDM_PBD # 0.169 × 0.250
total_GE <- GE_PBD + indirect_GE
total_BE <- BE_PBD + indirect_BE
# Tampilkan tabel
result <- data.frame(
Jalur = c(
"GE → PDM → PBD (tidak langsung)",
"BE → PDM → PBD (tidak langsung)",
"GE → PBD (total)",
"BE → PBD (total)"
),
Efek_Langsung = round(c(NA, NA, GE_PBD, BE_PBD), 3),
Efek_TidakLangsung = round(c(indirect_GE, indirect_BE,
indirect_GE, indirect_BE), 3),
Efek_Total = round(c(NA, NA, total_GE, total_BE), 3)
)
print(result, row.names = FALSE, na.print = "-")## Jalur Efek_Langsung Efek_TidakLangsung Efek_Total
## GE → PDM → PBD (tidak langsung) NA 0.090 NA
## BE → PDM → PBD (tidak langsung) NA 0.042 NA
## GE → PBD (total) 0.060 0.090 0.150
## BE → PBD (total) 0.365 0.042 0.407indirect_df <- data.frame(
Hipotesis = c("H6: GE -> PDM -> PBD", "H7: BE -> PDM -> PBD"),
Beta = round(as.numeric(sum_boot$bootstrapped_total_indirect_paths[, "Original Est."]), 3),
t_value = round(as.numeric(sum_boot$bootstrapped_total_indirect_paths[, "T Stat."]), 3),
p_value = round(as.numeric(sum_boot$bootstrapped_total_indirect_paths[, "Bootstrap P Val"]), 3),
CI_2.5 = round(as.numeric(sum_boot$bootstrapped_total_indirect_paths[, "2.5% CI"]), 3),
CI_97.5 = round(as.numeric(sum_boot$bootstrapped_total_indirect_paths[, "97.5% CI"]), 3)
)
indirect_df$Keputusan <- ifelse(indirect_df$p_value < 0.05, "Signifikan", "Tidak Signifikan")
print(indirect_df, row.names = FALSE)## Hipotesis Beta t_value p_value CI_2.5 CI_97.5 Keputusan
## H6: GE -> PDM -> PBD 0.090 4.421 0 0.051 0.132 Signifikan
## H7: BE -> PDM -> PBD 0.042 2.588 0 0.015 0.076 Signifikan8 Visualisasi Model
grViz("digraph PLS_SEM {
graph [layout = dot, rankdir = LR, bgcolor = '#F8F9FA',
fontname = 'Helvetica', splines = curved]
# Node settings
node [shape = rectangle, style = filled, fontname = 'Helvetica',
fontsize = 14, penwidth = 2, width = 2.5, height = 1]
GE [label = 'Reference Group\n(GE)',
fillcolor = '#1565C0', fontcolor = 'white', color = '#0D47A1']
BE [label = 'Brand Experience\n(BE)',
fillcolor = '#1565C0', fontcolor = 'white', color = '#0D47A1']
PDM [label = 'Purchase Digital\nMarketing (PDM)',
fillcolor = '#E65100', fontcolor = 'white', color = '#BF360C']
PBD [label = 'Purchase Brand\nDecision (PBD)',
fillcolor = '#1B5E20', fontcolor = 'white', color = '#154814']
{rank = same; GE; BE}
{rank = same; PDM}
{rank = same; PBD}
# EDGES / JALUR
# H1: GE → PDM (signifikan)
GE -> PDM [label = 'H1: β=0.360***',
color = '#E91E63', fontcolor = '#E91E63',
penwidth = 3, fontsize = 11, fontname = 'Helvetica-Bold']
# H2: BE → PDM (tidak signifikan)
BE -> PDM [label = 'H2: β=0.169***',
color = '#E91E63', fontcolor = '#E91E63',
penwidth = 3, fontsize = 11, fontname = 'Helvetica-Bold']
# H3: PDM → PBD (signifikan)
PDM -> PBD [label = 'H3: β=0.250***',
color = '#E91E63', fontcolor = '#E91E63',
penwidth = 3, fontsize = 11, fontname = 'Helvetica-Bold']
# H4: GE → PBD (tidak signifikan)
GE -> PBD [label = 'H4: β=0.060 (ns)',
color = '#9E9E9E', fontcolor = '#9E9E9E',
penwidth = 1.5, fontsize = 11, style = dashed,
fontname = 'Helvetica']
# H5: BE → PBD (signifikan)
BE -> PBD [label = 'H5: β=0.365***',
color = '#E91E63', fontcolor = '#E91E63',
penwidth = 3, fontsize = 11, fontname = 'Helvetica-Bold']
}
")9 Visualisasi Full SEM Path Diagram Dengan Semplots
model_full <- '
# MEASUREMENT MODEL (outer model)
GE =~ GE1 + GE2 + GE3 + GE4 + GE5
BE =~ BE1 + BE2 + BE3
PDM =~ PDM1 + PDM2 + PDM3 + PDM4 + PDM5
PBD =~ PBD1 + PBD2 + PBD3 + PBD4 + PBD5 + PBD6 + PBD7
# STRUCTURAL MODEL (inner model / jalur antar konstruk)
PDM ~ GE + BE
PBD ~ PDM + GE + BE
'
# Estimasi model
fit_full <- sem(model_full,
data = data,
estimator = "MLR")
# Cek ringkasan model
summary(fit_full, fit.measures = TRUE, standardized = TRUE)## lavaan 0.6-21 ended normally after 43 iterations
##
## Estimator ML
## Optimization method NLMINB
## Number of model parameters 46
##
## Number of observations 643
##
## Model Test User Model:
## Standard Scaled
## Test Statistic 650.516 620.958
## Degrees of freedom 164 164
## P-value (Chi-square) 0.000 0.000
## Scaling correction factor 1.048
## Yuan-Bentler correction (Mplus variant)
##
## Model Test Baseline Model:
##
## Test statistic 4484.089 4233.965
## Degrees of freedom 190 190
## P-value 0.000 0.000
## Scaling correction factor 1.059
##
## User Model versus Baseline Model:
##
## Comparative Fit Index (CFI) 0.887 0.887
## Tucker-Lewis Index (TLI) 0.869 0.869
##
## Robust Comparative Fit Index (CFI) 0.888
## Robust Tucker-Lewis Index (TLI) 0.871
##
## Loglikelihood and Information Criteria:
##
## Loglikelihood user model (H0) -14045.165 -14045.165
## Scaling correction factor 1.222
## for the MLR correction
## Loglikelihood unrestricted model (H1) -13719.907 -13719.907
## Scaling correction factor 1.086
## for the MLR correction
##
## Akaike (AIC) 28182.330 28182.330
## Bayesian (BIC) 28387.773 28387.773
## Sample-size adjusted Bayesian (SABIC) 28241.725 28241.725
##
## Root Mean Square Error of Approximation:
##
## RMSEA 0.068 0.066
## 90 Percent confidence interval - lower 0.062 0.060
## 90 Percent confidence interval - upper 0.073 0.071
## P-value H_0: RMSEA <= 0.050 0.000 0.000
## P-value H_0: RMSEA >= 0.080 0.000 0.000
##
## Robust RMSEA 0.067
## 90 Percent confidence interval - lower 0.062
## 90 Percent confidence interval - upper 0.073
## P-value H_0: Robust RMSEA <= 0.050 0.000
## P-value H_0: Robust RMSEA >= 0.080 0.000
##
## Standardized Root Mean Square Residual:
##
## SRMR 0.057 0.057
##
## Parameter Estimates:
##
## Standard errors Sandwich
## Information bread Observed
## Observed information based on Hessian
##
## Latent Variables:
## Estimate Std.Err z-value P(>|z|) Std.lv Std.all
## GE =~
## GE1 1.000 0.514 0.634
## GE2 1.058 0.086 12.314 0.000 0.543 0.645
## GE3 1.088 0.077 14.139 0.000 0.559 0.684
## GE4 1.063 0.092 11.599 0.000 0.546 0.589
## GE5 1.079 0.089 12.120 0.000 0.554 0.646
## BE =~
## BE1 1.000 0.548 0.649
## BE2 0.866 0.106 8.186 0.000 0.475 0.538
## BE3 0.876 0.077 11.386 0.000 0.480 0.573
## PDM =~
## PDM1 1.000 0.512 0.682
## PDM2 1.122 0.085 13.251 0.000 0.574 0.684
## PDM3 1.127 0.086 13.050 0.000 0.577 0.683
## PDM4 1.288 0.120 10.776 0.000 0.659 0.722
## PDM5 1.127 0.090 12.492 0.000 0.577 0.630
## PBD =~
## PBD1 1.000 0.563 0.651
## PBD2 0.965 0.080 11.998 0.000 0.543 0.690
## PBD3 0.916 0.075 12.157 0.000 0.516 0.685
## PBD4 1.062 0.067 15.794 0.000 0.598 0.686
## PBD5 0.894 0.070 12.815 0.000 0.504 0.656
## PBD6 0.991 0.056 17.586 0.000 0.558 0.677
## PBD7 0.954 0.075 12.766 0.000 0.537 0.653
##
## Regressions:
## Estimate Std.Err z-value P(>|z|) Std.lv Std.all
## PDM ~
## GE 0.397 0.126 3.160 0.002 0.399 0.399
## BE 0.168 0.107 1.566 0.117 0.180 0.180
## PBD ~
## PDM 0.303 0.075 4.060 0.000 0.275 0.275
## GE -0.194 0.091 -2.133 0.033 -0.177 -0.177
## BE 0.656 0.124 5.279 0.000 0.639 0.639
##
## Covariances:
## Estimate Std.Err z-value P(>|z|) Std.lv Std.all
## GE ~~
## BE 0.181 0.021 8.464 0.000 0.644 0.644
##
## Variances:
## Estimate Std.Err z-value P(>|z|) Std.lv Std.all
## .GE1 0.392 0.024 16.566 0.000 0.392 0.598
## .GE2 0.415 0.032 12.825 0.000 0.415 0.584
## .GE3 0.356 0.027 13.069 0.000 0.356 0.533
## .GE4 0.561 0.040 14.132 0.000 0.561 0.653
## .GE5 0.428 0.029 14.992 0.000 0.428 0.582
## .BE1 0.413 0.035 11.954 0.000 0.413 0.579
## .BE2 0.553 0.048 11.539 0.000 0.553 0.710
## .BE3 0.472 0.039 12.205 0.000 0.472 0.672
## .PDM1 0.302 0.020 15.185 0.000 0.302 0.535
## .PDM2 0.374 0.029 13.099 0.000 0.374 0.532
## .PDM3 0.380 0.025 15.495 0.000 0.380 0.533
## .PDM4 0.399 0.028 14.187 0.000 0.399 0.479
## .PDM5 0.506 0.035 14.354 0.000 0.506 0.603
## .PBD1 0.431 0.028 15.220 0.000 0.431 0.576
## .PBD2 0.325 0.023 13.993 0.000 0.325 0.524
## .PBD3 0.301 0.022 13.824 0.000 0.301 0.531
## .PBD4 0.402 0.023 17.195 0.000 0.402 0.529
## .PBD5 0.335 0.021 15.828 0.000 0.335 0.570
## .PBD6 0.368 0.021 17.177 0.000 0.368 0.542
## .PBD7 0.389 0.023 16.577 0.000 0.389 0.574
## GE 0.264 0.047 5.554 0.000 1.000 1.000
## BE 0.300 0.044 6.822 0.000 1.000 1.000
## .PDM 0.188 0.035 5.310 0.000 0.716 0.716
## .PBD 0.167 0.026 6.444 0.000 0.527 0.5279.1 Visualisasi Full Model PLS_SEM
semPaths(fit_full,
what = "std",
whatLabels = "std",
style = "lisrel",
layout = "tree2",
rotation = 2,
# Ukuran node
sizeMan = 6,
sizeLat = 10,
sizeMan2 = 3,
# Label
edge.label.cex = 0.8,
node.label.cex = 0.7,
label.cex = 0.8,
# Warna
color = list(
lat = c("#1565C0", # GE = biru tua
"#6A1B9A", # BE = ungu
"#E65100", # PDM = oranye
"#1B5E20"), # PBD = hijau tua
man = "#E8EAF6" # indikator = biru muda
),
edge.color = "#37474F",
border.color = "#263238",
# Tampilan
residuals = TRUE,
intercepts = FALSE,
nCharNodes = 0,
curve = 0.5,
curvePivot = TRUE,
fade = FALSE,
mar = c(2, 2, 4, 2),
title = FALSE)
# Tambah judul manual
title("FULL SEM PATH DIAGRAM\nPerilaku Pembelian Obat OTC Indonesia",
cex.main = 1.4, font.main = 2, col.main = "#212121")
# Tambah legend
legend("bottomleft",
legend = c("Reference Group (GE)",
"Brand Experience (BE)",
"Purchase Digital Marketing (PDM)",
"Purchase Brand Decision (PBD)"),
fill = c("#1565C0","#6A1B9A","#E65100","#1B5E20"),
cex = 0.8, bty = "n", text.font = 2)