5.2. Measurements
Persuasive system design (PSD) was measured using a 19-item scale (reduction, tunneling, tailoring, personalization, self-monitoring, rehearsal, suggestion, liking, social role, social facilitation, social comparison, normative influence, social learning, cooperation, trustworthiness, expertise, real-world feel, authority, and third-party endorsements). The scale was developed by the researchers using 19 design PSD principles out of 28 principles suggested by Oinas-Kukkonen and Harjumaa (2009). One item corresponds to each principle (e.g., Reduction: I can send messages on WhatsApp with few steps). However, nine principles (simulation, praise, rewards, reminders, similarity, competition, recognition, surface credibility, and verifiability) were ignored due to their inapplicability in WhatsApp.
Hofstede’s cross-cultural scale was measured using a 29-item scale adopted from Krüger (2016) [
56] to cover the 6 cultural dimensions: power distance (5 items), uncertainty avoidance (5 items), individualism (6 items), masculinity (4 items), long-term orientation (5 items), and indulgence (4 items).
Brand loyalty scale was measured using a 5-item scale [
57,
58]. A 5-point Likert ordinal scale (5—Highly Agree; 4—Agree; 3—Somewhat Disagree; 2—Disagree; 1—Highly Disagree) was utilized in this study.
5.4. Regression Moderation Analysis
To test the moderating effects of the national culture dimensions on the relationship between the persuasive system design (PSD) and user loyalty, an ordinary least squares regression moderation analysis, as suggested by Hayes (2017), was conducted to assess H1a, H1b, H1c, H1d; H2a, H2b, H2c, H2d; H3a, H3b, H3c, H3 d; H4a, H4b, H4c, H4d; and H5a, H5b, H5c, H5d.
To assess H1, the persuasive system design was entered as the main predictor (PSD), with power distance as the moderator (PD), and their interacting effect (PSD × PD) as a further predictor [
61].
H1a: The results obtained from the Netherlands’ respondents’ analysis (
Table 4) suggested that the interaction (PSD × PD) was not significant (b = 0.32, t (163) = 0.53,
p > 0.05). Therefore, H1a was not accepted. Power distance (PD) does not moderate the relationship between loyalty and persuasive system design for respondents in the Netherlands.
H1b: The results obtained from Germany’s respondents’ analysis (
Table 4) suggested that the interaction (PSD × PD) was significant (b = −0.76, t (121) = −1.99,
p < 0.05). The results showed that 36% of the variance in loyalty behavior was accounted for by the moderation model (R2 = 0.3638), with the regression slope being significant (F (121) = 3.9620,
p = 0.0488 < 0.05, LLCI = −1.5236, ULCI = −0.0041). Therefore, power distance (PD) does moderate the relationship between persuasive system design and loyalty for respondents in Germany.
In addition,
Figure 3 indicates that, in Germany, for people who are rated high in PD, there was no significant difference between the loyalty of those who rated high in the receptiveness to the PSD techniques and those who rated low in the receptiveness to the PSD techniques. However, for those who rated low in PD, the loyalty of those who rated high in the receptiveness to the PSD techniques was significantly higher than the loyalty of those who rated low in the receptiveness to the PSD techniques. Therefore, among users in Germany with low power distance, those who are more receptive to the PSD techniques show better loyalty than those who are less receptive, but this is not the case for users with high power distance, whose loyalty is high regardless of their receptiveness to the PSD techniques. Thus, H1b was accepted.
H1c: The results obtained from KSA’s respondents’ analysis (
Table 4) suggested that the interaction (PSD × PD) was not significant (b = −0.38, t (96) = −1.05,
p > 0.05). Therefore, H1c was not accepted. Power distance (PD) does not moderate the relationship between loyalty and persuasive system design for respondents in KSA.
H1d: The results obtained from Malaysia’s respondents’ analysis (
Table 4) suggested that the interaction (PSD × PD) was not significant (b = −0.49, t (92) = −1.36,
p > 0.05). Therefore, H1d was not accepted. Power distance (PD) does not moderate the relationship between loyalty and persuasive system design for respondents in Malaysia.
To assess H2, PSD was entered as the main predictor, with individualism as the moderator (IDV), and their interacting effect (PSD × IDV) as a further predictor [
61].
H2a: The results obtained from the Netherlands’ respondents’ analysis (
Table 5) suggested that the interaction (PSD × IDV) was not significant (b = −0.38, t (163) = −0.64,
p > 0.05). Therefore, H2a was not accepted. Individualism (IDV) does not moderate the relationship between persuasive system design and loyalty for respondents in the Netherlands.
H2b: The results obtained from Germany’s respondents’ analysis (
Table 5) suggested that the interaction (PSD × IDV) was not significant (b = −0.39, t (121) = −0.58,
p > 0.05). Therefore, H2b was not accepted. Individualism (IDV) does not moderate the relationship between persuasive system design and loyalty for respondents in Germany.
H2c: The results obtained from KSA’s respondents’ analysis (
Table 5) suggested that the interaction (PSD × IDV) was significant (b = −0.51, t (96) = −2.20,
p < 0.05). The results showed that 34% of the variance in loyalty behavior was accounted for by the moderation model (R2 = 0.3478), with the regression slope being significant (F (96) = 4.8590,
p = 0.0299 < 0.05, LLCI = −0.9764, ULCI = −0.0511). Therefore, individualism (IDV) moderates the relationship between persuasive system design and loyalty for respondents in KSA.
In addition,
Figure 4 indicates that, in KSA, for people who rated high in IDV, there was no significant difference between the loyalty of those who rated high in the receptiveness to the PSD techniques and those who rated low in the receptiveness to the PSD techniques. However, for those who rated low in IDV, the loyalty of those who rated high in the receptiveness to the PSD techniques was significantly higher than the loyalty of those who rated low in the receptiveness to the PSD techniques. Therefore, among users in KSA with low individualism, those who are more receptive to the PSD techniques show better loyalty than those who are less receptive, but this is not the case for users with high individualism, whose loyalty is high regardless of their receptiveness to the PSD techniques. Thus, H2c was accepted.
H2d: The results obtained from Malaysia’s respondents’ analysis (
Table 5) suggested that the interaction (PSD × IDV) was significant (b = −0.58, t (92) = −2.53,
p < 0.05). The results show that 56% of the variance in loyalty behavior was accounted for by the moderation model (R2 = 0.5600), with the regression slope being significant (F (92) = 6.4049,
p = 0.0131< 0.05, LLCI = −1.0390, ULCI = −0.1253). Therefore, individualism (IDV) does moderate the relationship between persuasive system design and loyalty for respondents in Malaysia. In addition,
Figure 5 indicates that loyalty in Malaysia increases when the user simultaneously has high PSD and low individualism. It shows that loyalty is low with low individualism and low PSD. Therefore, users in Malaysia who experience high PSD but have low individualism show better loyalty than users with high individualism. Therefore, H2d was accepted.
To assess H3, persuasive system design was entered as the main predictor (PSD), with masculinity as the moderator (MAS), and their interacting effect (PSD × MAS) as a further predictor [
61].
H3a: The results obtained from the Netherlands’ respondents’ analysis (
Table 6) suggested that the interaction (PSD × MAS) was not significant (b = 0.38, t (163) = 1.69,
p > 0.05). Therefore, H3a was not accepted. Masculinity (MAS) does not moderate the relationship between persuasive system design and loyalty for respondents in the Netherlands.
H3b: The results obtained from Germany’s respondents’ analysis (
Table 6) suggested that the interaction (PSD × MAS) was not significant (b = 0.68, t (121) = 1.13,
p > 0.05). Therefore, H3b was not accepted. Masculinity (MAS) does not moderate the relationship between persuasive system design and loyalty for respondents in Germany.
H3c: The results obtained from KSA’s respondents’ analysis (
Table 6) suggested that the interaction (PSD × MAS) was not significant (b = −0.17, t (96) = −0.69,
p > 0.05). Therefore, H3c was not accepted. Masculinity (MAS) does not moderate the relationship between persuasive system design and loyalty for respondents in KSA.
H3d: The results obtained from Malaysia’s respondents’ analysis (
Table 6) suggested that the interaction (PSD × MAS) was not significant (b = −0.51, t (92) = −1.86,
p > 0.05). Therefore, H3d was not accepted. Masculinity (MAS) does not moderate the relationship between persuasive system design and loyalty for respondents in Malaysia.
To assess H4, the persuasive system design was entered as the main predictor (PSD), with uncertainty avoidance as the moderator (UA), and their interacting effect (PSD × MAS) as a further predictor [
61].
H4a: The results obtained from the Netherlands’ respondents’ analysis (
Table 7) suggested that the interaction (PSD × UA) was not significant (b = −0.59, t (163) = −1.25,
p > 0.05). Therefore, H4a was not accepted. Uncertainty avoidance (UA) does not moderate the relationship between persuasive system design and loyalty for respondents in the Netherlands.
H4b: The results obtained from Germany’s respondents’ analysis (
Table 7) suggested that the interaction (PSD × UA) was not significant (b = 0.09, t (121) = 0.25,
p > 0.05). Therefore, H4b was not accepted. Uncertainty avoidance (UA) does not moderate the relationship between persuasive system design and loyalty for respondents in Germany.
H4c: The results obtained from KSA’s respondents’ analysis (
Table 7) suggested that the interaction (PSD × UA) was not significant (b = −0.10, t (96) = 0.30,
p > 0.05). Therefore, H4c was not accepted. Uncertainty avoidance (UA) does not moderate the relationship between persuasive system design and loyalty for respondents in KSA.
H4d: The results obtained from Malaysia’s respondents’ analysis (
Table 7) suggested that the interaction (PSD × UA) was not significant (b = −0.17, t (92) = −0.65,
p > 0.05). Therefore, H4d was not accepted. Uncertainty avoidance (UA) does not moderate the relationship between persuasive system design and loyalty for respondents in Malaysia.
To assess H5, the persuasive system design was entered as the main predictor (PSD), with long-term orientation (LTO) as the moderator, and their interacting effect (PSD × LTO) as a further predictor (Hayes, 2017).
H5a: The results obtained from the Netherlands’ respondents’ analysis (
Table 8) suggested that the interaction (PSD × LTO) was not significant (b = −0.50, t (163) = −1.37,
p > 0.05). Therefore, H5a was not accepted. Long-term orientation does not moderate the relationship between persuasive system design and loyalty for respondents in the Netherlands.
H5b: The results obtained from Germany’s respondents’ analysis (
Table 8) suggested that the interaction (PSD × LTO) was not significant (b = 0.09, t (121) = 0.19,
p > 0.05). Therefore, H5b was not accepted. Long-term orientation does not moderate the relationship between persuasive system design and loyalty for respondents in Germany.
H5c: The results obtained from KSA’s respondents’ analysis (
Table 8) suggested that the interaction (PSD × LTO) was not significant (b = 0.00, t (96) = 0.003,
p > 0.05). Therefore, H5c was not accepted. Long-term orientation (LTO) does not moderate the relationship between persuasive system design and loyalty for respondents in KSA.
H5d: The results obtained from Malaysia’s respondents’ analysis (
Table 8) suggested that the interaction (PSD × LTO) was not significant (b = −0.23, t (92) = −1.00,
p > 0.05). Therefore, H5d was not accepted. Long-term orientation does not moderate the relationship between persuasive system design and loyalty for respondents in Malaysia.
Figure 6,
Figure 7 and
Figure 8 show the hypotheses results for the four countries. As seen in those figures, masculinity, uncertainty avoidance, and long-term orientations do not indicate any significant moderating role in the four countries. According to the findings, only power distance and individualism have a significant moderating role. Power distance was found to have a significant moderating role in Germany, and individualism was found to have a significant moderating role in KSA and Malaysia.