Page 155 - 33-3

NTU Management Review Vol. 33 No. 3 Dec. 2023

higher than this benchmark, indicating that our data has no obvious multiple-collinearity
problem.
Lastly, we use Pearson correlation coefficient analysis to reflect the degree of close
relationships between the variables. The higher the absolute value of the correlation

coefficient between the two variables, the greater the mutual covariation. In general, when
there is a positive correlation between the two variables, X and Y, then Y will increase as
X increases; conversely, when there is a negative correlation between the two variables,

then Y will decrease as X increases. We use SPSS 22.0 to find the Pearson correlation
coefficients and their significance values. The results show that the correlation coefficient
of attractive alternatives and attention to switch other health apps is 0.174, and the p-value
is 0.031; therefore, the two variables achieve significant low correlation. Social influence
and intention to switch to other health apps have a correlation coefficient of 0.502 and

a p-value of 0; therefore, the two variables achieve a significant moderate correlation.
Dissatisfaction with the current health app and intention to switch to other health apps
have a correlation coefficient of 0.687 and a p-value of 0; therefore, both achieve a

significant moderate correlation. Procedural switching costs and intention to switch to
other health apps have a correlation coefficient of 0.212 and a p-value of 0.008; therefore,
both achieve a significant low correlation. Habits and attention to switch to other health
apps have a correlation coefficient of -0.318 and a p-value of 0; therefore, the two
variables achieve a significant low correlation. The interaction effect between procedural

switching costs and dissatisfaction with the current health app and intention to switch to
other health apps have correlation coefficients and the p-value is 0. Procedural switching
costs and intention to switch to other health apps have a correlation coefficient of 0.212,

indicating a significant low correlation. Dissatisfaction with the current health app and
intention to switch to other health apps have a correlation coefficient of 0.687, indicating a
significant moderate correlation.

5.2 Analysis of the Structural Model

After testing the reliability and validity of each construct, this study has confirmed
that each indicator has a certain degree of reliability and validity. Therefore, the
substantive relationship between the various indicators can be further verified. We use PLS

to check the structural model and estimate the structural model with path coefficients and

147