Dependent variable in regression model - sum of factors?

I got across below article due to one of our student used it for the brand loyalty measure.

I have a hard time to understand what is the dependent variable in this regression model - I did not find it explicitly mentioned, but it seems like a sum of all loyalty factors. Can it be so? My general logic does not allow me saying I proof my brand loyalty measurement questionnaire is valid if I check factor impact on their sum.

but maybe I am wrong, and I would like to hear other opinions (I reported it to publishers, but no response so far if they did any additional review...). Maybe you find it interesting to dig and share what you find about this paper statistics.

Tanks in advance for your opinions.
Last edited:
Adding a few citations from the paper, hopefully that makes it easier to understand the question.

Example of hypothesis (all in the same manner about all factors): “Higher level of emotional value will lead to higher level of brand loyalty.”

I have not found the description of what that “brand loyalty”, or later called simply “loyalty” is.

“The data were factor analysed using principal components analysis with varimax rotation.”

“Reliability was evaluated by assessing the internal consistency of the items representing each construct using Cronbach ’ s alpha. The reliability of each construct was as follows: Functional Value = 0.93; price worthiness = 0.92. Emotional value = 0.88; social value = 0.95; customer satisfaction = 0.70; brand trust = 0.88; commitment = 0.84; repeated Purchase = 0.96, involvement = 0.87.”

“The nine factors emerged with no crossconstruct loadings above 0.5, indicating good discriminant validity. The instrument also demonstrated convergent validity with factor loadings exceeding 0.5 for each construct. Consequently, these results confi rm that each of the fi ve constructs is unidimensional and factorially distinct and that all items used to operationalise a particular construct is loaded onto
a single factor.”

“The hypothesised relationships were tested using the multiple regression analysis of SPSS 11.5 for
Windows. The average scores of the items representing each construct were used in the data analysis. The R 2 was used to assess the model ’ s overall predictive fi t. Properties of the causal paths, including standardised path coefficients, t - values and variance, explained for each equation in the hypothesised model are presented in Figure 3 . The influence of perceived value (functional value, price worthiness, emotional value and social value), trust, customer satisfaction and repeated purchase commitment on loyalty has been proved by hypotheses H 1 , H 2a , H 2b , H 2c , H 2d , H 3 , H 4 , H 5 and H 6 .
As expected, repeated purchase ( b = 0.769, t -value = 7.159, p < 0.001) and functional value ( b = 0.138, t -value = 6.312, p < 0.001) have relatively strongest influence on loyalty, followed by commitment ( b = 0.127, t -value = 1.484, p = 0.148) and emotional value ( b = 0.108, t -value = 1.800, p = 0.082). Brand trust ( b = 0.095, t -value = 2.150, p < 0.05), price worthiness ( b = 0.046, t -value = 0.778, p = 0.443) ,customer satisfaction ( b = 0.034, t value = 1.523, p = 0.138) and social value ( b = 0.026, t -value = 1.207, p = 0.237) have a signifi cant positive effect on loyalty. Customers ’ involvement ( b = 0.057,
t -value = 2.622, p < 0.05) also has a signifi cant influence on loyalty. Therefore, hypotheses H 1 , H 2a , H 2b , H 2c , H 2d , H 3 , H 4 , H 5 and H 6 are supported. So the proposed model explained a significant percentage of variance in loyalty ( R 2 = 98.6 per cent, F value = 236.175, p < 0.001).”

Factor table and list of items attached.