CHAPTER 5 FACTORS AFFECTING CONSUMERS' DECISION TO
5.4. The Binary Logistic Regression
To evaluate the influence of factors affecting the purchase decision of Highlands Coffee brand, the binary logistic regression method with four independent variables is product quality, price, store atmosphere, and reference group. In addition, the control variable demographic is also included in the model. The dependent variable (Y) has the values 0 and 1 (Y = 1: consumers decide to purchase products of Highlands Coffee brand. Y = 0: consumers decide not to purchase products of Highlands Coffee brand).
Estimation results by Robust method (error correction) show that the model is statistically significant at 1% (Prob > chi2 = 0.000), meaning the results have 99% confidence. The criterion -2LL (-2 log likelihood) = -37,898 shows the good fit of the overall model. The adjusted R-squared coefficient is 0.661, which means that 66.1% of the dependent variable Y is explained by the independent variable X. The Wald chi2 = 52.11 is quite large. The correct prediction rate of the whole model is 93.49%, which is quite high. Thus, this allows the conclusion about the estimated model that is consistent with the survey data, this result allows to explain the correlation between the independent variables and the dependent variable. The results of the binary logistic regression analysis are presented in the following table 15-18:
55
Table 5-18. Binary logistic regression analysis results
Variables Coefficients Std.Error Sig. dy/dx
Gender 0.450 0.820 0.583 0.024
Age -1.630 0.852 0.056* -0.087
Education -0.145 0.295 0.622 -0.008
Income 0.158 0.444 0.722 0.008
Student 4.026 1.087 0.000*** 0.214
Self-
employed 2.172 1.642 0.186 0.116
Officer staff 3.238 1.204 0.007*** 0.172
Civil servant 1.330 1.070 0.214 0.071
Product
quality 0.587 1.237 0.635 0.031
Price 1.036 0.431 0.016** 0.055
Store
atmosphere 1.924 0.811 0.018** 0.102
Reference
group 5.373 1.524 0.000*** 0.286
Constant -27.895 11.209 0.013
Observations 215
Log pseudolikelihood -37.898
Prob > chi2 0.000
Pseudo R2 0.661
Wald chi2 52.11
Prediction probability (%) 93.49
Source: The author’s survey, 2022 Note: *; **; *** is the 10%, 5% and 1% statistical significance level
Factors “gender”, “education”, “income”: The estimated results show that they do not affect the decision to purchase products of Highlands Coffee brand, not statistically significant at 5% (Sig. > 0.05). This means that despite different individual factors, there is no difference in the purchasing decision. The results are consistent with the study of Ngan (2013); Duyen (2015) but Penolazzi et al.
(2012); (Van der Merwe & Maree, 2016) because the authors argue that gender, education level and income affect consumers' decision to buy coffee.
Specifically, male tend to consume more coffee than female.
56 Factor “age”: The estimated results show that the variable age has a negative impact on the purchase decision of Highlands Coffee brand, which is statistically significant at 10%. With a probability of initial product selection of 50%, other factors remaining constant. When the age of consumers increases by one age group, the probability of purchasing products of the Highlands Coffee brand will decrease by 8.7%. This is consistent with the fact that the use of branded products is more known and consumed by young people than older people because they think that using branded products is to express fashion and affirmation to determine their position in society. Therefore, the results are consistent with the study of Tamaki (2013); Pattanakulkamjorn & Sawang (2021) found that consumers of working age and social behavior with friends, they tend to look for shops with modern design, suitable for drinking coffee.
Factor “occupation”: Take the unskilled workers variable as the reference variable. The estimated results show that there is a difference between students and officer workers (except self-employed and civil servants) in the decision to purchase products of Highlands Coffee brand, which is statistically significant.
at 1%. The results also show that if the subject is a student, the probability of buying products of the Highlands Coffee brand will be higher than that of other subjects with a probability of 21.4%. This is consistent with the research of Pattanakulkamjorn & Sawang (2021) because private sector workers frequently visit coffee shops to conduct business outside of normal business hours.
Factor “product quality”: The estimated results show that the product quality variable is not statistically significant at 5% because of the value Sig.= 0.635 > 0.05. Therefore, the hypothesis H1 that product quality has a positive impact on consumers' decision to purchase products of Highlands Coffee brand is rejected. Previous studies have shown that product quality has a positive impact on the purchase decision of famous coffee brands and consumers tend to care more about product quality than popularity (Vlahović et al, 2012; Rosanti & Salam, 2021; Burge, 2013; Dat, 2019; Duyen, 2015).
However, in this study, this factor has no effect at the 5% level of significance because consumers believe that branded products are of the same quality during production. Finding the right product for their needs is easy, so product quality is not a decisive factor in purchasing products of the Highlands Coffee brand.
Factor “price”: The estimated results show that the price variable is statistically significant at the 5%. Therefore, the hypothesis H2 that reasonable price has a positive impact on consumers' decision to purchase products of Highlands Coffee brand is accepted. With a probability of initial product selection of 50%, other factors remaining constant. When that product is priced
57 in line with consumers' spending, the probability of buying products of Highlands Coffee brand will increase by 5.5%. The results are consistent with the study of Prihastomo et al. (2019), Effendy et al. (2019), Dat (2019), Duyen et al. (2015), Ngan (2013) because the authors believe that reasonable prices will promote consumers' purchasing decisions higher. This is true of the fact that consumers have different views, where the price will influence the consumer's evaluation of a product. Consumers will buy coffee products if the benefits of these products are greater than what has been spent to get them.
The factor of “store atmosphere”: The estimated results show that the store atmosphere variable is statistically significant at the 5%. Therefore, hypothesis H3 is that store atmosphere has a positive impact on consumers' decision to purchase products of Highlands Coffee brand. With a probability of initial product selection of 50%, other factors remaining constant. When the store atmosphere creates comfort for consumers, the probability of purchasing products of the Highlands Coffee brand will increase by 10.2%. The results are consistent with the study of Burge (2013), Effendy et al. (2019), Jindal et al.
(2013), Duyen et al. (2015), Hieu (2019), Tram (2016) because the authors argue that good atmosphere is the most important aspect for branded coffee shops. This is true with the fact that when consumers want to choose a comfortable place to chat and work, the store atmosphere is one of the factors that lead to that decision. In addition, the atmosphere will partly affect our mood and work performance, so the better the atmosphere, the more satisfied customers will be and decide to choose a store.
Factor “reference group”: The estimated results show that the reference group variable is statistically significant at the 1%. Therefore, the hypothesis H4 that reference group has a positive impact on consumers' decision to purchase products of Highlands Coffee brand is accepted. With a probability of initial product selection of 50%, other factors remaining constant. When there is feedback from the reference group about the product, the probability of purchasing products of the Highlands Coffee brand will increase by 28.6%. The results are consistent with the study of Pattanakulkamjorn & Sawang (2021), (Jelison, 2017), Duyen et al. (2015), Dat (2019) because the authors believe that the opinions of people who are important to consumers will affect their decision to buy coffee. Therefore, the evaluation of people around leads to consumers' decision to purchase products of the Highlands Coffee brand. In other words, the more support from friends and family, the higher the decision to buy.
58