Professional Documents
Culture Documents
a. Discuss data collection (what did they do well/could have done better)
Authors collected data from 3519 individuals, including 2044 internet banking users without offering
monetary incentives to avoid bias. The authors could have done better by providing more information
about the representativeness of the sample. They did not provide information on how the sample was
selected or on the characteristics of non-respondents, which could affect the external validity of their
findings. In addition, since the sample was drawn from an online portal, the results may not be
representative of the entire population of internet banking users in Poland. However, the authors did
well by using a structured survey instrument and by measuring all the variables in a consistent and
standardized manner.
b. Discuss the variables (what did they do well/could have done better)
The variables used in the study are categorized into different groups, such as perceived security,
internet experience, marketing exposure, mobile banking, ownership of different types of cards,
broadband access, and demographic factors. Some of the variables used in the study are quite well-
defined, such as perceived security, internet experience, ownership of different types of cards, and
broadband access. For instance, the perceived security of internet transactions is measured on a five-
point scale, ranging from very unsafe to very safe. However, some of the variables could have been
better defined or measured. For example, the variable measuring exposure to internet banking
advertisements is binary, indicating whether the respondent has come across at least one
advertisement or not. It does not take into account the frequency or type of advertisements that the
c. Discuss the choice of analytical techniques (e.g., negative binomial regression or logit model
or OLS regression)
In this study, the dependent variable is whether or not an individual has adopted internet banking,
which is a binary outcome (yes or no). Therefore, using a binary logistic regression is a suitable
choice for analyzing the factors that affect the adoption of internet banking in Poland. Furthermore,
binary logistic regression can account for the presence of multiple predictor variables and can provide
1|Page
GBA 7023 Take-Home exam
estimates of the strength and direction of the relationship between each predictor variable and the
outcome variable. This allows researchers to identify the key factors that influence the adoption of
Overall, the summary statistics presented in the paper provide a good overview of the sample
characteristics and key variables used in the analysis. However, there are a few potential concerns that
should be considered. First, it is unclear from the summary statistics whether any missing data or
outliers were removed from the sample. Second, it is not clear what units are used for the
"Area_of_Residence" variable. Finally, while the summary statistics provide a good overview of the
sample, they do not provide information on the distributional properties of the variables, such as
skewness or kurtosis.
The presented result is a regression analysis using a binomial logistic model to examine the predictors
of internet banking adoption. The coefficients, standard errors, z-statistics, and marginal effects of
each variable are reported in Table IV. The results indicate that perceived security is a strong
predictor of internet banking adoption, as individuals with lower perceived security have a
significantly lower probability of adopting internet banking. The regression model has a high
predictive power, as indicated by the higher log likelihood compared to the restricted log likelihood
and the correct classification of 85.08% of respondents. The Hosmer and Lemeshow Goodness-of-Fit
test also supports the suitability of the model. Overall, the results provide support for the hypotheses
that familiarity with the internet and marketing campaigns positively affect internet banking adoption,
2|Page