1. This is an exploratory study using K-means clustering. The purpose of this technique is to identify groups of respondents with similar characteristics. The number of groups chosen by the experimenter (K) is usually three as it is here and the key characteristic that differentiates the groups is self-assessed digital competence. High, medium and low competency groups are distinguished and the other characteristics of respondents in these groups are investigated. The results show that these other characteristics are not markedly different between the three groups. High competence is associated with more recent graduation (rather than age itself) as is to be expected. The other significant differences (between groups shown in Table 2) were difficult to interpret and are not adequately reported or explained.

2. The authors then use binary logistic regression but in order to do so they need to dichotomise the dependent variable. Although this is legitimate I did not get much out of the results shown in Table 4.

3. The text reporting the results contains too much detail - too many numbers. It should be just a summary of the results. The Tables themselves (see Table 4 for example) do not really convey a clear message, even though the ‘significant’ variables are highlighted.

4. It is a self-selecting sample. Because it is not random any calculation of sample size is invalid. Moreover because the sample is self-selecting one assumes that those who chose to respond to the invitation were people who believed they had a higher than average level of competence.

5. I noticed a few minor grammatical errors:-

a) <<. K-cluster means analysis>> other way round. K-means clustering

b) <> emergence

c) <>

One-way Anova; Fischer’s; I’m not sure how Monte Carlo simulations can be used to examine dependence, though they could be used to analyse instruments (the self-report scales used)

d) <> represented

The newly-introduced text in red has helped clarify the difference between the three clusters. Additionally the limitation resulting from the self-selecting sample has been noted.

I thank the authors for their response to my earlier comments.