The Development of Cognitive and Noncognitive Skills in Students in the Autonomous Province of Trento

Round 1

Reviewer 1 Report

7 Consider the use of the term "recent' as will become inaccurate over time and detract from the reference.

63 again, consider revising "more recently" 

78 Consider adding the sample size

112 Consider revising "extremely rich and of high quality" unless this is a shared view which would need to be cited and referenced. Beware of over-use of superlative adjectives.

253 Consider describing the type of interview (structured, unstructured etc)

404 Consider expanding upon "Interestingly". Suggesting that homework increases CS whereas household chores do not may be seen as an obvious finding.

478 Consider identifying this as a prime subject for further research

530 This is a refreshing discussion regarding limiting factors. Are there more you would like to add?

This is a very well written, interesting and comprehensive piece of research which I have thoroughly enjoyed reviewing. Your comprehensive approach is clear and your results concise. If you wished to discuss whether your research in more inductive than deductive it may add more context to the presentation of your findings.

A little more background regarding the geographical setting of your research (Northern Italy, mountainous region also with urban areas) would be very welcome.

Please take advice from the journal editors as to whether they will accept footnote referencing within the body text. As you have a comprehensive referencing section, you may wish to consider removing the footnotes.

I hope my comments are of use to you and look forward to seeing the published article.

Author Response

please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

The paper examines, through the statistical analysis of sample surveys, two empirical questions pertaining to the children acquisition, via schooling, of Non-Cognitive Skills (NCS) in the context of the Trento province in Italy:

1) Are the children' NCS impacting the children Cognitive skills (CS) (measured by the Italian INVALSI administrative data reporting children test scores in Italian language and mathematics) ?

2) Are some specific teaching practices (such as performing educational activities on the themes of citizenship, legality, and respect for the  environment) effective in improving NCS ?

These two questions are handled by constructing an integrated data set based partly on survey conduced at selected schools and secondary administrative data on the children test scores and family backgrounds and by running standard statistical analysis on the data. Question 1 is handled by regressing - using Generalized least square estimation- the CS variables (one regression for the math score, one regression for the reading scores) on various NCS variables and family control variables. The analysis suggests that NCS variables "explains" in a statistically significant way the CS variables. However, the usual caveat should apply to the interpretation given to these results. Namely, the fact that the coefficients that multiply the NCS variables in the regressions have statistically significant values does not indicate in any way that the children's NCS affect their CS ? The only thing that the results show is that the student who happens to have the specific favorable (say) NCS (for example "Emotional Stability") tend also to be those student who do well in math or in reading. But this does not say anything about causality, and can not serve as a recommendation for promoting the children "emotional stability" as a policy tool for improving his or her performance in mathematics. Perhaps it is the good performance in mathematics that brings to the child an "emotional stability". Perhaps there is an unobserved child's trait that determines jointly both his/her emotional stability and her performance in mathematics. I believe that the authors should therefore acknowledge this difficulty of inferring any causal relation from that first part of the analysis.
Switching now to the second part of the analysis - which examines the impact of specific teaching practices on the NCS acquisition using difference-in-differences methods - I also believe that some caution should be made in interpreting the results. If I understood correctly, the authors have partitioned their sample of schools in two groups: one (made of some 108 schools) where some specific teaching interventions on environment or citizenships have been implemented and some where these interventions were weak or absent. 
The authors have then treated these two groups as if they were differing only by the fact that one group has been the object of specific teaching practices and the other has not. But this is not so. The authors have not run a real experiment by choosing the two groups of school at random. Because of this, there might be, here again, unobserved heterogeneity that prevents one from making causal inference from the results. Perhaps the schools where  important teaching interventions on environmental matters or citizenships have been conduced are also schools with children with better NCS, and one may even imagine that the parents of these children may be more demanding of these kinds of interventions than the parents of children with lower NCS. I understand that the authors have done some descriptive analysis of the two samples that seem to suggest that the distributions of observable variables is not too different in the two groups of schools. But it is sufficiently different to suggest that the dynamic of acquisition of NCS has been better in the schools that had these teaching interventions than in the school that hadn't.

Hence, I believe that the authors should bring several caveats to their analysis so as to acknowledge this. 

Author Response

please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

I believe that the new version of the paper has correctly addressed some concerns raised in my earlier report. In my view therefore, the paper can be published as it is, modulo perhaps a last proofreading that could eliminate small typos.