Dear Otman, I think it would be better to ask a slightly different question. I believe GeoGebra can support the development of higher order cognitive skills when it is used in certain ways, especially when students construct mathematical objects for themselves and are able to move between different representations, such as geometric and algebraic. If you were to research this topic independently from GeoGebra you could then see whether the methods that researchers propose can be implemented in GeoGebra. I think there has been a lot more research into systems like Cabri Geometre but their approaches are probably transferable.
I haven't done any work - but can refer you to some work of mathematics colleagues
Aventi, B., Serow, P., & Tobias, S. (2014). Linking GeoGebra to explorations of linear relationships. In J. Anderson, M. Cavanagh & A. Prescott (Eds.), Curriculum in focus: Research guided practice (Proceedings of the 37th annual conference of the Mathematics Education Research Group of Australasia) (pp. 79–86). Sydney, NSW: MERGA.
First, I apologize for my bad english! During the period 2005--2007, I was director of a research concerning the impact of "Apprenti Geometre" upon learning. Probably you do not know "Apprenti Geometre". That sof can be freely downloaded at the URL www.crem.be. It is another (french) dynamic geometry software, conceived both for primary schools and secondary schools. It contains tools such that "Glisser" (in english "slip" or "slide"), "Tourner" (Turn), "Retourner" (flip), "Dupliquer" (Duplicate), "Decouper" (Cut), "Diviser" (Divide), "Fusionner" (Merge) beside the corresponding geometric transformations (Translate, Rotate etc.) So it combines the physical approach and the geometric, static approach. We wanted to test exactly what you are looking for: what is the influence of the process of learning of such a soft. So we devised an experimentation with groups of pupils using the soft and other groups not using it but confronted to the same activities. We tested groups of pupils of primary schools (11 to 12 years old) and of secondary schools (13 years). I conjecture that the conclusions would not be very different with other dynamic geometry softwares (geogebra, cabri, etc.) We had some conclusions but we must be very prudent for we did not master all the organisational circumstances, specially the composition of the groups: we had to take whole classes of pupils and the classes were clearly not always equivalent. Moreover many differences between groups were not really significant. So I consider that our conclusions should be taken cautiously and been used mainly to initialise other experimentations io order to confirm them --- or not. Which we could not do because of financial restrictions to the research! Now, you can read our complete report, I have put it to day in ResearchGate under the title "Impact d'Apprenti Géomètre sur certains apprentissages". (I hope you can read french!) See specially parts III, IV and V. I just sketch two observations that I consider important.
First the role of perception in learning geometry is clearly visible through some of the activities proposed to pupils. It is even possible to distinguish three levels of perception: atomic, local and global. The role of this factor is much too often under-estimated.
Second: the use of the software has an influence, specially at the secondary school but that influence is not necessary related to the technical learning of geometry. Instead of that what we observed is a question of attitude in front of a problem to solve. The pupils from the experimental group, even when they no longer had a computer at their disposal were much more active to solve a problem. In a few words they an attitude nearer to that of a mathematician trying to solve a problem than the pupils of the "traditional" group.
I would like to thank you Guy for the information you shared us in your research.I understand that the geometric tool you applied in your research lead to positive effects concerning geometric perception, despite the restrictions you mentioned.I believe that the Geogebra tool ,like many other tools, have many positive effects in many aspects ,including geometric perceptions,Howerver, the aspect of developing HOCS,maybe, still needs investigation.
I believe that your research and the researches that Mr Kevin suggested could surely help!
A research team that includes colleagues at Illinois State has a preliminary study finding that supports reasoning skills among grades 3 and 4 students, as shown in improvements in the grasp of measurement for volume with rectangular prisms. We will present work at the PME/NA (Psychology of Mathematics Education) conference coming up in Fall 2015 to be held at Michigan State University. We are working to share our work soon from the follow up study.
The results are really interesting, especially when we are talking about supporting results in,relatively, low grades .I believe that it could also support reasoning and other aspects of higher grade students.
I will keep following your full research paper that would come later.
This recent "systematic review and meta-analysis" might be useful for you:
Chan, K. K., & Leung, S. W. (2014). Dynamic geometry software improves mathematical achievement: Systematic review and meta-analysis. Journal of Educational Computing Research, 51(3), 311-325.