FLOSS success-focused researchNaturally, I plan to continue and extend my dissertation research toward a more comprehensive model of FLOSS success. It can be improved and extended along multiple dimensions. In particular, future research can use more complex types of analysis, such as dynamic effects and sensitivity analysis as well as multi-group analysis. Introducing a time component into the research model seems like another interesting direction. Performing time-series analysis and latent curve analysis will make possible to study temporal aspects of FLOSS success.
With a goal of building a more comprehensive and accurate model of FLOSS success, improving the scope of future research studies is a good idea. In particular, the studies will include carefully selected and validated latent constructs and measures (indicators), which should improve the accuracy of the proposed models and results of data analysis. Improving statistical rigor of the future studies, such as performing statistical power analysis and estimation of effect sizes in SEM-focused research, is a worthwhile goal as well.
Furthermore, future FLOSS success research should and will improve data quality (through data cleaning) as well as data availability (through data sources integration), which will offer much more accurate models and results. Finally, I plan to convert my dissertation research software (DISS-FLOSS) into two separate R packages (one for automated FLOSS data collection and another for SEM-based FLOSS research frameworks) as well as to implement additional functionality for each of both packages.
Comprehensive FLOSS research
Using experience, gained during my dissertation research, as well as developed methodology, approaches and software, as a foundation, I plan to extend my research efforts to a larger area of studying FLOSS phenomenon, ecosystem and their various aspects. This promises to produce many more interesting, useful and important research studies toward on overarching goal of better understanding of FLOSS and building a comprehensive theory and model of the FLOSS phenomenon.
Complex socio-technical systems research
My dissertation research on FLOSS and its success factors additionally revealed to me how interesting and useful quantitative social sciences research can be. Specifically, I am very interested in using in research advanced statistical and other approaches, such as SEM and the ones, based on ML and AI theory and practice. I believe that these methods and techniques have enormous perspectives in science and I will do my best to master them in order to be able to perform research studies at the most comprehensive level possible. I am very enthusiastic about potential of studying complex socio-technical systems and plan to be involved in such research as much as possible.
Startup and venture ecosystems research
During early phases of my dissertation research I planned to integrate collection and analysis of data, related to startup and venture ecosystems, in order to study relevant factors in the context of open source ecosystem, in general, and their impact on FLOSS success, in particular. While I implemented software for data collection, using AngelList and CrunchBase APIs, later I have decided to not include those data sets into my dissertation research data analysis due to narrowing the focus of my research model. Nevertheless, based on that experience, I plan to continue working on this research stream, recognizing its theoretical and practical value and potential benefits.
Data science and quantitative social science research
Integrating various data science approaches, methods and tools into quantitative social science research methodology and practice is another topic of my research agenda. I am especially interested in studying novel data science methods that could be used as valuable tools in quantitative social science context. One of examples of such topics is studying approaches to integrating big data methods (i.e., machine learning) with causal inference. This one was inspired by running across materials from the Arthur M. Sackler Colloquium “Drawing Causal Inference from Big Data”, which was backed by U.S. National Academy of Sciences and took place in Washington, DC on March 26-27, 2015 (please see this video playlist of talks/presentations).