\section{Research Design}
\label{sec:researchdesign}

To answer the two research questions presented in the previous section, we
conducted after-project surveys divided into three target groups of project
participants:

\begin{enumerate}
 \item \textit{MPOG Staff:} two government-side employees who have acted
directly in the platform development process. They were separately interviewed
by videoconference using the Hangouts platform. The interviews took an average
of 2 hours with 28 open questions divided by subject: Professional profile;
Organization, communication and development methodologies in the context of
government and project; Satisfaction with the developed platform; Lessons
learned.
 \item \textit{UnB students-developers:} 42 undergraduate students who
participated in any time of the project receiving scholarship. A
questionnaire with 45 closed and six open questions was sent through emails using
the Google Forms platform. The topics covered were: Organization, communication
and development activities between the respondents and the different groups of
the project; Learning acquired; Professional learning; Experience with FLOSS
projects. We received a total of 37 responses.
 \item \textit{Senior Developers:} eight advanced level researchers, MSc students or
IT market professionals who participated in some period of the project. A
questionnaire with 29 closed questions and 10 open questions addressed the
follow topics: Organization, communication and relationship between respondents
and distinct groups of the project; Development process; Experience with Free
Software. All eight recipients answered the questions.
\end{enumerate}

In a second round, we also collect post-mortem data from Redmine
(outside the SPB portal), Gitlab and Mailman (inside the SPB portal) - tools
used for management, communication and code versioning during the 30-month
project. The data of the tools provided by the SPB portal are open and available
for access at any time. For Redmine, we use a backup and instantiate the tool on
a server in UnB's research lab. The analysis of these data composes and ratifies
the evidences obtained in the previous round (surveys). The information obtained
represents, in terms of volume, interactions and the evolution of these
interactions between the government and academia teams, and, in terms of
development complexity, the platform size and quantity of software releases
delivered.


%Ambas possuem nível superior na área de TI e são funcionárias do governo a mais de 7 anos. Apenas uma delas segue atuando no mesmo ministério.
%Para o segundo grupo, um formulário online foi enviado aos 42 alunos e ex-alunos do curso de engenharia de software da UnB que participaram do projeto como bolsistas. A idade média dos 37 respondentes é de 25 anos e 91,9\% deles são do sexo masculino. Atualmente, 35,1\% seguem na universidade como alunos de graduação ou pós-graduação, 18,9\% atuam como desenvolvedor em empresa de pequeno porte e 18,9\% em empresas de médio ou grande porte, 10,8\% são empreendedores, 8,1\% estão desempregados e os demais atuam como professor ou funcionário público.
%O terceiro grupo é formado por 8 profissionais do mercado de TI ou alunos de mestrado (no período do projeto). Todos responderam ao formulário online enviado. As idades média é de 32 anos e 87,5\% são do sexo masculino. Eles possuem em média 11 anos de experiência no mercado de TI, e atualmente 25\% são estudantes de mestrado, 25\% empreendedores, 37,5\% são desenvolvedores freelancer e 62,5\% dos respondentes são funcionários de empresa. Já trabalharam em média em 5 empresas e participaram de 4 a 80 projetos. Eles participaram do projeto de colaboração estudado entre 7 a 24 meses.

% And finally, we analized Colab code before and after the project to evaluate how much effort was spent to use this software as a component of the platform.