A research design begins with a spark of inspiration or opportunity, develops through a planning phase, comes to fruition in the stages of realization, and adds to knowledge and influences decisions as it is integrated into collective scientific understanding. At each step in the research design lifecycle, your specification of M, I, D, and A shapes your choices and how others will learn from your work.
This part of the book works through stages of the research design lifecycle. While it is presented linearly, the steps are all intertwined by their shared connection to the research design. We describe in each entry how we can use the declaration, diagnosis, and redesign framework to make progress at each step. Not every research project will feature each and every stage. For example, prospective research designs like experiments and surveys often include pilot studies to learn about important unknown features of M before implementing the full studies. Retrospective studies, like textual analyses of speeches delivered to parliament, might not.
We divide the research design lifecycle into four broad categories: brainstorming, planning, realization, and integration. Brainstorming is the process of going from the kernel of a research idea to a tractable design. Planning includes all of the activities you undertake before data collection starts: conducting ethical reviews, obtaining approvals, organizing partnerships, securing funding, running pilots, gathering criticism, and filing analysis plans. Realization begins with the execution of the data strategy as planned, and continues through the inevitable changes that come with analytic challenges and scientific surprises. We trace realization through implementation, pivoting in response to unexpected developments, writing up results, reconciling planned and implemented designs, and responding to peer reviewers. In the final phase of a research project, the results are integrated into the scientific literature. Integration includes how the study will inform theories and decisions and also how the study will later be reanalyzed, replicated, and meta-analyzed.