Know the major theoretical constructs you are referring to in your research. Do not rely on secondary sources to explain theory: rely on primary sources for the theory and the numerous studies that have confirmed the strength of that theory.
As examples, connectivism and critical pedagogy are growing topics in C&RL submissions nowadays, but many authors have a cursory knowledge of how these theories originated and how they tie into even larger metatheories:
Connectivism draws extensively from the constructivism of Piaget and Vygotsky, yet connectivism has not been tested to any great extent so far to show its strength as a theory independent of constructivism. It is, rather, falsely assumed it is a strong, independent theory and is cited as such.
Qualitative research is composed of many different methods and epistemologies, too comprehensive for this site. Authors are encouraged to refer to these titles which can be found in most academic libraries:
Probably not, unless the same measurement has been used consistently and over time to develop strong theory. Descriptive statistics are qualitative in nature, because they describe the qualities of a group, phenomenon, responses, mean, etc. Since they are qualitative in nature, it is difficult to make generalizations about populations, yet we frequently see authors try to do this.
Many of the research studies C&RL reviewers see utilize convenience samples rather than random samples. Convenience samples have a high degree of several types of biases and are, therefore, extremely difficult to make generalizations to the broader research base with validity and reliability.
C&RL reviewers rarely see confidence internals included as a basis for the samples authors provide. This is an important aspect to determine whether the sample you are working with contains the population’s parameters. This, again, highlights the importance of using truly random samples rather than convenience samples in your research
Effect sizes are reported to show how much variance can be accounted for among significant variables within the total variance of your statistical model. C&RL reviewers rarely see effect sizes reported, but they are critical to reporting results in social and behavioral sciences.