First, because "practices" in "practices involved in the research process" isn't just plural because researchers have multiple different methods. Some researchers regard certain practices as unsound/invalid generally or in specific cases while others do not. The easiest example (in that it is quantitative) are disagreements over the use of statistical models, methods, and measures. A large portion of social science research relies on null hypothesis significance testing (NHST) a.k.a. "significance testing". Many researchers argue that there is no theoretical basis for this method and no empirical support either. As this method isn't just about the analysis of results but determines to some degree the research design, it is important to consider the when conducting social science inquiry.

Second, because methods evolve. A few decades ago, it would be hard or impossible to perform studies that are common today. Online research, for example, introduced a slew of methodological issues to consider, such as pipeline logic in online questionnaires, as well as new tools (such as online survey software services/SaaSs).

Third, because the sciences have become increasingly interdisciplinary. We find international conferences like HCII in which sociologists or psychologist give seminars on behavioral  research to engineers and computer scientists, machine learning in textual criticism, and quantum physicists working alongside neuroscientists. This interdisciplinary approach involves the transfer of methods from one field to others.

Finally, and most importantly, research isn't an algorithm. There is no "cookbook" containing all and only the sound "algorithms" to carry out a study. Every study involves particular issues that, while they are usually related to those of other studies, are still somewhat unique. Thus even when using common methods it is always important to understand why it is believed that these methods are sound and, granted they are, how they will be as used in any particular study.

As we can approach a specific research problem in a multifarious way, the researcher is expected to spell out his/her research protocol. then only the reader can relish it fully. there should not be any ambiguity in methodological issues discussed. the diversity of approaches is another reason for making social science research so dynamic.

First, because "practices" in "practices involved in the research process" isn't just plural because researchers have multiple different methods. Some researchers regard certain practices as unsound/invalid generally or in specific cases while others do not. The easiest example (in that it is quantitative) are disagreements over the use of statistical models, methods, and measures. A large portion of social science research relies on null hypothesis significance testing (NHST) a.k.a. "significance testing". Many researchers argue that there is no theoretical basis for this method and no empirical support either. As this method isn't just about the analysis of results but determines to some degree the research design, it is important to consider the when conducting social science inquiry.

Second, because methods evolve. A few decades ago, it would be hard or impossible to perform studies that are common today. Online research, for example, introduced a slew of methodological issues to consider, such as pipeline logic in online questionnaires, as well as new tools (such as online survey software services/SaaSs).

Third, because the sciences have become increasingly interdisciplinary. We find international conferences like HCII in which sociologists or psychologist give seminars on behavioral  research to engineers and computer scientists, machine learning in textual criticism, and quantum physicists working alongside neuroscientists. This interdisciplinary approach involves the transfer of methods from one field to others.

Finally, and most importantly, research isn't an algorithm. There is no "cookbook" containing all and only the sound "algorithms" to carry out a study. Every study involves particular issues that, while they are usually related to those of other studies, are still somewhat unique. Thus even when using common methods it is always important to understand why it is believed that these methods are sound and, granted they are, how they will be as used in any particular study.

I'm not entirely sure what you mean by "methodological theory." If you mean theory about how to use research methods that is one thing, but it is very different from the application of "substantive" theory. The latter comes from a specific research area and is used guide research, such as providing hypotheses for testing in quantitative research.

I myself would identify methodological theory topics such as the difference between constructivism versus post-positivism, etc. I personally am not convinced that those kinds of debates are relevant for most researchers most of the time. 

I agree with Andrew Messing's excellent response, and would add a fourth: that there is an important reflective and reflexive (per Schon 1987) element to qualitative research (and, I would argue, quantitative research, but that's a different discussion). One is an active process of thinking about your own history and experience, and how this might influence your participation in the research (the questions YOU might pose during interview, and how YOU might pose them) and thereby the findings of the research. The latter is best referred to Schon (1987): "Reflexivity is an interactive process that takes into consideration the relationship between self, other and context. Reflexivity expands the frame to include an examination of the underlying assumptions and priorities that shape interaction within a given time, place and situation (like a conflict). Being reflexive requires that parties examine their priorities before they react ..."

Methodological theory helps the investigators to situate personal and professional perspectives in research following such reflective and reflexive practices.