Research submissions to JMHC: perspectives from the associate editor.

I am extremely pleased to serve as the Journal of Mental Health Counseling (JMHC) associate editor for research, especially as the role affords me the opportunity to work with such fine colleagues as Drs. Jim Rogers, Alex Hall, and Bernie Jesolowski. I also look forward to working with the mental health counselors who submit their work for publication in JMHC; therein lies the true pleasure of editing. The opportunity to work with scholars doing cutting edge research and to read their contributions as they are drafted; fresh from the field, so to speak.

I am fortunate, too, in following in Jim Rogers' footsteps, he having previously served in my current role as research editor (2002-2005). His introductory column in the July 2002 issue of JMHC, in which he laid out his vision for strengthening research publications, leaves little to improve upon. Therefore, I will seek only to expand on his chief points in that editorial.

In terms of an overarching perspective, I agree with fully and will continue to support Rogers' (2002) idea that research articles in JMHC will reflect an appreciation for different methodologies relevantly serving to address research questions. As well, JMHC will encourage and seek to publish research submissions that are tied to clinical practice and that serve to build and test theory. Finally, although the JMHC editorial staff anticipates fielding submissions that possess a high level of competency with respect to research design, methodology, and statistical analyses, we recognize that research is definitely a "give and take" venture. As Gelso (1979) aptly points out, strengthening certain aspects of experimental design (i.e., increasing internal validity) will inevitably lessen the generalizability or clinical relevance of results (i.e., decrease external validity) and vice versa. Even good research has weaknesses.

My goal as an associate editor will be to foster the knowledge and learning of all mental health counselors who submit research-based works to the journal. Offering constructive feedback will be the mainstay provision so as to edify authors' work and strengthen their contributions. In my opinion, the role of editing is not only to serve a gatekeeping function, but also to help all mental health counselors, as a community, move forward together, learning and growing in our collective skill.

Rogers (2002) also made specific recommendations to authors of research submissions concerning increasing their attention to the theory that drives their research questions; to making accurate and explicit statements concerning the validity and reliability of instruments employed and constructs assessed; to report relevant sample-specific psychometric indices of reliability as a matter of course; and, to placing the significance of their findings within an appropriate context with respect to normative (e.g., absolute versus relative differences), sample (e.g., statistical power issues), and clinical benchmarks (e.g., clinical significance versus statistical significance). I cannot state any better than Rogers (2002) the importance of these issues. These recommendations will stand during my tenure as associate editor for research exactly as they are laid out in Rogers' remarks.

I will also note that Rogers' editorial made reference to a number of fine research and statistical resources (e.g., Pedhauzuer & Smelkin, 1991; Wilkinson et al., 1999). These are certainly recommended reading and reference for any mental health counselor submitting research articles to JMHC. I might also add a few other statistical (Pedhazur, 1997; Tabachnick & Fidell, 2001) and research design (Heppner, Kivlighan, & Wampold, 1999) resources to that list.

BUILDING A BETTER MOUSETRAP 

Moving a manuscript from initial submission to publication is a time-intensive effort for all involved. However, there are ways in which the process can be made easier. I hope that my affirmation of Rogers' (2002) comments will provide a significant "head start" to contributors. Indeed, any author who has addressed these aforementioned issues in their initial submissions will likely find the process of revision to be a much less arduous task!

In the following section, I would like to expound on Rogers' (2002) notion concerning the importance of theory-driven work.

Why Bother?

The armamentarium of the modern mental health researcher is impressive with respect to the research design and statistical tools she can bring to bear in research efforts. However, ultimately, whether one is repairing a plumbing fixture or conducting a research study, doing things right almost always means: (a) understanding the nature of the job to be accomplished; and (b) selecting the right tools for the task at hand.

"Why fuss?" some might say. "If it works, do it. I don't have the time to make sure everything is a perfect textbook job; that's not the way things work in the real world!" I understand this pragmatic point of view. And I will admit that, for example, you can actually substitute strips of newspaper for a plastic gasket when repairing a dripping pipe, but I wouldn't recommend it. Things tend to get leaky around the edges.

Likewise, using designs and statistical analyses in a project that do not appropriately fit the research question at hand will still yield results that you can try to publish; but again, I wouldn't recommend it.

No mental health counselor wants to entrust the standard of care she provides for her clients, the information she teaches to her students, or her credibility as a professional upon research literature that has a good chance of being leaky around the edges. So, although no piece of research will be perfect, it is important to make it as watertight as possible.

The Nature of the Job

A body of research, on any given topic of interest in mental health, is a living, growing, changing thing. Through publishing research articles, counselors contribute individual pieces to a body of knowledge. However, because the knowledge base in any given area is being simultaneously affected by a great many researchers, it is necessary for researchers to maintain a constantly updated "big picture" perspective of their area of interest so as to know what contributions are needed and worthwhile at any given time.

To understand what research contributions are needed and worthwhile is a complex and multi-faceted task. Three primary things a researcher must do are to gauge: (a) where current theory is at with respect to explaining and predicting the phenomenon of interest; (b) what empirically-supported facts have been established that explain and predict the phenomenon of interest; and (c) what the exact relationship is between extant theory and empirical research. With respect to the latter, invariably, theory almost always lags behind, is ahead of, or moves in a completely different direction than empirical research. Invariably, this situation has arisen because: (a) researchers are not taking time to get a complete big picture perspective before they conduct studies; and/or, (b) researchers are not doing theory-driven work. These "mistakes" are correctable though.

When an area of research has little or no a priori theory base to guide investigators, and perhaps has little extant empirical research to inform inductive reasoning about a phenomena of interest, qualitative methods and exploratory designs might be best employed. Qualitative methods and exploratory designs allow for the rich, indigenous description of phenomena, and help investigators to generate ideas about relevant questions to ask. In turn, this more foundational exploration of phenomena also allows for inductive thinking and theory-building to take place. Such methodologies as interviewing and content analyses can help to uncover common themes and relevant issues in an area of study. On the other hand, confirmatory style designs (e.g., comparative conditions) and/or statistical tools will be less helpful in this situation because investigators have little ability to know which variables to appropriately compare, what hypothetical predictions to make, or how to interpret findings when no theory is present.

However, when theory is present (even if it is errant in some ways) and when an empirical base of literature exists for investigators to draw upon, either exploratory or confirmatory methods may have a place. Theory provides scholars with a priori, supposed relations among variables; theory predicts what should happen among variables, and predicts their relation to outcomes. Essentially, theory is a proposed map of reality. Theory is to be tested and subsequently refined according to obtained empirical findings, through both deductive and inductive avenues.

Deductive vs. Inductive Approaches. Theory testing with hypothetico-deductive approaches (e.g., significance-testing) call for more confirmatory-style designs and statistical tools. Such efforts test those tenets put forth by a theory or model that suggest how a phenomenon works. The goal is to examine the theory map and find if indeed the proposed relations among variables and outcomes do indeed occur in reality the way the theory suggests they do.

Conversely, theory building via inductive methods, using exploratory designs, is employed to uncover or explain unknown relations among extant or new variables. Related, the statistical tools employed in these studies, either quantitative or qualitative, serve a different purpose. In exploratory designs, inductive reasoning is quite useful and necessary where a base of non-theory-driven empirical work is present (see Strong, 1991 for a discussion of naive empiricism). Examining an extent atheoretical empirical research base and sketching a theory map from available findings is a good first step in inductive reasoning (as an aside, this process is actually a good step even when a theory-driven body of empirical research is present because it helps investigators to get the best big picture perspective). Exploratory projects that seek to provide mortar among some loose or vaguely connected variables are very useful contributions. Sometimes, a well-focused qualitative study can inductively clarify and even add new pieces to a theory map that make previously disparate empirical findings suddenly fit within a newly apparent framework of understanding.

Regardless, whether using deductive or inductive avenues in research, the one goal common to all approaches concerns the building, testing, and refinement of theory. Collectively, the "nature of the job" is to develop (as much as possible) a base of stable and valid knowledge, where what is theorized about a phenomenon of interest reliably predicts what occurs in reality. Recursively, investigators also seek to find that what occurs in reality can be validly understood in the context of the theory that explains the phenomenon of interest. In short, we strive to get to the place where what we "think" is in alignment with reality, and where reality is in alignment with what we think.

This tenet of correspondence is a fundamental principle that mental health counselors, teachers, and researchers all strive to achieve. The mental health counselor strives to bring her clients' thinking, emotions, and behavior accurately in line with an evidence-based reality. Teachers strive to have students' knowledge and clinical behaviors brought accurately in line with an evidence-based reality. And of course, mental health researchers strive to bring their theories in alignment with an evidence-based reality that allows them to both accurately explain and predict phenomena. This is the nature of the job.

Selecting the Right Tools for the Task at Hand

Funny thing about tools; although frequently created for a specific task, one can almost always find a way to improvise with them. In a pinch, who has not hammered in a nail with a pair of pliers? This ambiguity of tool use also exists in research and makes things hard for investigators; there are many different designs and statistical tools that could be brought to bear on any given research question. To make matters worse, sometimes, one statistical tool can be used to answer two entirely different research questions! For example, regression procedures can be used to either inductively explain phenomena or to deductively and parsimoniously maximize the prediction of phenomena.

All this is to say that statistics and design issues are extremely fluid and flexible-like putty. They can serve several different purposes and address several kinds of research questions. The task of the researcher, once she understands the nature of the job she wants to accomplish, is to select the best design and statistical tools to answer the research questions at hand.

It is beyond the scope of this editorial to match methodological and statistical tools to specific jobs. Many venerable scholars have forgotten more about these issues than I will ever learn. I will refer the reader to Rogers' and my earlier references for works that offer detailed explanations of such issues. I can, however, offer a few basic heuristics that might be helpful to mental health researchers as they prepare projects intended for submission to JMHC.

Touchstone Questions

The first touchstone question will always be: "What is the exact nature of the research question I seek to ask?" For example, is the research question a comparative one (e.g., is A better than B or a control group?), exploratory (e.g., what are the relations among A, B, C and Z?), time-oriented (e.g., do scores on X change across different administration points?), efficacy related (e.g., does intervention X change characteristic Y), or confirmatory (e.g., does Q mediate, moderate, or cause Z?). The answer to this question helps to determine what design to employ, the type of hypotheses to be forwarded, and which statistical tools to employ. Of course, all research questions should either be driven by available theory and empirical literature or else seek to build theory.

A second touchstone question might be "Which design can best address and answer the research question at hand?" In great respect, answering this question well is highly contingent upon the answer to the first question and the theory being examined. For example, a study investigating the effect of an intervention might employ some type of pre-post design. However, the specific selection of which type of pre-post to employ is influenced by a myriad of factors such as subject availability, a desire to control for baseline effects, or a desire to guard against the outcome measure itself creating a bias. For example, an investigator might employ a Solomon four-design to address some of these issues.

A third touchstone question might be "What is the nature of the data (variables) with which I will work?" Continuous, non-fixed data (e.g., scores on a test) would be considered and treated in a different manner than, say, fixed, nominal data (e.g., eye color, sex). The nature of the data and extant theory help investigators to decide what tools are best for statistical analyses. For example, why artificially categorize (e.g., high, medium & low) continuous data (e.g., test scores) for use in an ANOVA procedure and lose its descriptive power, if in theory and reality such data are not applied to real life cases in a categorized fashion?

The fourth touchstone question might be "Which statistical procedures will best answer the research question at hand?" For example, if a research question is exploratory in nature (e.g., what are the relations among A, B, C and Z?), and is undertaken because available theory is incomplete, a mixed-method design and data analysis might be in order. That is, an investigator might collect both quantitative data (e.g., on instruments that tap constructs present in extant theory) and qualitative data (e.g., open-ended questionnaire data to address unknown gaps in the theory). The investigator might then elect to use regression-like analyses to see what variance in a construct of interest is captured by known theoretical variables. She then might elect to perform a content analysis on the questionnaire data to see if obtained emergent themes correspond to known theoretical variables or if her findings suggest the presence and influence of new, previously unknown variables. She may even be able to quantify these new thematic qualitative data and see what quantitative relation these new variables have to established variables in the theory.

Regardless, investigators' methods, research designs, and statistical analyses all must have a common, consistent thread running through them-all must work in a complementary fashion to effectively address the research question. This is selecting the right tool for the task at hand.

CONCLUSIONS

We all sometimes have a hard time gaining clarity on the complex, multifaceted concepts of research. Because this is so, I strongly encourage all authors of research works submitted to JMHC to specifically include a section on "Relevance to Clinical Practice." Just as research must be done with the goal of testing and refining theory, research must serve as the evidence base for clinical work. Research, theory and practice are the traditional trumvariate of the mental health profession.

In my mind, the processes of research and practice are exactly alike. There is utter artistry and creativity present in research efforts, there is confusing ambiguity in working with probability and chance, there are professional judgments necessary to determine if the data and results that appear before our eyes are to be trusted and acted upon, there are definite consequences when errors in judgment are made or when researchers fail to understand data and results in the context within which they present themselves.

Sound familiar? It should. These characteristics are all very similar to those that I and other mental health counselors associate with clinical practice. These exact same principles apply: artistry, creativity, ambiguity, risk, professional judgment, consequences of judgment error, and needing to understand clients in the contexts of their environment, culture, and personal idiosyncrasies.

The best advice I can offer to JMHC contributors is to think about research submissions in the exact same way they think about sound, methodical clinical practice. For example: (a) always collect a good clinical history on the client before you proceed (i.e., do a thorough literature review and clarify the exact nature of the big picture perspective); (b) clarify the presenting problem and collaboratively agree upon treatment targets with the client (i.e., assess current theory and empirical literature to determine what specific research questions are most appropriate to address at this point in time); (c) proceed intentionally, with a treatment plan based upon your theoretical orientation and established standards of care (i.e., choose a priori those methods and data collection procedures that address the research question in a theoretically relevant way); (d) collect and analyze relevant outcome data specific to the treatment target and interventions used (i.e., use valid instruments in research to measure constructs of interest and employ statistical tools that address the research question at hand): (e) evaluate clinical outcomes based upon what your theoretical orientation, previous experience, the literature, and the client's cultural factors suggest should have happened for the client (i.e., interpret statistical results with normative-, sample-, and significance-based contexts in mind and determine of results are in alignment with theory); (f) based upon observed clinical outcomes, make adjustments as needed either to your theoretical orientation or your interventions (i.e., confirm theory, inductively argue to change theory, or determine that your experiment was not a good test of theory as it is posed); (g) repeat as necessary until your theoretical orientation, therapeutic interventions, and clinical outcomes are all in alignment for the clientele you counsel (i.e., keep investigating until your theory map and reality are one in the same).

Research and practice are much closer in nature than is apparent at first face. Much the same way that the emotionally moving classical music we hear and experience is so very different yet exactly identical to its highly structured sheet music that we can hold and see, research and practice are one in the same-and both are driven by theory.

I very much look forward to the opportunity to share ideas with all contributors to JMHC. I am happy we will journey together in scholarship.

REFERENCES

Gelso, C. J. (1979). Research in counseling: Methodological and professional issues. Counseling Psychologist, 8, 7-35.

Heppner, P. P., Kivlighan, D. M., Jr., & Wampold, B. E. (1999). Research Design in Counseling (2nd ed.). Belmont, CA: Brooks Cole.

Pedhazur, E. J. (1997). Multiple Regression in Behavioral Research (3rd ed.). Orlando, FL: Harcourt Brace.

Pedhazur, E. J. & Schmelkin, L.P. (1991). Measurement, design, and analysis: An integrated approach. Hillsdale, NJ: Erlbaum.

Rogers, J. R. (2002). Looking back and moving forward: Research in the Journal of Mental Health Counseling. Journal of Mental Health Counseling, 24, 195-198.

Strong, S. R. (1991). Theory-driven science and naive empiricism in counseling psychology. Journal of Counseling Psychology; 38, 204-210.

Tabachnick, B. G., & Fidell, L. S. (2001). Using Multivariate Statistics (4th ed.). Needham Heights, MA: Allyn and Bacon.

Wilkinson, L., & the APA Taskforce on Statistical Inference (1999). Statistical methods in psychology journals: Guidelines and explanations. American Psychologist, 54, 594-604.

Loreto R. Prieto, Ph.D., is with the Department of Counseling, The University of Akron, Akron, OH. E-mail: [e-mail].