Advancing the Counseling Profession Through Contemporary Quantitative Approaches

Abstract

To advance the counseling profession, researchers must look to contemporary research designs that provide more thorough, consistent, or reliable results. In this article, we describe four unique quantitative methods: single‐case research design, dyadic data analysis, profile analysis, and nonlinear analysis. We detail the utility, application, and recommendations for use associated with each approach, including advances in the methods reviewed for future consideration in counseling research.

Reducing student anxiety using neurofeedback‐assisted mindfulness: A quasi‐experimental single‐case design

“…Completing the study within these parameters, while adhering to the preestablish treatment protocol, resulted in only two demonstrations of effect (Baseline to Intervention and Intervention to Postintervention). Thus, while increasing ecological validity (Ledford et al., 2018b), we decreased our internal validity, which required a more conservative interpretation of our results (Gast et al., 2018; Wood et al., 2021). Another limitation related to the design is that although participants were only directed to sit quietly for 5 min during the baseline phase, the auditory feedback they received from the NFB headband might have been passively training the brain to be calm.…”

“…A final limitation related to our design is potential carryover effects. While carryover effects have historically been a limitation in SCRD, in counseling research they are often expected as a sign of practical significance (Ray, 2015; Wood et al., 2021). That said, while carryover effects anticipated and observed in this study appeared to support the practical significance of the intervention, they made it difficult if not impossible to establish a true return to baseline measurement following the intervention, further limiting our ability to make causal inferences about the treatment effect.…”

“…For this project, we implemented an A–B–A with follow‐up quasi‐experimental SCRD due to the goodness‐of‐fit with the time limitations placed on the project by the research site's academic calendar (Wood et al., 2021). Although A–B–A designs are acceptable for evaluating counseling interventions (Cox et al., 2015; Gast et al., 2018; Ray, 2015) and are often used in counseling research (Brubaker & La Guardia, 2020; Cox et al., 2015; Greenleaf et al., 2019; Ikonomopoulos et al., 2015; Meany‐Walen et al., 2016; Ratanavivan & Ricard, 2018; Schomaker & Ricard, 2015; Smith et al., 2015; Swan et al., 2015; Swank et al., 2015), they are not without limitations.…”

“…According to current standards (What Works Clearinghouse, 2022), to qualify as the most rigorous designs (i.e., experimental), SCRDs must contain at least three demonstrations of effect (i.e., A–B–A–B or multiple baseline). Because A–B–A designs lack at least three demonstrations of effect, they are considered quasi‐experimental, and results are interpreted with additional caution (Gast & Ledford, 2018) and as exploratory instead of causal (Wood et al., 2021). That said, less rigorous withdrawal/reversal designs such as A–B–A can provide important information regarding counseling interventions (Gast et al., 2018; Ray, 2015; Ray et al., 2010), especially those that are newer and lack research support (Gallo et al., 2013) like in this study.…”

Examining coping and nonsuicidal self‐injury among adolescents: A profile analysis

“…Profile analysis is a special application of the multivariate analysis of variance (MANOVA) design, which subjects DV scores to the same scaling technique (e.g., z‐score), thereby creating what is known as commensurate level data (i.e., ratio; Tabachnick & Fidell, 2019). The advantage of profile analysis over MANOVA or descriptive discriminate analysis is that findings from profile analysis yields both statistical and visual evidence, creating greater utility of research findings for counseling professionals (Schmit et al., 2019; Wood et al., 2021). Within this approach, three profile analysis specific null hypotheses are evaluated using a single statistical analysis: (a) parallelism, (b) flatness, and (c) level.…”