AB designs

[ASKurz]

Please leave suggestions for single-case or small-n studies using an AB design.

[ASKurz]

Consider Mancuso & Miltenberger (2015; https://doi.org/10.1002/jaba.267), Using habit reversal to decrease filled pauses in public speaking. Here's the abstract:

This study evaluated the effectiveness of simplified habit reversal in reducing filled pauses that occur during public speaking. Filled pauses consist of “uh,” “um,” or “er”; clicking sounds; and misuse of the word “like.” After baseline, participants received habit reversal training that consisted of awareness training and competing response training. During postintervention assessments, all 6 participants exhibited an immediate decrease in filled pauses.

This is an n = 6 multiple baseline AB design where the primary outcome is counts per minute. This article was used in Kate's class (file path: /BA Seminar/4. Behavior Analysis in Post-secondary Education/Mancuso & Miltenberger, 2016.pdf) and I used it in Statistical Analysis (fall 2021).

[ASKurz]

Consider Mason et al (2014; https://doi.org/10.1016/j.rasd.2013.12.014), Peer mediation to increase communication and interaction at recess for students with autism spectrum disorders. Here's the abstract:

Recess plays an integral role in the social and emotional development of children given the time provided to engage in interactions with others and practice important social skills. Students with ASD, however, typically fail to achieve even minimal benefit from recess due to social and communication impairments as well as a tendency to withdraw. Implementation of evidence-based interventions such as peer-mediated social skills groups, are necessary to ensure recess is an advantageous learning environment for students with ASD. A multiple-baseline design across participants was used to determine if a functional relationship exists between a social skills instructional program combined with peer networks with school staff as implementers and increases in level of communicative acts for participants with ASD at recess. Results indicate all participants demonstrated an immediate increase in the number of communicative acts with the introduction of the intervention. Implications for practice are discussed.

This is a nice n = 3 multiple baseline AB design where the primary outcome is unbounded counts. This article was discussed in Moeyaert (2018), within the context of a multilevel meta-analysis. Highlighting the Poisson/NB likelihoods would contrast with Moeyaert's strategy. This data set could also make for an easier introduction, which could lead into the Mancuso & Miltenberger data and its offset structure.

[ASKurz]

Consider Isaksson et al (2021; https://doi.org/10.1016/j.jbtep.2021.101637), Radically open dialectical behavior therapy for anorexia nervosa: A multiple baseline single-case experimental design study across 13 cases." Here's the abstract:

Background and objectives: No treatment for adult anorexia nervosa (AN) has shown sufficient effectiveness or superiority to other treatments. Overcontrol has been suggested as a viable mechanism to target in the treatment of patients with AN. Radically open dialectical behavior therapy (RO DBT) is developed for disorders related to maladaptive overcontrol. Our objective was to evaluate the outcome of RO DBT for AN in a clinical outpatient setting.

Methods: Thirteen adult female patients with mild to moderate AN provided written consent and entered a multiple baseline single-case experimental design study. Median age at eating disorder (ED) onset was 15 years and the median duration of the ED was 10 years. Individual changes were assessed weekly during a baseline phase (A) of four to six weeks, and during the subsequent 40-week RO DBT phase (B). Additional assessments were conducted before and after treatment, and at a six-month follow-up. Primary outcome was ED psychopathology. Secondary outcomes were psychosocial impairment, quality of life, social connectedness, and adaptive control strategies.

Results: Eight patients (62%) completed treatment. All completers were in full remission after treatment, with BMI ≥18.5 kg/m2 and ED psychopathology within one standard deviation of the community mean. Improvements occurred after introducing RO DBT, not during baseline.

Limitations: Participants were female with mild to moderate AN, limiting generalizability to severe AN or males.

Conclusions: The study provides preliminary support for using RO DBT in adult outpatients with AN and overcontrol. Further studies should replicate these findings.

Isaksson et al used an n = 13 multiple-baseline design to evaluate radically open dialectical behavior therapy for anorexia nervosa. The data are collected over a good number of time points and the authors showcased three primary outcomes in Figures 1 and 2. A nice thing about their three outcome measures is they all call for different likelihoods:

• BMI would be good for the Gaussian or gamma
• Restrictive eating would be good for an ordinal model
• Flexibility (as rated on a visual analogue scale) would be good for the zero-one-inflated beta

[ASKurz]

Consider Ruiz et al (2020; https://doi.org/10.3389/fpsyg.2020.00356), A multiple-baseline evaluation of acceptance and commitment therapy focused on repetitive negative thinking for comorbid generalized anxiety disorder and depression. Here's the abstract:

Repetitive negative thinking (RNT) is a core feature of generalized anxiety disorder (GAD) and depression. Recently, some studies have shown promising results with brief protocols of acceptance and commitment therapy (ACT) focused on RNT in the treatment of emotional disorders in adults. The current study analyzes the effect of an individual, 3-session, RNT-focused ACT protocol in the treatment of severe and comorbid GAD and depression. Six adults meeting criteria for both disorders and showing severe symptoms of at least one of them participated in the study. A delayed multiple-baseline design was implemented. All participants completed a 5-week baseline without showing improvement trends in emotional symptoms (Depression Anxiety and Stress Scale – 21; DASS-21) and pathological worry (Penn State Worry Questionnaire; PSWQ). The ACT protocol was then implemented, and a 3-month follow-up was conducted. Five of the six participants showed clinically significant changes in the DASS-21 and the PSWQ. The standardized mean difference effect sizes for single-case experimental design were very large for emotional symptoms $(d = 3.34)$, pathological worry $(d = 4.52)$, experiential avoidance $(d = 3.46)$, cognitive fusion $(d = 3.90)$, repetitive thinking $(d = 4.52)$, and valued living $(d = 0.92$ and $d = 1.98)$. No adverse events were observed. Brief, RNT-focused ACT protocols for treating comorbid GAD and depression deserve further empirical tests.

This is an $n = 6$ multiple baseline AB design with 2 primary outcomes which are sum-scores of Likert-type items. It might make a good example for the beta-binomial likelihood.