PSY 3220 Module 6 Single-Subject, Quasi-Experimental and Developmental Research Paper: Designs Beyond the Randomized Experiment Example

When Random Assignment Is Not an Option: Other Ways to Study Phones and Focus

Student Name

Bachelor of Arts in Psychology, Thomas Edison State University

PSY-3220: Research in Experimental Psychology

Mentor Name

Month Day, Year

When Random Assignment Is Not an Option: Other Ways to Study Phones and Focus

Introduction

My main experiment randomly assigns adults to phone conditions. But some questions cannot be answered that way. An instructor might want to know whether a no-phones policy helps her class, without being able to randomly assign students. A counselor might want to know whether a strategy helps one student. A researcher might ask whether the effect changes with age. This paper compares three families of designs suited to these situations.

Single-Case Designs

Kazdin (2011) describes single-case designs, which study one person or a small number of people intensively, measuring behavior repeatedly over time. Instead of comparing groups, they compare phases. In an ABAB reversal design, a stable baseline (A) is followed by an intervention (B), then a return to baseline and a second intervention phase. If behavior changes each time the intervention is introduced and withdrawn, the pattern makes a strong case that the intervention caused the change. In a multiple baseline design, the intervention is introduced at different times across people, behaviors or settings, which avoids withdrawing a helpful intervention. Kazdin stresses stable baselines, enough data points in each phase and clear visual patterns.

How Well Single-Case Research Is Done

Smith (2012) systematically reviewed several hundred single-case studies published over a decade and compared them with emerging standards for the design. Multiple baseline designs were the most common. Many studies fell short of standards such as having enough data points in each phase, replicating the effect at least three times and reporting how data were analyzed. Smith urged researchers to adopt the standards so that single-case findings would be accepted as strong evidence.

Quasi-Experimental Designs

Shadish et al. (2002) define quasi-experiments as studies that manipulate or compare conditions without random assignment. The nonequivalent groups design compares existing groups, such as two class sections, one with a policy and one without, ideally measuring both before and after. The interrupted time series design measures an outcome many times before and after a change. Because groups are not randomly formed, these designs face threats to internal validity, including selection, in which groups differ from the start, history, in which other events occur at the same time, and maturation, in which people change naturally over time. The authors describe ways to strengthen inference, such as pretests, multiple comparison groups and careful matching.

Developmental Designs

Questions about age require their own designs. Cross-sectional designs compare different age groups at one time, which is efficient but confounds age with generation: older adults may differ because they grew up without smartphones, not because of age itself. Longitudinal designs follow the same people over time, which tracks real change but is slow and loses participants. Cross-sequential designs combine the two, following several age groups over time, so that age and generation effects can be separated.

An ABAB Study of My Own Studying

I designed a single-case study of myself. For ten study sessions (A), I will keep my phone on the desk and record pages read and a short quiz score. For ten sessions (B), the phone will be in another room. I will then return to the desk condition for ten sessions and again move the phone away for ten. If quiz scores rise each time the phone leaves and fall each time it returns, the pattern supports a causal effect for me. Following Smith, I planned at least ten data points per phase and graphed the results for visual analysis.

Comparing Two Course Sections

An instructor teaching two sections of an online course could ask one section to study with phones in another room for a month and compare quiz scores with the other section, after measuring both sections' scores in the month before. A pretest helps address selection, and checking for other differences, such as one section having an easier unit, addresses history.

A Cross-Sequential Plan

To ask whether the phone effect differs by age, a cross-sequential study could enroll learners aged 20, 35 and 50 and test them three times over four years, separating changes with age from differences between people who grew up with and without smartphones.

Choosing a Design

Each design trades strength of causal inference against practicality. The randomized experiment remains strongest for my main question, but these designs answer related questions it cannot.

Visual Analysis

Single-case data are usually judged by plotting each session and examining changes in level, trend and variability between phases, along with how quickly behavior changes when a phase begins. Kazdin notes that clear, immediate changes that repeat across phases are more convincing than small or delayed shifts. I will graph my sessions and ask a classmate who does not know the hypothesis to judge whether the phases differ.

Ethical Notes

A quasi-experiment comparing class sections must not disadvantage students. The comparison section would receive the same materials and could adopt the strategy after the study if it proved helpful.

Matching Design to Question

The randomized experiment asks whether phone location affects comprehension for adult learners in general. The single-case study asks whether it affects me. The section comparison asks whether a policy helps a real class, and the cross-sequential study asks whether the effect changes with age. Each answers a different question, and together they would give a fuller picture.

Limits

A single-case study of myself is not blind and cannot generalize to others. Quasi-experiments cannot fully rule out selection.

Conclusion

Kazdin showed how repeated measurement within one person can support causal conclusions, Smith showed the standards single-case studies must meet and Shadish, Cook and Campbell showed how quasi-experiments can be strengthened against threats to validity. An ABAB study, a section comparison and a cross-sequential plan extend my research to questions random assignment cannot reach.

References

Kazdin, A. E. (2011). Single-case research designs: Methods for clinical and applied settings (2nd ed.). Oxford University Press.

Shadish, W. R., Cook, T. D., & Campbell, D. T. (2002). Experimental and quasi-experimental designs for generalized causal inference. Houghton Mifflin.

Smith, J. D. (2012). Single-case experimental designs: A systematic review of published research and current standards. Psychological Methods, 17(4), 510-550. https://doi.org/10.1037/a0029312