Mixed-Design ANOVA: Acupuncture vs. Massage for Back Pain
This paper presents a research proposal designed to assess the relative effectiveness of acupuncture and massage therapy in treating chronic back pain. Using a mixed-design ANOVA framework, the study treats time (before, mid-point, and post-treatment) as a within-subjects variable and treatment modality as a between-subjects variable, with 30 participants recruited from a spine health clinic. The proposal outlines the sampling procedure, variable definitions, pain measurement instrument (the McGill Pain Questionnaire), and the statistical criteria for hypothesis testing. It also addresses assumptions of normality and sphericity, discusses potential sources of bias, and considers how findings could inform clinical decision-making for spine specialists.
- Introduction and Research Question: Background on back pain and study rationale
- Methods and Participant Selection: Clinic setting, eligibility, and sampling procedure
- Variables and Procedures: Defining independent, dependent, and measurement variables
- Statistical Analysis and Hypothesis Testing: ANOVA testing, F-statistics, and post-hoc plans
- Assumptions, Error Control, and Limitations: Normality, sphericity tests, and Type I error control
- Discussion and Clinical Implications: Clinical relevance, limitations, and study significance
✍️ How to write this paper — guide, tools & examples ▾
What makes this paper effective
- The proposal clearly articulates a research question and translates it into formal null and alternative hypotheses for each factor and the interaction term, demonstrating methodological rigor.
- The rationale for choosing mixed-design ANOVA over alternatives (e.g., two-way repeated measures ANOVA) is explicitly stated, showing awareness of the distinctions between within-subjects and between-subjects designs.
- The paper addresses statistical assumptions (normality, sphericity) and names the corrective procedures (Shapiro-Wilk, Mauchly's test), reflecting practical awareness of Type I error risk.
- Limitations are acknowledged honestly, including the difficulty of controlling for confounding treatments outside the study setting.
Key academic technique demonstrated
The paper exemplifies research design justification — the writer does not simply choose a statistical test but explains why it fits the variable structure of the study. Linking the characteristics of the dependent and independent variables to the conditions required by mixed-design ANOVA is a sound technique for methods sections in quantitative research proposals.
Structure breakdown
The paper follows a standard quantitative research proposal format: introduction with background and research question, formal hypotheses, methods covering sampling and setting, a procedures section defining variables and measurement tools, a results section describing planned analyses, and a discussion addressing implications and limitations. References follow APA style.
Introduction and Research Question
Chronic back pain has become a serious health concern for professionals across the United States. The American Chiropractic Association estimates that approximately 31 million Americans experience chronic back pain at any given time, and that over 80% of adults are likely to experience some form of back pain at some point in their lives (ACA, 2015). Back pain currently stands as the leading cause of disability in the country, with the economy losing over $450 billion annually as a direct result (ACA, 2015). It is therefore important that health researchers focus on devising effective ways of preventing and treating chronic back pain.
The proposed study focuses on the treatment dimension of back pain. Acupuncture and massage are two of the leading treatment modalities for this condition, and the study is designed to determine which of the two approaches is more effective in addressing back-related problems over time. The research question guiding the study is:
"Which of the two treatment approaches is more effective in treating back-related problems over time?"
The study also aims to determine whether there are any significant interactions between the time taken in treatment and the treatment approach selected. The corresponding null and alternative hypotheses are as follows:
H₀: µA = µB — there is no difference in the means of the two factors.
Hₐ: µA ≠ µB — there is a significant difference in the means of the two factors.
H₀: µA1 = µA2 = µA3 — there is no difference in the means of factor A (acupuncture).
Hₐ: µA1 ≠ µA2 ≠ µA3 — there is a significant difference in at least two of the means of factor A.
H₀: µB1 = µB2 = µB3 — there is no difference in the means of factor B (massage).
Hₐ: µB1 ≠ µB2 ≠ µB3 — there is a significant difference between at least two of the factor B means.
H₀: C₁₂ = 0 — there is no interaction between factors 1 and 2 (time and treatment approach).
Hₐ: CAB ≠ 0 — there is a significant interaction between factors 1 and 2.
The hypotheses will be tested using the mixed-design ANOVA test. This test is appropriate for the study because the goal is to identify mean differences between groups split on two different factors — time and treatment modality — where the former is a within-subjects variable and the latter is a between-subjects variable (Lane, n.d.). If both independent variables were within-subjects factors composed of unrelated, independent groups, a two-way repeated measures ANOVA would have been preferred. However, research questions involving one within-subjects factor and one between-subjects factor lend themselves more effectively to the mixed-design ANOVA (Sukal, 2013).
Methods and Participant Selection
The study will be conducted at the Cleveland Center for Spine Health in New York. Thirty participants will be selected to take part in the study. To be eligible, a patient must be between 35 and 40 years of age and must have attended regular sessions at the center during the two months immediately preceding the study. This two-month eligibility requirement ensures that potential participants and their health statuses are well-known to the therapists at the clinic, enabling therapists to make informed decisions about whether a patient is suitable to participate in a given treatment program during the eight-week study period.
A memo will be displayed in all high-traffic areas of the clinic, explaining to potential participants how both they and the clinic stand to benefit from taking part in the study. Those willing to participate and satisfying the eligibility requirements will be required to register with the director and to specify the treatment modality they wish to join. Eligible participants will also be asked to indicate the number of massage and acupuncture sessions — if any — they have had in the three weeks immediately preceding the study.
The registration window will remain open for two weeks, after which it will be closed. Fifteen of those with the highest number of sessions in the acupuncture program, and 15 of those with the highest number in the massage program, will then be selected to participate in their respective programs. Selected participants will be notified and provided with details of how the study will be conducted via email. This systematic sampling procedure is intended to increase objectivity and minimize the risk of bias. The study will run for eight weeks beginning October 24, 2015. Participants' pain levels will be measured before the start of the program, four weeks into the program, and upon completion of the program.
Variables and Procedures
In this study, back pain is the dependent variable, while time and treatment approach are the independent variables. Time is a within-subjects variable, meaning that the same participants are measured at each level; treatment approach is a between-subjects variable (Clark-Cutter, 2004). Time will be categorized into three measurement points: before the program, four weeks into the program, and eight weeks into the program — making it a categorical, ordinal variable. Treatment approach will be defined in terms of the therapeutic method in which a participant is enrolled and will be measured as a categorical, nominal variable with two categories: acupuncture program (Treatment 1) and massage program (Treatment 2).
The dependent variable, back pain, will be defined in terms of how much a person hurts and suffers as a result of their back problems — specifically, how much their daily activities, movements, work, and leisure are affected. The McGill Pain Questionnaire, a 13-item instrument, will be used to measure participants' pain levels before and after treatment. It requires patients to describe the intensity and quality of their pain as: (i) mild, (ii) discomforting, (iii) distressing, (iv) horrible, or (v) excruciating, based on how much they are hurting and how much their daily activities have been affected. Numerical values of −2, −1, 1, 2, and 3 will be assigned to these five responses respectively, and the overall pain intensity score will be calculated by summing the numerical values across all 13 questions. This makes the dependent variable a continuous, interval-level measure.
References
ACA. (2015). Back pain facts and statistics. The American Chiropractic Association. Retrieved October 23, 2015, from
Clark-Cutter, D. (2004). Quantitative psychological research: A student's handbook. Psychology Press.
Lane, D. M. (n.d.). Online statistics education: A multimedia course of study. Rice University. Retrieved October 23, 2015, from http://onlinestatbook.com/
Sukal, M. (2013). Research methods: Applying statistics in research. Bridgepoint Education, Inc.
Create your account
Always verify citation format against your institution’s current style guide requirements.