Praxis 5331 Research Methods & EBP: What to Know

Research methods are a quietly high-yield slice of the Praxis 5331. Roughly 5–10% of items require you to interpret a study design, apply an evidence hierarchy, or judge the strength of a treatment claim. It maps directly to ASHA's Evidence-Based Practice standard, and it's the area grad programs teach in a single semester and then never revisit. Here's the focused review.

What the 5331 actually tests about research

• Evidence hierarchy. Which level of evidence is strongest for answering a clinical question.
• Study design recognition. Given a described study, identify whether it's a RCT, case series, cohort, or single-subject.
• Basic biostatistics. Mean, median, standard deviation, effect size, confidence interval — what they mean, not how to calculate them.
• Reliability vs. validity. Test–retest, inter-rater, construct, criterion.
• EBP triad. Best available evidence + clinical expertise + client values/preferences.

The evidence hierarchy — memorize this ranking

1. Systematic reviews and meta-analyses — highest level for treatment questions.
2. Randomized controlled trials (RCTs) — the gold standard for a single study.
3. Cohort studies — follow groups over time, no randomization.
4. Case-control studies — compare groups with and without an outcome retrospectively.
5. Case series and case reports — descriptive, no control group.
6. Expert opinion — lowest level of external evidence.

On the Praxis, when a question asks "which of the following provides the strongest evidence for treatment X?" the answer is nearly always the systematic review or the RCT — never the case report.

SLP-specific research designs

• Single-subject experimental designs (SSED). Used heavily in SLP because our caseloads are small and heterogeneous. Includes ABAB (withdrawal), multiple baseline, and alternating treatments designs. Considered strong evidence when properly controlled, despite small N.
• Group treatment studies. RCTs and quasi-experimental designs. Common in aphasia, stuttering, and dysphagia intervention research.
• Diagnostic accuracy studies. Sensitivity, specificity, positive and negative predictive value — used to evaluate screening tests.

Reliability vs. validity — the differential

• Reliability = consistency. If you gave the test twice, would you get the same result? Types: test–retest, inter-rater, internal consistency.
• Validity = accuracy. Does it measure what it claims to measure? Types: content, construct, criterion (concurrent and predictive).

A test can be reliable without being valid (consistently wrong) but cannot be valid without being reliable. On the Praxis, when a stem asks about "consistency across raters," the answer is inter-rater reliability. When it asks whether a test predicts a real-world outcome, the answer is predictive validity.

Sensitivity and specificity — high yield on screening items

• Sensitivity. Of people who have the disorder, what percentage does the test correctly identify? High sensitivity = few false negatives. Priority for screening.
• Specificity. Of people who don't have the disorder, what percentage does the test correctly rule out? High specificity = few false positives. Priority for diagnosis.

Praxis rule of thumb: screening tools should be sensitive, diagnostic tools should be specific.

Effect size, p-value, and confidence intervals

• p-value. Probability the observed result happened by chance. p < .05 is the conventional threshold for statistical significance.
• Effect size (Cohen's d). How big is the difference? 0.2 = small, 0.5 = medium, 0.8 = large. Clinically more meaningful than a p-value alone.
• Confidence interval (CI). The range in which the true value likely falls. A CI that crosses zero (or 1 for ratios) means no statistically significant difference.

The Praxis rarely asks you to calculate. It asks you to interpret. If a study reports p = .04 but a Cohen's d of 0.15, the correct read is "statistically significant but clinically small."

The EBP triad — the mandatory answer

Any question that asks "how should the SLP decide which treatment to use?" is testing the EBP triad. The correct answer will include all three components:

1. Best available external evidence
2. Clinical expertise
3. Client values, preferences, and cultural context

Answers that pick only research, only clinical intuition, or only client preference are wrong — even when they sound reasonable.

Reading a Praxis research stem — the 3-step approach

1. Identify the study design in the stem (RCT, single-subject, cohort, etc.).
2. Note the outcome being measured (diagnostic accuracy, treatment effect, prevalence).
3. Match the design to the appropriate strength claim. Don't overreach beyond what the design supports.

Common trap answers

• "Because a p-value was significant, the treatment is clinically meaningful." — Wrong. Effect size matters.
• "A case study proves the treatment works." — Wrong. Case studies suggest, not prove.
• "The SLP should use the treatment with the strongest evidence, regardless of client preference." — Wrong. EBP is a triad.
• "Higher reliability means the test is more accurate." — Wrong. Reliability is consistency, not accuracy.

Where research shows up across content areas

You'll see research questions embedded in items about assessment selection (which tool has better psychometrics?), treatment planning (which intervention has the strongest evidence?), and ethics (documenting evidence in your rationale). It rarely gets its own obvious section — that's what makes it easy to under-prepare for.