Exploring Research in Music Education & Music Therapy

If you're a music teacher, therapist, or grad student who freezes at the words 'research methods,' this is your field guide. Phillips translates statistics, sampling, and APA style into plain steps, and grounds every design in real music studies. You finish knowing how to read a journal article critically—and how to plan your own study without faking your way through the methods section.

1. 1Part 1, Chapter 1: The Nature of Research

   Research is just disciplined decision-making

   Phillips opens the book by saying: stop arguing, start gathering. Research, he defines, is the systematic and objective process of gathering, recording, and analyzing data for the purpose of making decisions—answering questions or solving problems. He then splits the enterprise in two: scientific, quantitative research that seeks to explain causes and effects, and naturalistic, qualitative research that seeks to understand the meaning of human experience. The whole process unfolds as a spiral with five phases—problem identification, literature review, methodology, data analysis, and interpretation and dissemination.

   The idea

   The definition sounds bureaucratic until you notice what it rules out: intuition, tradition, and authority all fail the test of systematic, objective data. What surprises newcomers is the spiral shape—research isn't a straight line from question to answer, because interpretation usually loops you back to a sharper problem. And the quantitative split isn't a turf war; it's two different questions, one asking what causes what, the other asking what the experience means to the people living it.

   Try this: Take one classroom or clinical hunch you hold and write it as a question that data could actually answer this semester.

2. 2Part 1, Chapter 2: The Research Problem

   A problem you can't investigate isn't a problem

   Phillips insists a research problem must be researchable—investigable through the collection and analysis of data—and significant to the field. The fix is a well-written purpose statement that names the variables of interest, the population, and the setting. His own example: 'The purpose of this study is to investigate the effect of music listening on anxiety levels of preoperative patients in a hospital setting.' Suddenly the vague topic has a lever to pull and a place to pull it.

   The idea

   The purpose statement is a contract with yourself. Naming the variables, population, and setting forces you to decide, before you collect a single data point, what you'll measure, on whom, and where. What surprises people is how much work the word 'researchable' does—it kills questions that are matters of taste or faith rather than data. Significance is the second filter: a study nobody in the field needs is effort spent answering a question no one asked.

   Try this: Rewrite your topic as a purpose statement with variables, population, and setting—then check whether each one is something you can actually observe.

3. 3Part 1, Chapter 3: Review of the Literature

   Eighty percent of your review should come from the source

   Primary sources—original research reports—should make up at least 80% of the literature review; leaning on secondary sources like summaries and reviews is a trap, because every retelling smooths away the details you need. The literature also includes action research, conducted by practitioners in their own settings to improve practice or solve a specific problem.

   The idea

   The 80% rule is really a warning about telephone-game decay. Secondary sources tell you what a study concluded; primary sources tell you how it was done, on whom, and with what measures—the details that decide whether it applies to your classroom. What surprises people is that reading the literature isn't a hoop before the research; it's where you discover that practitioners can be researchers, and that the problem in front of you may be the most significant one available.

   Try this: This week, when you review any document or claim, spend your time checking it directly against the original source rather than relying on summaries or secondhand accounts, aiming for roughly four-fifths of your review time on the source itself.

4. 4Part 1, Chapter 4: The Research Plan

   Four designs, one honest question: how much control do you have?

   Phillips lays out the main quantitative designs: descriptive, correlational, causal-comparative, and experimental, alongside historical research. The choice hinges on a single question—can you manipulate anything? Descriptive work simply portrays what exists; correlational work measures how variables move together; causal-comparative work compares groups that already differ; experimental work assigns and intervenes. Whatever the design, the sample has to be chosen deliberately, and Phillips walks through the options: random sampling, stratified sampling, and purposive sampling.

   The idea

   Design names aren't jargon; they're honesty labels. Calling a study experimental when you couldn't randomly assign overstates what it can prove, while calling a descriptive survey 'causal' misleads readers before the first table. What surprises people is that the design decision comes before data collection, not after—retrofitting a design claim onto data you already gathered is how findings get discredited. Sampling choices carry the same weight: who gets in determines who your conclusions cover.

   Try this: For your next project question, write down which of the four designs fits and what you'd have to control to claim it.

5. 5Part 1, Chapter 5: Experimental Research

   Random assignment and the Solomon four-group gold standard

   In a true experimental design, subjects must be randomly assigned to treatment and control groups, and the researcher must manipulate the independent variable—no self-selected groups, no 'we tried it with the advanced choir.' When random assignment isn't possible, the design slides into quasi-experimental territory, and claims must soften accordingly. Then comes the rigorous endgame: the Solomon four-group design, described as the most rigorous experimental design because it controls for the effects of pretesting. Some groups get the pretest, some don't, and the design untangles whether the test itself changed the outcome.

   The idea

   Two requirements do the heavy lifting: random assignment balances the groups before treatment, and manipulation of the independent variable proves the researcher caused the change. What surprises people is the pretest problem—measuring people can change them, so a study that pretests everyone may be measuring the test's effect, not the treatment's. The Solomon four-group design spends extra groups to buy that certainty, which is why it sits at the top of the rigor hierarchy rather than the middle.

   Try this: Before your next intervention, list how you'll assign participants and whether your pretest itself could change their behavior.

6. 6Part 1, Chapter 6: Descriptive Research

   The survey: music research's workhorse

   The survey is the most common type of descriptive research in music education and music therapy. Data comes from a sample of a population through questionnaires or interviews, often built on Likert-type scales that turn opinions into comparable numbers. Alongside the survey sits its quieter cousin: observational research, the systematic viewing of people's actions with recording, analysis, and interpretation of behavior. One method asks people what they do; the other watches what they actually do—and the gap between the two is where the interesting findings live.

   The idea

   Descriptive research gets dismissed as 'just surveys,' but it's the foundation everything else stands on: you can't explain a cause until you've accurately described a condition. What surprises people is how much craft a Likert-type item takes—wording, scale points, and ordering all shape the answers. And observation isn't hanging around with a clipboard; systematic viewing with recording and analysis is what separates an impression from a finding.

   Try this: Draft one survey question about your program, then spend ten minutes observing the same behavior and compare what people say to what they do.

7. 7Part 1, Chapter 7: Historical Research

   History is a method, not a mood

   Phillips insists this is research, not nostalgia: historical research is the systematic search for facts relating to the past with the goal of interpreting and explaining those facts. The box splits cleanly in two. Primary sources—original documents or artifacts from the period under study—carry the most weight. Secondary sources—accounts written after the fact by someone not present at the event—get handled with care, because memory and agenda bend them.

   The idea

   The word 'systematic' is doing real work: historical research follows the same discipline as any other design—questions, sources, analysis, interpretation—just pointed at the past. What surprises people is that a document's age matters less than its proximity to the event; a later account written by an eyewitness can be primary, while a contemporary summary by someone who wasn't there is secondary. Interpretation is the goal, not decoration—facts collected without explanation are an archive, not a study.

   Try this: For one claim about your program's history, find a primary source from the period and one later account, and compare what each gets right.

8. 8Part 1, Chapter 8: Qualitative Research

   When you are the instrument

   In qualitative work, the researcher is the primary instrument for data collection—observing, interviewing, analyzing text and images, all in natural settings. Phillips maps four major traditions so the immersion has a shape: case study for the in-depth examination of a single individual, group, or setting; ethnography for prolonged engagement in a cultural setting; phenomenology for lived experience; grounded theory for generating theory from the data itself. Gathering continues until data saturation—no new information emerges from additional data gathering. To keep one observer's lens honest, findings get triangulated: multiple sources or methods, observations, interviews, and video analysis, verifying each other.

   The idea

   Being the instrument sounds like a weakness—no standardized test would brag about that—until you see the upside: a human instrument can adjust, probe, and notice what a questionnaire never thought to ask. The cost is rigor, which is why saturation and triangulation exist. Saturation tells you when to stop; triangulation checks whether your single lens distorts. What surprises people is that grounded theory doesn't test a theory—it builds one out of the data, the reverse of the quantitative habit.

   Try this: Observe one rehearsal or session twice this week and write down what a second source—an interview or recording—confirms or contradicts.

9. 9Part 1, Chapter 9: Philosophical Inquiry

   Arguing about beauty without a single data point

   Philosophical inquiry in music education addresses questions of aesthetics, values, and the nature of musical experience. There's no instrument to administer and no group to randomize. The method is logical argument—claims built on reasons, tested against counterexamples, refined through criticism.

   The idea

   Empirical research can tell you whether music lowers anxiety; it cannot tell you whether it should be required, or why a chord moves you. Those are philosophical questions, and Phillips treats them as a legitimate research tradition with its own standards—rigor of argument in place of rigor of measurement. What surprises people is that this isn't a lesser method for people who avoid statistics; it's the only method suited to questions about values and the nature of experience, and every empirical study quietly depends on some philosophical position anyway.

   Try this: Write one paragraph defending why music belongs in your curriculum using reasons, not studies—then find the weakest link in your own argument.

10. 10Part 1, Chapter 10: Quantitative Data Analysis

    p < .05 and the two ways to be wrong

    Phillips starts with the ground floor—descriptive statistics, the measures of central tendency (mean, median, mode) and the measures of variability (standard deviation, range, variance) that describe what the data look like. Then the inferential floor: t-tests, ANOVA for comparing the means of two or more groups, and chi-square. The t-test determines whether the difference between two group means is significant, judged against the conventional threshold of p < .05—less than a 5% probability the results are due to chance. And every decision carries two ways to be wrong: a Type I error rejects a true null hypothesis; a Type II error fails to reject a false one.

    The idea

    The p < .05 convention is a social agreement, not a law of nature—it's the field's shared line for how unlikely a result must be before chance stops being the best explanation. What surprises people is that the two errors pull in opposite directions: demand stronger evidence and you commit fewer Type I errors but more Type II ones, missing real effects. Describing before inferring matters too; a mean without a standard deviation hides whether your data cluster tightly or scatter wildly.

    Try this: Take any two group averages you care about and ask what a t-test would need: how big is the gap, and how spread out are the scores?

11. 11Part 1, Chapter 11: Qualitative Data Analysis

    Open, axial, selective: coding your way to a core category

    The work starts with open coding—initial identification of concepts, naming what's happening line by line. Then axial coding makes connections between categories, showing which concepts relate and how. Finally, selective coding identifies a core category, the spine that organizes everything else. It's slow, iterative, and nothing like running a t-test—but the discipline is just as strict.

    The idea

    Coding is qualitative research's answer to statistics: a systematic procedure that turns unstructured text into analyzable structure. The three stages move from the ground up—concepts first, connections second, a central organizing category last—which keeps the theory anchored in the data instead of imported from it. What surprises people is how physically demanding it is; you read every transcript repeatedly, and the analysis happens in the margins, not in software. The core category is the payoff: one idea that makes the rest cohere.

    Try this: Take notes from one interview or observation and label every few lines with a concept name; then see which labels connect.

12. 12Part 1, Chapter 12: Writing the Research Report

    120 words to sell your study

    Phillips gives the shape first—the standard structure of a research report runs introduction, review of literature, method, results, and discussion, with an abstract and references. Style comes from the Publication Manual of the American Psychological Association, the common style guide for writing research reports in music education and music therapy. And the abstract gets a hard limit: a concise summary of no more than 120 words.

    The idea

    The report structure isn't bureaucracy; it's a promise about where readers will find each kind of claim—questions up front, methods in the middle, evidence in results, meaning in the discussion. APA style matters for the same reason: shared conventions let readers find things and compare studies without friction. What surprises people is the abstract's discipline—a 120-word ceiling forces you to decide what your study actually is, which is often the moment you finally understand it yourself.

    Try this: Write a 120-word abstract for a study you'd love to run—if you can't, the study isn't clear in your head yet.

13. 13Part 2, Chapter 13: Research in Music Education

    The rehearsal decoded: task, response, feedback

    Cornelia Yarbrough and Harry Price studied sequential patterns of instruction in music rehearsals and found the answer in the pattern itself: effective teachers used a sequence of task presentation, student response, and specific positive or corrective feedback. The finding gave every rehearsal a repeatable unit to examine. On the measurement side, Edwin Gordon's Music Aptitude Profile was designed to measure tonal and rhythm imagery in students from grades 4 through 12, giving researchers a tool to track musical potential across a decade of schooling.

    The idea

    The Yarbrough and Price pattern matters because it turns 'good teaching' from a personality trait into an observable sequence you can count and train. Task, response, feedback—each element is specific, and the feedback's specificity is what makes the loop work. What surprises people is that the study came from watching rehearsals systematically, not from theory; the pattern was sitting in every effective rehearsal, waiting for someone to code it. Pair it with Gordon's aptitude work and you get the field's two halves: measuring potential and improving instruction.

    Try this: Record one rehearsal this week and tally your cycles of task presentation, student response, and specific feedback—count the gaps.

14. 14Part 2, Chapter 14: Research in Music Therapy

    Music in the incubator and the metronome in the gait

    Jayne Standley studied the effects of music on physiological responses of premature infants, finding that music stimulation can increase oxygen saturation levels. Michael Thaut pursued rhythmic auditory stimulation—RAS—for gait training in patients with Parkinson's disease. Both studies appear in the pages of journals like the Journal of Music Therapy, the field's leading research outlet.

    The idea

    These two studies show music therapy's distinctive burden: the outcome has to be physiological or functional, not just pleasant. Oxygen saturation and stride symmetry are numbers a skeptic can check, which is exactly why they persuade hospital administrators and neurologists. What surprises people is the mechanism's simplicity—Thaut's patients improve by entraining to a beat, using a rhythmic system the brain already runs. The research doesn't argue that music is magical; it demonstrates, measure by measure, what music can move.

    Try this: Try walking to a steady metronome beat for two minutes and notice how your stride settles—then imagine what ten weeks of that trains.

15. 15Part 6, Chapters 15 and 17

    The proposal is the study, written in advance

    The research proposal contains the problem statement as its introduction, the review of literature, and the proposed methodology, with references behind them. Phillips is blunt about its dual job: the proposal serves as a blueprint for the study and a means of obtaining approval or funding, guiding the research project from first question to final analysis.

    The idea

    A proposal forces every hard decision—variables, population, design, analysis—before the pressure of data collection makes shortcuts tempting. What surprises people is that the blueprint function and the approval function reinforce each other: a proposal clear enough to fund is clear enough to follow when the study hits its inevitable snag. Writing the methods section in advance is also a honesty device, committing you to a plan before results exist to be cherry-picked.

    Try this: Draft the methods section of your next project before you collect anything, and treat it as a contract you can't quietly amend.

16. 16Chapter 19: Evaluating Research

    Interrogate every report you read

    Phillips hands readers a checklist. Criteria for evaluating a research report include clarity of the problem, appropriateness of the design, adequacy of the sample, and accuracy of the data analysis. Then come the evaluative questions: Were the measuring instruments valid and reliable? Is the sample representative? Were the conclusions supported by the data? The book's final move points forward: more mixed-methods research, combining quantitative and qualitative approaches within a single study, to give a more complete understanding of complex phenomena in music education and music therapy.

    The idea

    Evaluation questions turn reading from consumption into cross-examination—every published claim must survive questions about its problem, design, sample, instruments, and analysis. What surprises people is how often a study fails not at the statistics but at the seams: a valid measure on an unrepresentative sample, or conclusions a step ahead of the data. The mixed-methods closing follows naturally, since complex questions about humans making music rarely yield to one method alone.

    Try this: Pick one study you cite often and answer Phillips's evaluation questions for it in writing—note where your confidence wobbles.

Music education and music therapy graduate students facing their first research methods course, plus practitioners considering action research in their own classrooms or clinics. Also ideal for anyone who reads JRME or JMT articles and wants to judge them critically.

FAQ