Uni Reading List
Study guides /

Choose an education research guide and plan evidence you can defend

A manageable educational study starts with a question and a credible observation, not a favorite method. A questionnaire, an interview and a classroom improvement cycle serve different purposes. Choose one main guide, make your intended inference explicit, and learn to report what the evidence leaves uncertain as carefully as what it shows.

Course topic: Educational Research Methods

Study route and exercise: University Reading List editorial project. How we curate · Sources checked:

Download blank worksheet (Markdown) ↓

Scope and responsibility

This route supports a first small project or preparation for a research-methods course. Ordinary percentages and careful reading are sufficient for the exercises; some quantitative designs require further statistics. It compares Action Research, the eighth edition of Bell, Waters and Johnson, and the seventh edition of Creswell and Guetterman. All tasks use invented material and require no participant recruitment or classroom intervention.

Choose your starting point

I need free guidance for a practitioner question

Choose Action Research when the aim is to examine and improve your own educational practice through a documented cycle. Its free full text supports planning; access to a classroom is neither needed for this paper exercise nor permission to conduct a study.

I need to organize my first bounded project

Choose Bell, Waters and Johnson for a practical sequence from question and ethics through evidence and report. Use the eighth edition’s contents to locate the method you actually need rather than collecting every type of data.

I must compare quantitative, qualitative and mixed designs

Choose Creswell and Guetterman when you need to explain why a question calls for a particular design. Its broader design taxonomy earns its space after you can name a problem, purpose and unit of observation.

Compare the three books

Action Research

J. Spencer Clark; Suzanne Porath; Julie Thiele; Morgan Jobe · 2020

Free practitioner-inquiry spine

Choose it when
You want a cycle connecting a local concern, literature, planning, data, reflection and sharing.
Use another route when
You need a comprehensive survey of research designs or a basis for claiming a universally effective intervention.
Selected reading
Use 1–2 for practitioner inquiry, 3–4 for questions and preparation, 5–6 for evidence, 7 for reporting; 8 is a practitioner perspective.
Limits and trade-offs
The authorized web book and downloads are free. Its 2020 ebook ISBN is 9781944548292. Local improvement and causal impact are different aims; one cycle does not by itself establish general effectiveness.

Try before choosing

Use Action Research chapters 1–4 to turn a fictional difficulty finding a deadline into a question, a possible action and an observation/reflection cycle. Write the design only; do not recruit learners.

Check your work: Question, action and evidence are distinguishable. The design includes a way to notice failure and an ethics check, rather than assuming practitioner involvement proves effectiveness.

Official free reading ↗

Book record and access options →

Doing Your Research Project: A Guide for First-Time Researchers

Judith Bell; Stephen Waters; Helen Johnson · Edition: 8 · 2024

Practical handbook for a first project

Choose it when
You need a work plan and accessible support for literature, documents, questionnaires, interviews or observation.
Use another route when
You need detailed statistical theory or a specialist qualitative methodology beyond a first project.
Selected reading
Read 1–3, then 4–6 for literature. Select 7 or 9–12 according to evidence needs; finish with 13–14 for interpretation and reporting.
Limits and trade-offs
The selected 2024 eighth-edition print ISBN is 9780335252428; the paid eBook has a different ISBN and uses VitalSource. A free reading app is not a free textbook.

Try before choosing

Use the literature-review route to create a two-source ledger for a feasible project: exact source, supported claim and how that claim changes the design. Use sources you can actually locate.

Check your work: Every quoted or paraphrased claim has a retrievable source and location. An attractive topic list or a fabricated author/year citation is not a literature review.

Book record and access options →

Educational Research: Planning, Conducting, and Evaluating Quantitative and Qualitative Research

John W. Creswell; Timothy C. Guetterman · Edition: 7 · 2024

Integrated process and design reference

Choose it when
Your course requires a reasoned choice among quantitative, qualitative and mixed methods, with explicit links between questions and designs.
Use another route when
You only need a short project timetable and would be overwhelmed by a larger design reference.
Selected reading
Use 2–4 for problem, literature and purpose; 5–8 for paired collection/analysis; 12 for surveys, 16 for mixed methods and 17 for action research.
Limits and trade-offs
Pearson dates publication to June 2024 and copyright to 2025. Print ISBN 9780138161194 differs from Pearson+ access ISBN 9780138173784. This route preserves the selected edition and does not assume digital entitlement.

Try before choosing

Use the seventh-edition survey route to draw invited/responded/yes/no/missing boxes for the worked example before reading further about analysis. Explain each denominator in words.

Check your work: The yes/no/missing leaf boxes total 80; 40/80 and 28/40 answer different questions. If this explanation is difficult, return to measurement and basic proportions before selecting an analysis technique.

Book record and access options →

Check your preparation

Is “improve communication” a research question?

It is an aim. Specify who, what communication task, which observable indicator and what period. Decide whether you want a description, an account of experience or an impact claim.

Does 70% of respondents mean 70% of everyone invited?

No. That requires assumptions about nonrespondents or a defensible adjustment. Always state the denominator and the number missing.

Does adding interviews automatically create a strong mixed-methods study?

No. Explain how the two kinds of evidence answer linked questions and how they will be integrated. A quotation cannot repair an unknown population rate.

Six steps with a concrete output

These are your own study marks, not an assessment. They are not sent to this site.

  1. 1. Turn a concern into an answerable question

    Selected reading: Action Research 1–3; Bell 1–2; Creswell 2 and 4.

    Make this: For an invented college information page, replace “is it clear?” with a question naming an audience, a task, an indicator and a time window. State whether the aim is diagnosis or causal evaluation.

    Check your work: The indicator can be observed and could disappoint your expectation. It does not silently equate finding a deadline with understanding the whole course.

  2. 2. Build a small literature-and-decisions record

    Selected reading: Action Research 3; Bell 4–6; Creswell 3.

    Make this: Create a five-row planning table with a claim to investigate, search terms, inclusion rule, evidence needed and a reason the evidence might not transfer. Use placeholders rather than inventing published findings.

    Check your work: A reader can reproduce your intended search and see how the question could change. A literature review is not a list of favorable quotations.

  3. 3. Define the measure and the missing cases

    Selected reading: Action Research 4–5; Bell 9; Creswell 5 and 12.

    Make this: Draft one neutral yes/no item about finding a deadline without help. Define “help,” invitation eligibility and nonresponse. Complete the fictional calculation below before proposing any success claim.

    Check your work: You keep invitation, response and reported success counts separate. The item measures a self-report of one task, not achievement or overall satisfaction.

  4. 4. Add a different kind of evidence for a reason

    Selected reading: Action Research 5–6; Bell 7 and 10–12; Creswell 7–8.

    Make this: Write a documentary-check protocol for the invented page and three interview prompts that could explore why the deadline is hard to locate. Explain how each source would complement the item, using no real participants.

    Check your work: Prompts do not presume failure or blame. A document check concerns page content; an interview concerns experience; neither supplies missing survey outcomes automatically.

  5. 5. Design a proportionate inquiry cycle

    Selected reading: Action Research 2 and 4–6; Creswell 16–17.

    Make this: Draw a paper-only cycle: question, baseline evidence plan, proposed page revision, comparable follow-up measure, reflection. Add one risk to comparison and one decision rule for revising the plan.

    Check your work: The plan distinguishes improvement from proof of causation. Explain how changed invitations, seasonal demands or item wording could alter a comparison.

  6. 6. Write the results you are entitled to report

    Selected reading: Action Research 7; Bell 13–14; Creswell 9.

    Make this: Write a short mock abstract using only the invented counts, the measure definition, missing-response bounds and next evidence needed. Add a limitations paragraph and a separate local-review checklist for any future real study.

    Check your work: No fictional finding is presented as an actual classroom result. Unknown responses remain unknown, and proposed collection is not described as already approved or performed.

Worked exercise

A clear response rate does not identify the invited group’s success rate

For a wholly fictional planning exercise, 80 eligible learners receive a yes/no item: “On this page, did you find the assignment deadline without help?” Forty respond; 28 say yes and 12 say no. A draft report claims at least 60% of all 80 succeeded. What can these counts establish?

Assumptions

Each person can respond once, all 80 invitations are valid, the 40 recorded responses are complete, and no information is available about the other 40. Treat the item as a self-report of one task. Do not assume that missing responses are random or that an answer measures learning.

  1. The response rate is 40/80 = 50%. Reported success among respondents is 28/40 = 70%. These percentages have different denominators and answer different questions.
  2. If all 40 missing answers would be no, the invited-group reported-success proportion is 28/80 = 35%. If all would be yes, it is (28 + 40)/80 = 85%. Without an assumption or more evidence, both extremes remain compatible with the counts.
  3. At least 60% of 80 means at least 48 yes answers. The observed 28 need at least 20 additional yes answers among the 40 missing. The data do not tell us whether that condition holds.

Answer and interpretation

Report “28 of 40 respondents said yes; 40 of 80 invitees responded.” The overall proportion lies between 35% and 85% under the stated assumptions, so the at-least-60% claim is not established. This is a missing-response bound, not a confidence interval or a measured improvement.

Test the result

The 28 yes plus 12 no plus 40 missing sum to 80. The threshold requires 48 − 28 = 20 missing yes answers; 19 would yield 47/80 = 58.75%, while 20 gives 60%. A neutral follow-up might reduce missingness, but it would not by itself validate the wording or prove that a page revision caused improvement.

Change an assumption

More responses can still leave the threshold unproved

Keep the original 80 invitees and the same original-task question. Of the 40 missing people, a fictional follow-up obtains 20 new answers: 18 yes and 2 no. Each person is counted once, answers refer to the original task rather than a later attempt, and the remaining 20 are still unknown. Reassess the claim that at least 60% of all invitees reported success.

  1. Combine the records: yes=28+18=46, no=12+2=14, missing=20; total=80. The response rate is 60/80=75%. Yes among respondents is 46/60≈76.7%, a different denominator.
  2. The invited-group bounds narrow to 46/80=57.5% and (46+20)/80=82.5%. Better response coverage has narrowed the uncertainty, but the lower bound is still below 60%.
  3. The threshold remains 48 yes answers. At least two of the remaining 20 would need to say yes. Both zero and two additional yes answers remain compatible with the supplied records.

Answer and interpretation

Report 46 yes out of 60 respondents, with 60 of 80 responding. The 60% invited-group claim is still not established; its nonresponse bounds are now 57.5–82.5%. The increase from 70% to about 76.7% among respondents does not by itself show an improvement in the page or learning.

Test the result

Check 46+14+20=80 and 48−46=2. If only one missing person would say yes, 47/80=58.75%, still below the threshold. No random-missingness assumption or confidence interval has been introduced.

Common mistakes

Letting the tool choose the question

Choosing a survey because a form is convenient can produce data that do not address your concern. Start with the intended claim, then choose a measure and design.

Treating silence as a negative or a positive

Nonresponse is missing information unless a defensible definition says otherwise. Show bounds or justify assumptions rather than quietly changing the denominator.

Equating a local cycle with a universal effect

An action-research cycle can inform local reflection. Generalization, causal attribution and ethical authorization each require their own argument and process.

Before moving on

  • State a question whose population, observation, period and intended inference are explicit.
  • Reproduce all denominators, bounds and the 60% threshold, and explain why no confidence interval was calculated.
  • Submit a paper-only protocol that connects sources of evidence to decisions, keeps limitations visible and makes no claim to real participant results.

Sources and edition checks

Sources checked:

These are proposed study activities, not official course assignments or a claim that every book has been read cover to cover.

How we curate →