The Research Paper: Sources, Synthesis, and Citation

A research paper answers a question with evidence and a claim of your own. The work is to decide what the evidence supports, explain why, and let readers check your reasoning.

Know what you are making

Original work at this level means reaching your own reasoned answer. You do not have to discover a fact no one has ever known. Imagine joining a discussion about school phone rules. Your sources have already offered findings and arguments; your paper explains which findings matter and how they shape your answer.

Make the assignment into a plan

Before searching, record the deadline, length, topic limits, required source types, citation style, format, and checkpoints. Check the rules on feedback and AI use. These choices affect the work: an interview takes time to arrange, and a required scholarly article may need library access.

Research moves through background reading, a focused question, searching, source evaluation, notes, synthesis, a thesis, an outline, drafting, revision, editing, citation checks, and submission. The stages loop. A weak paragraph can reveal a missing source. A strong source can show that your question needs to change. Keep a date for starting the draft so that searching does not consume the whole project.

Work backward from submission

Suppose a paper is due Friday, May 14, and you begin Monday, March 8, in a calendar year such as 2027. You have 67 days between those dates. Reserve time for the finished draft to receive feedback, then place earlier tasks behind it:

• March 12: question approved.
• March 19: working source list ready.
• April 2: main reading and notes complete.
• April 9: outline ready.
• April 23: full draft complete.
• May 7: revised draft ready after feedback.
• May 11: citations checked and proofreading complete.
• May 14: final submission.

The April 23 draft deadline matters. It leaves time to change the argument, not just correct typing. A short project uses the same stages in a few days; a sustained project needs more sources, fuller analysis, and several drafts.

Keep a record that helps tomorrow's work

A research log records the date, searches, reading, decisions, and next task. An entry might say, “March 10: searched school phone restrictions and attainment; found two studies with different designs; next, compare their measures.” This is more useful than “Did research.” It saves repeated searches and supplies evidence for a later reflection on how your ideas changed.

Starting with “I will prove bans work” invites you to collect agreement and miss objections. Start with what you need to find out. When reading changes your answer, update the plan and explain the decision in the log.

“Phones in school” names a subject, but it does not tell you what to find out. A useful research question gives your reading a purpose and keeps a long paper within reach.

Move from interest to inquiry

A topic is a subject area. An issue is a debate within it. A research question names what you will investigate. A thesis is the answer you eventually defend. Keep these roles separate: a question opens the inquiry; a thesis grows from the evidence.

Start with background reading in an encyclopedia, textbook, or reliable overview. Notice recurring terms, debates, and researchers. Then limit the topic by the group, place, period, aspect, or effect that matters. You do not need a limit in every category.

For example, “phones in school” becomes “phone bans,” then “all-day phone bans in high schools.” A question can now ask, “To what extent do all-day phone bans in high schools improve students’ exam results and well-being?” It specifies a policy, a group, and outcomes that research can examine. It also allows a mixed answer: a policy could affect exam results and well-being differently.

Test the question before committing

A strong question passes five tests:

1. Focused: the answer fits the available length.
2. Open: it calls for explanation or judgment beyond one fact or a bare yes or no.
3. Arguable: informed readers could reasonably reach different answers.
4. Researchable: you can obtain relevant evidence in the time available.
5. Worthwhile: you can explain why the answer matters.

“What is a phone ban?” needs a definition. “Are phones bad?” leaves “bad” undefined and covers too much. “Should schools ban phones?” becomes more useful when you define success, such as better exam results or less distraction. Changing the opening to “how” helps only when the rest of the question has a clear focus.

Adjust the scope to the evidence

Too much material signals a need to narrow. A 2,500-word paper cannot assess every form of digital technology in every school. Too little evidence calls for broadening. “Did one school's ban raise grade 10 math scores in 2024?” may depend on records you cannot access. Widening the setting to high schools gives you published studies to compare.

Break the main question into two to four sub-questions. For the all-day-ban question, ask what happens to exam results, what happens to well-being, and why findings differ between schools. Each sub-question gives you search terms and a place to organize notes. A sub-question that does not help answer the main question needs to go.

Literary research uses the same focus but asks how a text creates meaning. “What happens in Frankenstein?” invites plot summary. “How does Shelley’s use of several narrators shape sympathy for the creature?” connects a method to an effect. The answer needs close reading and critical discussion, not a list of events.

Searching works best when you decide what evidence your question needs before choosing words for the search box. A library database, a book catalog, and a general search engine find different kinds of material.

Know what a source can offer

These roles depend on the question. A newspaper report from 1918 interprets an epidemic, but it is original evidence for a question about how newspapers described that epidemic. A scholarly article can report original research and discuss earlier studies.

Books offer depth; government and research reports provide data; news gives recent events and public views. Scholarly articles usually address researchers, document sources, and undergo peer review. Combine the publication types that your question needs.

Choose a route into the material

Use a library catalog for books. School and public libraries may provide article databases such as JSTOR, ProQuest, EBSCO, and Gale. Google Scholar searches scholarly publications across the web. General search engines help locate organizations, reports, and news. A librarian can help you select a database and access a full article.

Pull the main concepts out of the question and give each a group of search terms. For phone policy, try “cellphone,” “mobile phone,” and “smartphone,” along with “ban,” “restriction,” and “policy.” For the learners, try “high school,” “secondary school,” and “adolescents.” For results, try “test scores,” “achievement,” and “attainment.” Synonyms and international terms keep one choice of wording from hiding useful research.

Combine concepts and adjust results

In databases, AND requires the joined concepts; OR accepts alternatives; NOT excludes a term. Quotation marks request a phrase, and parentheses group alternatives. Many databases use an asterisk for word endings: restrict* can find “restrict,” “restriction,” and “restrictions.”

A starting query is ("phone ban" OR "mobile phone" OR smartphone) AND (school OR classroom) AND (achievement OR "test scores"). Check the search tool's help: syntax differs. Google commonly treats a space like AND and uses a minus sign to exclude a word. site:gov targets U.S. government sites; filetype:pdf requests PDF files. A file type says nothing about quality.

Too many results? Add an outcome, an exact phrase, or an appropriate date or source-type filter. Too few? Remove a concept or add synonyms. A peer-reviewed filter narrows publication type but does not evaluate each argument for you.

Follow the conversation

Read a useful article's references to find earlier work. Use “Cited by” to locate later work that responds to it. Record productive searches so you can repeat them.

Wikipedia can supply background and references; an AI tool can suggest terms. Find and read the actual sources before citing them, since generated references can be false. Search for evidence that challenges your expected answer. Move toward drafting when each sub-question has solid evidence and new searches mainly return familiar studies and debates.

A source can be accurate yet irrelevant to your question. It can also have a clear interest in persuading you and still reveal something useful about that interest. Evaluate reliability and usefulness separately.

Look beyond the page

The CRAAP test asks about currency, relevance, authority, accuracy, and purpose. Check when the material appeared, whether it answers your question, who created it, what evidence supports it, and why it was published. Recency matters for changing phone habits; an older novel remains primary evidence about its own language.

Do these checks outside the page as well as within it. Lateral reading means opening other sources to investigate the author, organization, and claim. A polished website can make impressive claims about itself. Independent reporting and the original research help you test them.

SIFT gives four moves: Stop; Investigate the source; Find better coverage; Trace claims, quotations, and media to their original context. Suppose a post says a policy transformed learning. Pause before accepting it, identify the publisher, find fuller reporting, and trace the statement to the research it relies on. The original may concern a smaller group or a different outcome.

Match evidence to the claim

Consider three sources for research on school phone bans:

• Goodyear and colleagues' 2025 study included 1,227 students in 30 English secondary schools. It compared schools with different policies at one time and found no evidence of better well-being or attainment under restrictive policies. It is useful for comparing policies and outcomes, but that design cannot establish that a policy caused an outcome.
• Pew Research Center's 2026 survey found that 77% of U.S. adults supported classroom bans. That measures public opinion. It cannot tell you whether bans improve learning.
• Suppose a company selling lockable pouches claims on its blog that “bans transform schools” without supplying data. Its commercial interest matters. The blog can be evidence of marketing; its effectiveness claim needs checking against the original study, if one is identified.

The distinction is between what a source measured and what you want to claim. Public support and educational success are different outcomes.

Weigh quality and independence

Peer review means other experts examined a study before publication. It is a valuable check, not a guarantee of a final answer. Read the methods and limits: who was studied, where, when, and how? A finding about one age group may not transfer to another.

Bias is a perspective or slant; unreliability concerns the trustworthiness of information. An opinion column can be useful evidence of a position. Check its factual claims and name its perspective when you discuss it.

Three news reports repeating one press release provide one underlying piece of evidence. Trace their sources before treating repetition as confirmation. A .org address or professional layout does not establish expertise, and a .edu page may contain student work.

Finish your evaluation with a short, specific judgment: “Useful comparison across schools; cannot establish cause; setting may limit its application elsewhere.” This gives you a reason for using the source and a limit to carry into the paper.

Published sources cannot answer every local question. To find out how people experience a rule, what they do, or how they speak in a setting, a project may need evidence you collect yourself.

Match the method to the question

A survey can show how common a view or reported habit is among respondents. An interview gives room to explain reasons and experiences. Observation records behavior. A collection of texts or transcripts lets you examine language closely. Choose the method that can answer the question, then explain that choice in the report.

For an interview, find someone with relevant knowledge and prepare five to eight open questions. “How has the ban changed your school day?” allows unexpected answers. “Don't you agree the ban is unfair?” steers the response. Take notes as well as making any permitted recording.

Survey questions should be short, neutral, and about one thing. “Is the ban strict and helpful?” cannot capture a student who thinks it is strict but unhelpful. Split the two judgments. Try the survey with a few people before collecting the final answers and fix confusing wording.

Protect the people taking part

Before collecting data, explain the purpose, seek informed consent, and make clear that participants can stop. Obtain the school or parent permissions required for work with minors. Ask permission before recording, keep data secure, and honor agreements about names and anonymity. Avoid questions likely to cause distress. These decisions belong in the research plan before the first interview.

Report the sample honestly

The sample is the group that actually takes part. Explain how people were selected and who responded. Your friends may be easy to contact, but they may share views that are uncommon elsewhere in the school.

Suppose 120 students answer a survey sent to a school of 800. Of the respondents, 78 support a phone ban. The response rate is 120 divided by 800, or 15%. Support among respondents is 78 divided by 120, or 65%. Those percentages answer different questions. Write, “65% of the 120 respondents supported the ban.” The 680 students who did not reply could hold different views.

Keep the records behind the findings

Record dates, settings, questions, and participant labels. Transcribe speech faithfully, including hesitations when they matter to your question. Keep raw data so that you can check your analysis; use anonymized extracts or appendices as appropriate.

Counts and percentages suit survey responses. Repeated themes can organize interview findings. In a 10-minute podcast transcript, 18 of one host's 25 questions are open questions: 72%. You can discuss what those questions do in that conversation. This sample does not establish how all hosts speak.

A report on collected evidence normally explains the question and background, methods, findings and analysis, conclusions, references, and relevant appendices. Link the findings to published research. State limits such as a small sample, self-selection, or a short observation period where those limits affect the conclusion.

Good notes preserve both the source's meaning and your own response. They should make it easy to tell who said what and where you can find it again.

Read for a purpose

Screen a source through its abstract, introduction, headings, and conclusion. Decide which sub-question it can answer, then read the relevant passages closely. Read a source fully when your argument depends heavily on it; a useful sentence can mean something different within the whole argument.

Begin a working bibliography with the first source. Record the author, title, publisher or journal, date, volume and issue, page range, DOI or URL, and web access date. A DOI is a persistent identifier for a publication. Check details imported by a reference manager such as Zotero against the actual source.

Separate three voices in your notes

A note can hold exact words, a restatement, or your own thought. Label the type at the moment you write it:

• Quotation: copy the words accurately, enclose them in quotation marks, and record the page or other locator.
• Paraphrase or summary: label it, record the source and locator, and preserve the claim's strength.
• Your idea: use a tag such as “ME” so you will not later mistake it for a published finding.

Pasting source text into a notes file without quotation marks creates a future problem. Weeks later, fluent wording may look like your own. A citation identifies the source, but it does not tell you which words you copied.

Build a note you can use

For example:

“Source: Goodyear et al., 2025. Sub-question: well-being. Type: paraphrase. Point: 1,227 students in 30 English secondary schools; restrictive policies were not associated with better mental well-being. Locator: article summary. ME: compare this design with studies tracking a policy change.”

The source, note type, topic, locator, and response have different jobs. Before using a more detailed claim from the article, record the page or section where that claim appears. A locator should lead back to the passage you actually read.

Tag notes by sub-question as well as source. When drafting a section on well-being, you then retrieve several authors' evidence together. Keeping each source in a separate folder alone can encourage a paper that reports one author after another.

Turn a list into an annotated bibliography

An annotated bibliography gives each source a short account of its content, quality, and planned use. An entry on Beland and Murphy could explain that the study examines exam performance after phone bans in four English cities, appears in a peer-reviewed journal, and helps the attainment section but not the well-being section. “Useful source” gives no reason for keeping it.

After reading several sources, write a short “what I think now” paragraph from memory. State what seems supported and what you still need to check. This helps you move from collecting other people's sentences to developing an argument. Highlighting alone does not create that record of understanding.

Synthesis explains what sources mean together. It starts when you stop asking only “What does this author say?” and ask “How does this finding change my answer?”

Put sources beside the same question

A summary reports one source's main ideas. Synthesis connects sources through agreement, disagreement, a gap, a change over time, or a difference in method. Your organizing idea leads the paragraph; the authors supply evidence for it.

A synthesis matrix makes these connections visible. Put sources in rows and sub-questions in columns. Record brief notes in the cells, including important limits. Reading across one row follows one source. Reading down a column compares evidence about the same issue.

An empty cell means you have not entered evidence for that issue. It is not a finding that there was no effect. Go back to the source if the missing information matters.

Explain disagreement fairly

These studies both concern school phone policies and attainment, yet they do not reach the same finding. Listing the findings is a start. Synthesis asks what could account for the disagreement.

Compare the outcome measured, participants, setting, date, and design. Beland and Murphy track results around policy adoption; Goodyear and colleagues compare schools at a single time. Those designs ask about change in different ways. The difference may help explain the findings, but it does not establish the reason for the disagreement by itself.

A careful synthesis is, “Evidence on attainment is mixed: an earlier study found gains after bans, particularly among lower achievers, while a later cross-sectional study found no association between restrictive policies and better attainment. Their different designs make direct comparison difficult.” The judgment belongs to the writer and follows from comparing the studies.

Weigh the evidence

Do not settle a disagreement by counting sources. Three brief stories based on one report add no independent test. One well-designed study that directly answers your question can carry more weight than several weaker sources. Explain your reasons: relevant group, suitable measure, stronger method, or a limit that matters.

Keep disagreement visible even when it challenges your working thesis. A paragraph beginning “Studies disagree about attainment” creates room for both findings and your evaluation. A sequence beginning “Source one says,” then “Source two says,” can leave the reader to work out the relationship.

Frames such as “Both studies suggest…,” “While one finds…, the other…,” and “The difference may reflect…” help signal a relationship. Follow the frame with the particular evidence and your judgment; the connecting words cannot supply the reasoning for you.

A thesis gives the paper a direction that the evidence can support. An outline then turns that answer into a sequence of smaller claims.

Make an answer that can be tested

A thesis is the paper's central claim. It answers the research question and gives readers a position to assess. A working thesis is a first version that can change as you read and draft.

For the question about all-day phone bans, a thesis could be, “All-day phone bans may raise exam results for some students, but the evidence does not show that they improve well-being on their own, so schools should treat a ban as one tool within a wider policy.” The claim answers both outcomes and keeps its confidence within the evidence.

Use four checks:

• Does it answer the research question?
• Could an informed reader disagree?
• Does the evidence support it?
• Is it specific enough to guide the paper?

“This paper will discuss phone bans” announces a topic. “Many schools ban phones” states a fact. Neither says what your research leads you to argue. Adding “I believe” does not turn an announcement into a supported thesis.

When evidence challenges the thesis, revise it. For example, a claim that every student benefits must change if the strongest evidence identifies gains mainly among lower-achieving students. Hiding that distinction weakens the argument.

Organize by claims

A section heading such as “Effects on exam results” is a topic. Under it, write the claim the section will defend and list the evidence, analysis, and objection it needs. This prevents the outline from becoming “Study A, Study B, Study C.” The same study can serve different sections when it supplies different relevant findings.

For a 2,500-word paper, a possible budget is:

The total is 2,500. Giving more room to the two outcomes reflects their importance to the question. The budget is a planning tool: a section that needs more reasoning may need words moved from background.

Build the order and the objection

Choose an order that helps readers follow the reasoning: by theme, cause and effect, problem and solution, comparison, or time. Ask why each section belongs after the previous one. An outline should show the path to the thesis, not just the order in which you found sources.

Include the strongest objection to your answer. If supporters argue that a clear ban is easier to enforce, present that reason fairly. You might concede the practical benefit while explaining that ease of enforcement and improved well-being are separate claims. A reply must address the actual objection; choosing a weak version merely makes the paper look one-sided.

As you outline, mark any claim with no evidence. That is a specific search task to complete before, or during, drafting.

A quotation earns its place when the exact words matter. In literary research, a word or image may be the evidence you analyze. In a general research paper, a finding often works better in paraphrase.

Give the quotation a job

Build a short quotation into three moves: introduce it, quote it, and explain it. A signal phrase identifies the speaker and the relation to your point. “Reports” describes a finding; “argues” introduces a position; “admits” suggests reluctance. Choose a verb the passage supports. MLA commonly uses the present tense for discussion of a source; APA commonly uses the past tense for reported research.

For example: The creature links his violence to suffering: “I was benevolent and good; misery made me a fiend” (Shelley, ch. 10). The verb “made” presents his change as the result of experience and directs attention toward his treatment.

The explanation analyzes the wording instead of merely saying that the creature changed. Check the edition you are using and give its page or other reliable locator.

Fit the words into your grammar

A short phrase can fit your own sentence: The creature describes himself as once “benevolent and good” (Shelley, ch. 10). Read the sentence aloud with the quotation in place. Your grammar must work without changing the source silently.

Use square brackets for an inserted clarification or a necessary change: The creature says that “misery made [him] a fiend” (Shelley, ch. 10). The change from “me” to “[him]” shows which wording is yours. Use [sic] after an error only when you need to show that the error belongs to the original.

An ellipsis marks omitted words. In MLA, use three spaced periods: “. . .”. A cut must keep the original meaning. Removing “not” from “The plan is not a solution” reverses the claim; an ellipsis cannot make that honest.

Keep marks, locators, and punctuation in order

Copy quotations exactly. A quotation inside a quotation uses single marks inside double marks in American English. Normally commas and periods sit inside closing quotation marks; when an in-text citation follows a short quotation, the sentence period goes after the closing parenthesis.

A block quotation starts on a new line, has an indent of about 1.27 cm, and has no enclosing quotation marks. Its citation follows the final punctuation. A block takes substantial space; discuss the passage so readers know why its full wording is needed.

Credit the source you actually read

If you meet a quotation inside another writer's work, look for the original. If it remains unavailable, identify the secondary route. Suppose you read Lee's words only in an invented article by Ortiz. MLA uses “qtd. in Ortiz 14”; APA uses “Lee, 2019, as cited in Ortiz, 2024, p. 14.” The final list contains Ortiz, the source you consulted. Do not imply you inspected Lee's original.

Strings of quotations can crowd out your reasoning. After each quotation, explain what its wording contributes to your claim.

Using your own words is only half of accurate paraphrase. You must also preserve whose claim it is, what it means, and how strongly the evidence supports it.

Choose the amount of detail

Paraphrase and summary still borrow ideas. Both need credit even though you wrote the sentences.

Rebuild the meaning

Read until you can explain the passage. Cover it, write from understanding, and compare your version with the original. Check the factual meaning, the strength of the claim, and any distinctive phrasing. Add a citation and a signal phrase where needed to show the borrowing's boundaries.

Suppose the invented researcher Ortiz writes on page 12, “Teachers in the study reported that students talked to one another more at lunch after the ban, although many students said they felt anxious when they could not check messages.”

A paraphrase can reorganize the idea around mixed social effects: “Ortiz describes a divided picture of the ban's social effects: teachers described more lunchtime conversation, while students reported worry about missing messages (12).” This keeps both perspectives and their reporting roles.

Changing “teachers” to “educators” and “anxious” to “nervous” while keeping the sentence's pattern is patchwriting. A citation does not turn a near-copy into independent wording. Close the passage and explain the relationship between its ideas before trying again.

Keep the claim's strength

Small changes can create large errors:

• “May improve” leaves an outcome uncertain; “will improve” predicts it firmly.
• “Students reported feeling worried” records their accounts; “the policy caused anxiety” claims a cause.
• “In these schools” limits the setting; “in all schools” extends the finding.

Keep the group, conditions, qualifications, and attribution needed for a fair account. You can retain names, technical terms, and ordinary words such as “teachers.” Awkward synonyms are not evidence of originality. Put a distinctive borrowed phrase in quotation marks if you keep it.

Show where the source stops

Start a long paraphrase with “According to Ortiz” or another clear signal. Across several sentences, renew the attribution where a reader might lose track. Then mark the move to your analysis: “This contrast matters because…” Readers should be able to distinguish the source's point from your conclusion.

A summary of a whole study keeps the relevant group and result while dropping secondary details: “Goodyear and colleagues found no evidence of better mental well-being under restrictive phone policies in their comparison of English secondary schools.”

Academic integrity means giving credit, reporting honestly, and submitting work that represents your own learning. It applies to ideas and data as well as words.

Recognize the different problems

Plagiarism presents someone else's words, ideas, images, or other work as your own. It can involve copying without credit, quotation without quotation marks, patchwriting, missing citations, or submitting another person's writing. Reusing your earlier work without permission is often called self-plagiarism. Fabrication invents evidence: a nonexistent source, quotation, interview, or data set.

Careless notes can create plagiarism without a deliberate plan to deceive. Mark quotations, record locators, and save source details while you read. Those habits are more reliable than reconstructing credit after the draft is finished.

Decide what needs credit

Cite borrowed quotations, paraphrases, summaries, statistics, specialized facts, images, charts, and ideas from conversations. Common knowledge depends on the audience: “Mary Shelley wrote Frankenstein” is widely established information; a particular study's finding needs its source.

Your own ideas and collected data do not need an external citation, but identify them clearly. Explain how you gathered your data so readers can judge it. If you adapt someone else's chart, credit the chart's source as well as explaining your changes.

For a paper on phone rules, a specific 2026 Pew figure about support needs citation. Your proposal that schools should monitor outcomes by student group is your judgment; the studies used to justify that proposal still need credit.

Use help and AI within the rules

Check the assignment's rules before using generative AI or outside help. Where a tool is prohibited, do not use it. Where permitted, disclose how it contributed. Citation does not make a prohibited use acceptable, and an AI-generated passage cannot honestly be presented as writing you produced yourself.

When you discuss AI output as a source, MLA begins with a description of the generated content or prompt, then the tool, model, company, date, and stable conversation link where available. APA's retrievable-chat form gives the company, full date, italic chat title, “[Generative AI chat],” tool or model, and chat URL. These are source references, not authorship credits for your paper. For factual evidence, find and read the original publications that a tool suggests.

Feedback can point to a weak paragraph or ask for clearer reasoning. It becomes a different kind of help when someone rewrites the assessed work. Follow the permitted boundaries and record help that must be acknowledged.

Keep evidence of your process

Save notes, outlines, research logs, and dated drafts. They help you track changes and explain how the paper developed.

A similarity score measures matching text, not honesty. Correct quotations and references can match; missing matches can leave dishonest work undetected. Inspect the passages and their attribution. A high or low score alone does not settle whether work has been plagiarized.

Citations connect a claim in your paper to the evidence behind it. They give credit, let readers check the evidence, and make the basis of your argument visible.

Keep the two parts connected

Most systems pair a short citation in the text with a full entry at the end. The short form identifies the right entry; the full form supplies enough detail to locate the work. A page, chapter, section, or paragraph directs the reader to the particular passage.

A reference list records cited sources. A bibliography may include wider reading. Follow the assignment's requested list and the style's exceptions: APA personal communications appear in the text only because readers cannot retrieve them.

Collect details before formatting

For each source, find who created it, its title, where it appeared, when it appeared, and how to find it. Use a book's title and copyright pages, the article itself, or the web page's author and publication details. Save volume, issue, page range, DOI, and URL when present. An access date records when you viewed web content; it is a different date from publication.

The same source changes shape across styles. For a format illustration using page 143 of Ward and colleagues' 2017 article, MLA writes “(Ward et al. 143),” while APA writes “(Ward et al., 2017, p. 143).” The locator here illustrates punctuation; a real citation must identify the passage you used.

Handle missing information honestly

A named organization can be an author. If no author is identifiable, start the reference with the title and use an appropriate short title in the text. MLA avoids repeating an organization as both author and publisher when the same body performs both roles. With no date, MLA omits that element; APA uses “n.d.”; the Harvard form taught here uses “no date.”

Do not invent page numbers for an unpaginated page. Use an available chapter, paragraph, or section locator when the style calls for one. MLA may need only the author for an unpaginated paraphrase; an APA direct quotation needs a usable alternative locator.

Use a style consistently

MLA is common in English and the humanities; APA in the social sciences; Chicago in history. Harvard has several institutional versions. Use the required system and edition: MLA 9, APA 7, Cite Them Right Harvard, and Chicago 18 are the versions illustrated here.

Citation generators can import names or titles incorrectly, omit a date, or confuse an article with a whole website. Compare their output with the actual source and the required style. Finally, match citations to entries in both directions, allowing the style's text-only exceptions. A perfect-looking entry cannot help readers if its in-text name points somewhere else.

MLA makes the path from a claim to a passage short: author and locator in the text, then publication details in Works Cited. Use the MLA 9 form consistently.

Write the in-text citation

For a format example using page 45, write “(Caulfield and Wineburg 45).” There is no comma or “p.” If your sentence names the authors, give only “(45).” With three or more authors, use the first author's last name and “et al.”: “(Ward et al. 143).” A short quotation's citation comes before the sentence period.

No named author? Use a shortened title that matches the entry, such as “("Phone Rules" 3).” No page numbers? Do not invent any: an author-only citation may work. For numbered chapters, use a label, such as “(Shelley, ch. 10).” The locator must refer to the source you read, not a page in a different edition.

Build the full entry

The core elements run in this order: author; source title; container title; other contributors; version; number; publisher; date; location. Omit elements that do not apply. A container is the larger work holding the source: an article sits in a journal, and a web page sits on a website.

Book:

Caulfield, Mike, and Sam Wineburg. Verified: How to Think Straight, Get Duped Less, and Make Better Decisions about What to Believe Online. U of Chicago P, 2023.

Reverse only the first author's name. Italicize the book title. MLA shortens “University” and “Press” in publisher names to “U” and “P.”

Journal article:

Ward, Adrian F., et al. “Brain Drain: The Mere Presence of One’s Own Smartphone Reduces Available Cognitive Capacity.” Journal of the Association for Consumer Research, vol. 2, no. 2, 2017, pp. 140-54, https://doi.org/10.1086/691462.

The article title takes quotation marks; the journal title is italic. Three or more authors become the first author plus “et al.” in this entry too. The DOI keeps its full link form. MLA can shorten the second number in a page range: 140-154 becomes 140-54.

Web page:

Gottfried, Jeffrey. “About Half of Americans Support All-Day School Cellphone Bans.” Pew Research Center, 29 July 2026, www.pewresearch.org/short-reads/2026/07/29/more-americans-now-support-than-oppose-all-day-school-cellphone-bans/.

Record the title displayed on the page, even when the URL contains older wording. MLA commonly drops the URL's https prefix. Longer month names are abbreviated, as in “Mar.” and “Sept.”

For an online novel, record the edition actually consulted. An entry for the Gutenberg edition is: Shelley, Mary. Frankenstein; or, The Modern Prometheus. 1818. Project Gutenberg, 1 Oct. 1993, www.gutenberg.org/ebooks/84. The early year records the original publication; the later date belongs to this online edition.

For a supposed interview with a named expert, the pattern is: “Rivera, Ana. Personal interview. 12 Mar. 2027.” Protect identities promised anonymity in your own research.

Arrange the paper and list

Begin Works Cited on a new page with its heading centered. Alphabetize by the entry's first element, usually the author's last name. Double-space and use a hanging indent of about 1.27 cm for lines after the first. A usual MLA paper has name, teacher, course, and date at the top left of page 1, with last name and page number at the top right; it normally has no separate title page.

Check author names, title marks, containers, and punctuation. Listing sources in order of first mention makes entries harder to find and does not follow MLA's alphabetical system.

APA shows when research appeared as well as who produced it. Its author-date system uses different punctuation from MLA, and its reference entries use initials and sentence-style titles.

Distinguish a citation from a reference

A paraphrase needs author and year: “(Beland & Murphy, 2016).” In running text write, “Beland and Murphy (2016) found…” The ampersand belongs inside parentheses; “and” belongs in the sentence. Three or more authors use the first author's last name plus “et al.” from the first citation.

A direct quotation needs a locator. For a punctuation illustration using page 143, write “(Ward et al., 2017, p. 143).” Use “pp.” for multiple pages and an appropriate paragraph or section locator for an unpaginated quotation. A locator is optional for a paraphrase but can help with a long work.

Follow who, when, what, where

Book:

Caulfield, M., & Wineburg, S. (2023). Verified: How to think straight, get duped less, and make better decisions about what to believe online. University of Chicago Press.

Use initials, put the year after the author, and capitalize the first word of the title and subtitle plus proper nouns. The title is italic.

Journal article:

Ward, A. F., Duke, K., Gneezy, A., & Bos, M. W. (2017). Brain drain: The mere presence of one’s own smartphone reduces available cognitive capacity. Journal of the Association for Consumer Research, 2(2), 140–154. https://doi.org/10.1086/691462

The article title has sentence capitalization and no quotation marks. The journal name retains its capitals. Italicize the journal and volume, but not the issue. List up to 20 authors in a reference entry even though the short citation may use “et al.”

Web page:

Gottfried, J. (2026, July 29). About half of Americans support all-day school cellphone bans. Pew Research Center. https://www.pewresearch.org/short-reads/2026/07/29/more-americans-now-support-than-oppose-all-day-school-cellphone-bans/

Use the page's full publication date when supplied. If the author is an organization with the same name as the site, omit the repeated site name. With no date use “(n.d.).” With no author move the title into the author position and use a short title in the citation.

Handle unretrievable information and your data

An email or informational interview with an expert that readers cannot retrieve is a personal communication. A supposed example is “(A. Rivera, personal communication, March 12, 2027).” It appears only in the text, with no References entry.

Participants in your own study are different: their words are your data. Explain your method and identify quoted speech as a participant's response. Use agreed pseudonyms and preserve anonymity. Do not make personal-communication citations or reference entries for each participant.

Format a student paper

A student title page gives the title, your name, affiliation, course number and name, instructor, and due date. Put the page number at the top right, starting on the title page. Begin a new page headed References, centered and bold. Alphabetize by the first author's last name, double-space, and use hanging indents of about 1.27 cm.

Check the common carryovers from MLA: full first names, headline-style article titles, quotation marks around article titles, and a missing year. A four-author article still needs all four authors in its reference entry.

Harvard places author and year in the text. Footnote styles place a note number there instead. Both need enough information for readers to trace the source, but they arrange that information differently.

Use a particular Harvard guide

Harvard is a family of author-date styles. The forms here follow Cite Them Right. Follow your institution's own guide when it specifies a different Harvard version.

For a page-format illustration, write “(Caulfield and Wineburg, 2023, p. 45).” Name both authors with “and.” With three authors, name all three; with four or more, use the first author's last name and “et al.” Give pages for quotations and ideas from particular pages; a whole-source summary needs no page locator.

Book:

Caulfield, M. and Wineburg, S. (2023) Verified: how to think straight, get duped less, and make better decisions about what to believe online. University of Chicago Press.

Unlike APA, this form uses “and,” has no period immediately after the parenthesized year, and uses the Harvard guide's title capitalization.

Journal article:

Ward, A.F., Duke, K., Gneezy, A. and Bos, M.W. (2017) ‘Brain drain: the mere presence of one’s own smartphone reduces available cognitive capacity’, Journal of the Association for Consumer Research, 2(2), pp. 140–154. Available at: https://doi.org/10.1086/691462.

The article title uses single quotation marks in this style. The journal is italic, and the page range follows “pp.”

Web page:

Gottfried, J. (2026) About half of Americans support all-day school cellphone bans. Available at: https://www.pewresearch.org/short-reads/2026/07/29/more-americans-now-support-than-oppose-all-day-school-cellphone-bans/ (Accessed: 5 November 2026).

The access date records your visit. It is separate from the publication year. Alphabetize the reference list and check punctuation against the chosen guide.

Compare without mixing styles

Copying an MLA name order into an APA entry, then adding Harvard punctuation, produces no consistent style. Change all the relevant features when converting an entry.

Understand footnotes and bibliography

Chicago notes and bibliography places a raised number in the text and a source note at the foot of the page or end of the paper. MHRA, used in literary studies, also uses notes. Your main argument remains in the body; notes identify sources and may carry brief side comments.

For a Chicago 18 format illustration, a first note can read: “1. Mike Caulfield and Sam Wineburg, Verified: How to Think Straight, Get Duped Less, and Make Better Decisions about What to Believe Online (University of Chicago Press, 2023), 45.” A later note shortens it: “2. Caulfield and Wineburg, Verified, 52.” Use the real page for the passage you cite.

The bibliography reverses the first name and uses periods: “Caulfield, Mike, and Sam Wineburg. Verified: How to Think Straight, Get Duped Less, and Make Better Decisions about What to Believe Online. University of Chicago Press, 2023.” Check which notes and list entries count toward your assignment's word limit; moving analysis into footnotes does not improve the body argument.

A long paper develops its answer through linked sections. Each paragraph has a job in that argument, and each source enters because it helps with that job.

Start with a section you can develop

Drafting need not follow the final reading order. Start with the section whose evidence you understand best. You can write the introduction later, when the answer has become clearer. Add citations as you draft so that the link between source and claim stays intact.

An introduction brings readers into a specific issue, shows why it matters, states the question or gap, and gives the thesis. A long paper may also sketch its organization. An opening such as “Since the dawn of time, people have used technology” spends words without defining the problem. Begin with the particular debate your paper addresses.

Build a paragraph around your claim

A developed body paragraph moves through claim, evidence, analysis, and a link to the thesis or next point. For example:

“School phone bans may have uneven effects on attainment. Beland and Murphy (2016) found gains after bans in schools in four English cities, driven by the lowest-achieving students. That pattern makes differences between student groups important when judging a policy's value. A school evaluating its own rule should therefore examine those groups as well as the overall result.”

The first sentence makes the paragraph's claim. The second supplies the finding. The third explains its relevance. The fourth draws a consequence for policy evaluation. The finding does not by itself explain why scores rose, so the paragraph does not assert a cause such as improved concentration.

After evidence, ask “What does this show for my question?” Description reports what a source says. Analysis explains why that information supports, limits, or changes your argument. If changing your thesis would leave a paragraph untouched, check whether the paragraph needs more analysis or a clearer purpose.

Keep the thread visible

Use headings where the assignment and style allow them. At a section boundary, show the relationship between points. “Evidence about attainment leaves a separate question: how do restrictions relate to well-being?” makes the shift clear. “Another source says” signals a new author but no logical move.

Choose words that match the evidence. “One study found gains after bans” is different from “Bans improve learning everywhere.” Use “suggests,” “in this sample,” or “may” when the evidence requires them. Firmly state rules and established facts; adding uncertainty to every sentence blurs the argument.

Finish the reasoning

A conclusion states the answer the paper has reached and its significance. It can identify a meaningful limit or a further question. It should not introduce a major new study that the body never examined or repeat the introduction sentence for sentence.

Label tables and figures, mention them in the text, and cite their sources. Put extensive raw data in an appendix when appropriate. Keep the word budget in view, but complete the draft before cutting it heavily. A full draft lets you see which background or repeated summary can go while the reasoning remains.

A literary research paper argues how a work creates meaning. Its main evidence is the language and form of the work itself; context and critics help you develop and test your reading.

Begin with the texts

A focused comparison might ask, “How do Frankenstein and Never Let Me Go use first-person narration to win sympathy for beings created for other people's purposes?” The question names the works, a method, and an effect. It directs reading toward voice and readers' judgments.

Read closely before consulting critics. Note your first ideas and the passages behind them. Choose short quotations that allow you to analyze particular words, images, narrative choices, or structure. Use the present tense for events within a work: the creature “tells” his story. Retelling what he does provides plot information but does not explain how the telling shapes sympathy.

Bring context into the passage

Context includes conditions of writing, reception, literary tradition, and debate. Use it when it changes the reading of particular evidence. Frankenstein appeared in 1818, amid interest in electricity and experiments involving bodies. That context can help explain the imaginative setting of Victor's project, but it does not settle how to interpret every scene.

A paragraph of historical facts without a link to language or form interrupts the analysis. Instead, identify the relevant context, return to the passage, and explain what the connection helps you see. An author's life may inform a question, but it cannot substitute for evidence about what the text means.

Make critics part of a discussion

Find critical views in scholarly editions, book-length studies, JSTOR, and the MLA International Bibliography. Record the critic's actual argument and evidence. Use a critic when you can agree and extend, qualify, or challenge the reading.

For a supposed critical view, imagine a critic treats a narrator's account as completely trustworthy. Your response needs a passage where the narrator's language complicates that confidence. A useful frame is, “This reading explains the narrator's appeal for sympathy, but it leaves the contradiction in this passage unresolved.” Follow it with close analysis. Adding a critic's name to your claim does not itself develop a debate.

Compare throughout the argument

Organize sections around shared ideas such as narrative access, creators' responsibility, and endings. Within each section, examine both works and make the relation clear. Two long halves, each devoted to one novel, can leave the comparison until the conclusion, where there is too little room to develop it.

Fair attention does not require exactly equal word counts in every paragraph. It requires enough evidence and analysis to justify the comparison. Explain the important similarity or difference instead of assuming readers will infer it from two summaries.

Name the editions you consulted and check each quotation against them. Italicize novels and plays; put short poem titles in quotation marks. Follow any requirement to separate primary literary works from secondary criticism in the final list.

Finishing a paper means improving the thinking as well as the sentences. Work from large decisions to small corrections so that you spend time on material that will remain in the final paper.

Revise the argument first

Revision changes the thesis, structure, evidence, and reasoning. Editing improves sentences and word choices. Proofreading checks final spelling, punctuation, and format. Correcting every comma before deciding whether a section belongs can waste a great deal of work.

Make a reverse outline: write one line stating each paragraph's claim, in draft order. Read the lines alone. Repeated claims may need merging; a claim unrelated to the thesis needs changing or cutting; a missing sub-question reveals a gap. If paragraphs 4 and 9 both explain the same benefit, compare their evidence and build one stronger paragraph.

Check the assignment and rubric requirement by requirement. Mark where the draft actually meets each one. A section that describes sources may still fail a requirement to evaluate them. Recheck the introduction's thesis after revision; the body may now defend a more precise answer.

Ask for useful feedback

Give a reader a question such as “Where do you stop following my reasoning?” or “Which claim needs stronger evidence?” Leave time to act on the reply. You decide what to change, but repeated confusion at one passage is a signal to examine that passage carefully.

Audit the evidence

Check quotations word for word, verify numerical claims, and compare paraphrases with the originals. Trace every in-text citation to its matching entry, then trace the entries back to the text. Respect exceptions such as APA personal communications. Remove uncited works from a reference list unless the requested bibliography allows wider reading.

A citation such as “(Ortiz 12)” fails its purpose if there is no matching Ortiz entry. A reference entry cannot repair wording that misstates what Ortiz found; accuracy and formatting are separate checks.

Cut without removing the reasoning

Check what the word limit includes. A 2,890-word draft facing a 2,500-word limit needs a cut of 390 words, about 13.5% of the draft. One plan is to remove 150 words of excess background, 120 of repeated source summary, and 120 of wordy sentences. That reaches the limit while preserving analysis.

Replace filler such as “it is important to note that” with the point itself. Give vague nouns a clear subject and choose direct verbs: “The study shows” is clearer than “The study is an indication of.” Check pronoun references and verb tenses.

Proofread the final form

Read aloud or change the display to notice familiar errors. Reading sentences from the end backward can help you focus on spelling. A spell-checker may miss “form” where you meant “from.”

Check title, identifying details, page numbers, spacing, headings, final list, appendices, and any required word count. Save a clearly named final file, keep earlier drafts and a backup, and leave time for submission problems.

A senior project, or capstone, connects research with independent work that shows new learning. A paper explains the evidence; a product or experience lets you apply and test what you learned.

Make a feasible proposal

Schools set their own requirements, but a senior project often has a paper, a product or field experience, a process portfolio, and a presentation. Read the handbook for deadlines, hours, mentor requirements, and assessment rules before making commitments.

Choose something that interests you, stretches your skills, and fits the time and resources available. A proposal normally states the question, product, skills to learn, timeline, materials and costs, and possible mentor. Make the link between the paper and product clear enough that a reader can see why each belongs.

A mentor offers knowledge, feedback, and guidance. Contact an appropriate adult early, check the school's eligibility rules, agree on meetings, and record them. Guidance should help you make decisions and learn skills while the assessed work remains yours.

Connect research to action

Suppose you research cafeteria food waste, then run an approved two-week trial offering smaller portions. For an illustrative metric example, weekly waste falls from 60 kg to 45 kg. The reduction is 15 kg; dividing 15 by the starting 60 gives 0.25, or 25%.

The paper explains relevant published research, how the trial worked, and what happened. A short trial at one school leaves other possible influences, such as menu or attendance changes. The result supports discussion of that trial; it does not establish that the same reduction will follow everywhere.

A product may also be creative: a film, event, story collection, design, or built object. Explain how research shaped particular choices. Credit borrowed text, images, music, and data in the product as well as in the paper.

Keep a portfolio as the work happens

A portfolio can hold the proposal, dated hours log, mentor records, source notes, drafts, product evidence, and reflections. “March 3, two hours: compared waste measurements with cafeteria staff; changed the weighing process after inconsistent readings” shows work and a decision. A log rebuilt from memory at the end often misses both.

Reflection goes beyond sequence. “I collected responses and wrote the paper” names events. A reflection explains a change: “The first survey combined two judgments, so I split the question. The revised answers showed me why neutral wording matters.” Connect a setback to what you learned and what you would do differently.

Present the question, evidence, and learning

For a 12-minute talk, one possible plan gives 1 minute to the opening, 2 to the question and its importance, 5 to methods and findings, 3 to learning, and 1 to the close. Use a few clear visuals with credits and practice aloud with a timer.

Prepare for questions about limits, setbacks, and next steps. If the product no longer fits the research, discuss the change with your mentor or teacher early and update the plan. A well-kept portfolio gives you specific evidence to explain your decisions.