Here's a secret nobody tells you in your first methods course: professional psychologists almost never read a journal article straight through, the way you'd read a chapter. They can't — a working researcher might need to process thirty papers a week. Instead they exploit the fact that every empirical article in psychology has the same anatomy, so they know exactly where each kind of information lives and go straight to it. That standardization is not an accident and it's not bureaucratic fussiness. It's a scientific value, nearly a century old, and once you understand why it exists, APA style stops being a list of arbitrary rules and starts being a map. This lesson gives you the map: the anatomy of an article, a strategic reading method, the citation mechanics that graders actually check, and the honesty norms — plagiarism, paraphrasing, AI — that the whole system runs on.
The anatomyWhat each section is really for
An empirical article runs in a fixed order: Title page, Abstract, Introduction, Method, Results, Discussion, References. The order is not arbitrary — it's an argument unfolding, and each section answers one question a skeptical reader would ask next.
The title announces the variables — a good psychology title usually names the constructs and hints at the relationship between them ("Real men don't eat (vegetable) quiche: Masculinity and the justification of meat consumption"; Rothgerber, 2013 — the parenthetical joke is optional but appreciated). The Abstract compresses the entire study — question, method, key result, takeaway — into roughly 150–250 words. It is, statistically speaking, the only part most people will ever read, which is why authors sweat over it more than any other paragraph in the paper.
The Introduction answers why should anyone care about this question? It reviews what's already known, exposes the gap, and ends on the study's specific hypotheses — always at the end, never the beginning. The Method answers what exactly did you do? It's the replication recipe: who the participants were, what they were given, in what order, with enough precision that a stranger could rerun the study. When Lesson 15 talks about the replication crisis, this is the section replicators live in. The Results answers what did you find? — statistics reported plainly, with no interpretation. The Discussion finally answers so what?: what the findings mean, how they connect back to the theories from the Introduction, what the study can't tell us (limitations), and what should happen next. The References section — which begins on its own page — is the paper trail: every source cited in the text, and nothing that isn't.
The classic guide to this structure is Daryl Bem's chapter "Writing the empirical journal article," which describes the paper's hourglass shape (Bem, 2004). The Introduction runs broad → narrow: open with a question a smart non-psychologist would care about ("Why is meat-eating so tangled up with manhood?"), funnel down through the prior literature, and land on the tip — your specific, testable hypothesis. The Method and Results are the narrow waist: maximally specific, welded to your data. Then the Discussion re-widens: interpret the finding, reconnect it to the theory you opened with, and close as broadly as you began. If a reader can't feel the funnel narrowing across your Introduction, your paragraphs are in the wrong order — that's the diagnostic.
Which section?Sort the sentence
The most common drafting error is putting the right sentence in the wrong section — interpreting inside the Results, or smuggling new numbers into the Discussion. Tap each sentence to see where it belongs and why.
Which section does this sentence belong in? Tap one.
Reading strategicallyHow to actually read one
Now the part that will save you the most hours in this course: do not read an article front to back on your first pass. Jordan and Zanna (1999), in a guide written precisely for students meeting journal articles for the first time, make the case that reading a research article is a skill of its own — the article is a dense, technical argument, and you need the shape of the argument before the details can mean anything.
So make two passes. On the first pass, read the title and Abstract to get the study in miniature. Then jump to the last paragraph of the Introduction, where the hypotheses live. Then skip ahead to the opening of the Discussion, where the authors state in plain language what they think they found. Ten minutes in, you now know the question, the prediction, and the claimed answer — the whole argument in skeleton form. Only then, on the second pass, read the Method and Results, checking the evidence against the claim you already understand. Reading the Results before you know the hypothesis is like watching a replay of a game when you don't know which team is which.
What are you checking on that second pass? Two things you already own from earlier lessons. First, the claim: is this study making a frequency claim, an association claim, or a causal claim — and does the design actually license it? A correlational Method paired with a causal-sounding Discussion is the single most common mismatch in the published literature, and you are now equipped to catch it. Second, the validities — construct, external, statistical, internal — the interrogation framework from Lesson 03, drawn from Shadish, Cook, and Campbell (2002). You don't need every validity to be strong; you need to know which ones the claim leans on.
Claim: What is the one-sentence claim, and is it descriptive, associational, or causal? Does the design (measured vs. manipulated variables, random assignment or not) support that verb — "predicts" is not "causes"? Construct validity: How was each variable actually operationalized, and would you accept that as a measure of the concept? Sample: Who were the participants, and how far are the authors generalizing beyond them? Numbers: How big is the effect, not just how significant? Candor: Do the limitations the authors admit match the limitations you spotted — or are the real ones missing?
Why APA style existsStandardization as a scientific value
APA style did not descend from a committee of pedants. It began in 1929 as a seven-page article in Psychological Bulletin, written by a group of psychologists and journal editors trying to solve a real problem: every author formatted papers differently, which made articles slow to read, expensive to typeset, and hard to compare (Bentley et al., 1929). They called their guidelines a "standard of procedure, to which exceptions would doubtless be necessary" — a working convention, not scripture. That seven-page memo grew into the Publication Manual, now in its 7th edition at 400-plus pages (American Psychological Association [APA], 2020), and along the way it was adopted far beyond psychology — nursing, education, business, and communication all write in APA.
The underlying value is worth naming, because it explains almost every rule in the manual: in science, format is standardized so that content can be compared. When every article puts the hypotheses at the end of the Introduction and the statistics in the Results, a reader can find, verify, and build on any study's evidence without first decoding the author's personal organizational whims. Uniform citation rules serve the same end — they make every claim traceable. A citation is not decoration; it's a hyperlink in prose, and the reference list is the guarantee that every link resolves. That's why the golden rule is absolute: everything cited in the text appears in the References, and everything in the References is cited in the text.
And because APA style is a working convention rather than scripture, it evolves — nowhere more visibly than in its language guidelines.
Early editions of the manual said little about how to describe people; APA 7 devotes an entire chapter to it (APA, 2020). Use the singular "they" when gender is unknown or irrelevant ("each participant rated their mood") — now explicitly endorsed, retiring the clunky "he or she." Choose between person-first language ("a person with schizophrenia") and identity-first language ("an autistic person") based on how the community in question describes itself — many disability communities actively prefer identity-first, so ask rather than assume. Be specific: report age ranges rather than "the elderly," name the actual racial and ethnic groups sampled rather than gesturing at "minorities." The point isn't politeness — it's precision. "The elderly" is vague measurement; "adults aged 65–79" is data. Notice that this standard has changed edition over edition and will change again: bias-free language is a live scientific norm, updated as the field learns, exactly like any other methodological standard.
The manual's 2020 revision made several changes that older websites — and older professors' handouts — still get wrong. Student papers no longer need a running head, just a page number top-right. One space after a period, not two. In the reference list, list up to 20 authors before resorting to an ellipsis (the old cap was 7); for 21 or more, list the first 19, an ellipsis, then the final author. DOIs are formatted as full https URLs (https://doi.org/10.xxxx), not "doi:". And in-text, use "et al." for three or more authors from the very first citation — the 6th edition's spell-them-all-out-once rule is gone. Small changes, but they're precisely the ones graders scan for (APA, 2020).
The mechanicsCitations that don't lose points
Every idea you borrow gets an in-text citation, and the format shifts with the number of authors and the way the citation sits in the sentence. A parenthetical citation rides at the end in parentheses with an ampersand; a narrative citation makes the authors the subject of your sentence, spells out "and," and puts only the year in parentheses. This table covers roughly 95% of what an undergraduate paper needs. (Author names in the generic rows are placeholders — the forms are what matter.)
| Situation | Parenthetical | Narrative (in the sentence) |
|---|---|---|
| One author | (Neff, 2003) | Neff (2003) found… |
| Two authors | (Kim & Osei, 2021) | Kim and Osei (2021) reported… |
| Three or more authors | (Alvarez et al., 2022) | Alvarez et al. (2022) showed… |
| Group / organization | (American Psychological Association [APA], 2020) | The American Psychological Association (APA, 2020) states… |
| Direct quotation | (Neff, 2003, p. 224) | Neff (2003) defined it as "…" (p. 224) |
| Same author, same year | (Neff, 2003a, 2003b) | Neff (2003a) proposed… Neff (2003b) then validated… |
| Source you couldn't access | (Freud, 1915, as cited in Smith, 2020) | Freud (1915, as cited in Smith, 2020) argued… |
Two of those rows deserve a second look. Direct quotations always carry a page number — a quote without one is an automatic markdown in most rubrics. And the "as cited in" form is for the rare case where you read Smith's account of Freud but not Freud himself: only Smith, the source you actually read, goes in your reference list. Use it sparingly — it's an admission, not a technique.
Now the reference list itself. Every entry, whatever the source type, answers four questions in fixed order — who, when, what, where:
| Element | Rule | In the example |
|---|---|---|
| Who | Authors, surname + initials, in the original order (never alphabetize within a paper's authors — order encodes credit) | Neff, K. D. |
| When | Year in parentheses | (2003). |
| What | Article title in sentence case — capitalize only the first word, the first word after a colon, and proper nouns | The development and validation of a scale to measure self-compassion. |
| Where | Journal in italics and Title Case, volume italicized, issue in parentheses, pages, DOI as an https URL | Self and Identity, 2(3), 223–250. https://doi.org/10.1080/15298860309027 |
The DOI — digital object identifier — is the piece students most often mangle. It's a permanent ID assigned to a published work, and APA 7 wants it as a full clickable URL. Journal URLs rot; DOIs don't. If a work has one, include it; never invent one, and never trust one an AI tool hands you without checking that it resolves (more on that in a moment). Entries are alphabetized by first author's surname and formatted with a hanging indent.
HonestyPlagiarism, paraphrasing, and the AI question
Everything above is mechanics. This section is the ethics under the mechanics, and it follows directly from Lesson 04: scholarship runs on accurate credit. Plagiarism is presenting someone else's words or ideas as your own — and the definition students most often miss is that changing a few words is not paraphrasing. Roig (1997) gave undergraduates an original paragraph plus rewritten versions, some properly paraphrased and some superficially disguised, and asked which were plagiarized. Up to half the students judged lightly reworded copies — same sentence structure, a few synonyms swapped — to be acceptable paraphrases. They're not. That pattern even has a name, patchwriting, and citation doesn't cure it: if the sentence architecture is someone else's, the writing is someone else's.
The reliable test for honest paraphrasing: read the passage, close the source, explain the idea from memory in your own sentence structure, then cite it — because the idea still isn't yours even when the words are. If you can't write it with the source closed, you don't understand it yet, which is useful information. Quote directly only when the original wording itself matters (a definition, a famous phrase), and remember the page number.
Self-plagiarism surprises people: you can plagiarize yourself, by recycling your own previous paper — or big chunks of it — into a new assignment without disclosure (APA, 2020). The logic is the same as for data: each submission implicitly claims to be new work. If you want to build on your own earlier paper, ask your instructor; the answer is often yes, with a citation.
And then there's AI. A frank word, because you will use these tools and pretending otherwise helps no one. Large language models are genuinely useful for some parts of this work: explaining a concept you're stuck on, critiquing a draft you wrote, generating counterarguments to stress-test your thinking. They are hazardous for two specific things. First, submitting AI-drafted prose as your own writing is the same act as submitting a friend's prose as your own writing — the label on the tool doesn't change the transaction. Check your instructor's policy; in this course, undisclosed AI-written text is treated as plagiarism, and disclosed, appropriately-used assistance is not a scandal. Second — and this one bites even honest users — language models fabricate citations. They produce plausible-looking references, correct-sounding authors, realistic DOIs, for papers that do not exist. Every reference an AI hands you must be verified against the actual source before it enters your paper. A fabricated citation in your reference list is your fabrication once you submit it.
There's an official rule, so don't guess it. APA treats a large language model as software, with the company as author (McAdoo, 2023). The reference looks like this:
OpenAI. (2023). ChatGPT (Mar 14 version) [Large language model]. https://chat.openai.com/chat
In text: (OpenAI, 2023). Because nobody else can retrieve your exact chat, APA says to describe the prompt you used in the text or put the full exchange in an appendix. And keep the hierarchy straight: an AI is not a peer-reviewed source — it can be a cited tool, but it can never stand in for evidence. Evidence comes from the literature, which is why you learned to read it first.
The craftClarity beats jargon
One last thing, because following every rule above can still produce an unreadable paper. Good scientific writing is not a separate, duller dialect of English. Sword (2012) analyzed the prose of a large sample of published academic articles across disciplines and found that the most effective writers do the opposite of what students assume: they use concrete examples, active verbs, first-person statements ("we predicted"), and even the occasional story — while the stereotypical academic mush of nominalizations and passive hedges mostly signals insecurity, not rigor. Gernsbacher (2018), writing specifically about psychology, ties clarity to scientific values: plain words, active voice, minimal abbreviations, and concrete examples don't just make a paper pleasant — they make it more transparent, easier to evaluate, easier to replicate, and easier to remember. Writing clearly is a methodological virtue, not a stylistic garnish.
The practical version for your own reports: prefer the short word to the long one when they mean the same thing; let people do things in your sentences ("participants completed the scale," not "the scale was completed"); and when you catch yourself writing a sentence you don't fully understand, stop — that sentence is patchwriting waiting to happen. If you can explain your study to a curious friend at lunch, you can write the paper. The APA skeleton tells you where everything goes; clarity is what makes anyone want to read it.
SourcesCited in APA 7
American Psychological Association. (2020). Publication manual of the American Psychological Association (7th ed.). https://doi.org/10.1037/0000165-000
Bem, D. J. (2004). Writing the empirical journal article. In J. M. Darley, M. P. Zanna, & H. L. Roediger III (Eds.), The compleat academic: A practical guide for the beginning social scientist (2nd ed., pp. 185–219). American Psychological Association.
Bentley, M., Peerenboom, C. A., Hodge, F. W., Passano, E. B., Warren, H. C., & Washburn, M. F. (1929). Instructions in regard to preparation of manuscript. Psychological Bulletin, 26(2), 57–63. https://doi.org/10.1037/h0071487
Gernsbacher, M. A. (2018). Writing empirical articles: Transparency, reproducibility, clarity, and memorability. Advances in Methods and Practices in Psychological Science, 1(3), 403–414. https://doi.org/10.1177/2515245918754485
Jordan, C. H., & Zanna, M. P. (1999). How to read a journal article in social psychology. In R. F. Baumeister (Ed.), The self in social psychology (pp. 461–470). Psychology Press.
McAdoo, T. (2023, April 7). How to cite ChatGPT. APA Style Blog. https://apastyle.apa.org/blog/how-to-cite-chatgpt
Neff, K. D. (2003). The development and validation of a scale to measure self-compassion. Self and Identity, 2(3), 223–250. https://doi.org/10.1080/15298860309027
OpenAI. (2023). ChatGPT (Mar 14 version) [Large language model]. https://chat.openai.com/chat
Roig, M. (1997). Can undergraduate students determine whether text has been plagiarized? The Psychological Record, 47(1), 113–122. https://doi.org/10.1007/BF03395215
Rothgerber, H. (2013). Real men don't eat (vegetable) quiche: Masculinity and the justification of meat consumption. Psychology of Men & Masculinity, 14(4), 363–375. https://doi.org/10.1037/a0030379
Shadish, W. R., Cook, T. D., & Campbell, D. T. (2002). Experimental and quasi-experimental designs for generalized causal inference. Houghton Mifflin.
Sword, H. (2012). Stylish academic writing. Harvard University Press.