Where to Find Reliable Sources on Abortion Research
When you're writing a research paper on this topic, the first problem you run into is that almost every source has an angle. That's not a criticism — it's just the reality of the field. Your job is to triangulate between different types of data so the conclusions you draw actually hold up. Let me walk through what I've learned from actually writing papers on this and advising students who were in the same spot. The biggest mistake people make is starting with advocacy literature. Organizations like the Guttmacher Institute and the Center for Reproductive Rights publish useful data, but their framing is directional. That's fine if you know how to read against it. You use their statistics, but you verify them against primary government sources whenever possible. In the United States, the CDC's Abortion Surveillance System is your baseline. It publishes annual reports going back decades. The data is self-reported by states, which means coverage varies. About 80 to 85 percent of states report consistently year over year. The remaining states either don't collect the data or don't send it in. When you cite CDC figures, always note which states are included in the denominator. A reader who knows what they're looking at will spot a missing-data problem immediately.
On the international side, the World Health Organization's Abortion Care Guideline and their annual estimates are the standard reference. They use a modeling approach that accounts for both legal and illegal abortions. The model has been criticized, and fairly so. Their 2022 revision changed some of the assumptions around what counts as an induced abortion versus a spontaneous miscarriage in their datasets. If you're doing a literature review, you need to understand whether a study's numbers come from direct observation, from hospital records, or from WHO's model. These three methods produce very different results, and they're not interchangeable. I spent two weeks last year trying to reconcile a 2019 Guttmacher figure with a contemporaneous CDC report for a paper on post-Roe state-level policy changes. The numbers for the same year diverged by roughly 15 percent on total procedure counts. The gap came down to two things: CDC excluded states that didn't report that year, and Guttmacher's methodology imputes missing data using a regression model. Neither is wrong, but they answer slightly different questions. I ended up presenting both figures with the methodological caveat attached. That's the honest approach.
Primary Source Types and Where They Live
Legislative text is the most underused source in student papers. State bills, enacted statutes, and court opinions are free on sites like Constitute Project, StateLegislation.com, and CourtListener. These give you the actual law, not a news summary of the law. A bill's text tells you about gestational limits, mandatory waiting periods, and ultrasound requirements in a way that a newspaper article never will. Court opinions are essential. The full-text archive at Justia and the free Supreme Court database are adequate. I always pull the majority opinion and the dissent side by side. The dissent in Dobbs, for instance, contains a detailed factual recitation about maternal mortality that the majority largely sidesteps. Both are legitimate interpretive moves. Citing only one makes your paper look like you picked a side rather than engaged with the reasoning. Peer-reviewed medical literature lives in journals like American Journal of Public Health, Obstetrics & Gynecology, The Lancet, and Contraception. Access usually requires institutional logins. If you're a student, your library portal will get you through. A useful trick: search Google Scholar with the site operator limited to your university domain. That surfaces open-access versions your institution already pays for.
Get the Full Details
Common Pitfalls That Make Papers Look Amateur
Using anecdotal evidence as a data point. A viral story on social media is not a dataset. I see this constantly. A single account of someone traveling across state lines for care is emotionally compelling but methodologically worthless. Use it only if you're writing a qualitative paper, and even then, get IRB approval first. Mixing up correlation and causation in policy analysis. A state passes a targeted regulation of abortion providers and clinic numbers drop. That doesn't mean the regulation caused the drop. It could be concurrent economic conditions, a hospital closure, or a provider retirements. You need a difference-in-differences design or at minimum a pre-post comparison with a control state to make a causal claim. I ran into this exact problem when a classmate tried to argue that Pennsylvania's 2014 ultrasound mandate caused a decline in first-trimester procedures. The raw numbers showed a drop, but the control group of neighboring Ohio (which didn't pass a similar law) showed a similar decline that year. The apparent effect vanished once you controlled for regional trends.
How to Structure Your Literature Review Without It Becoming a List
Don't organize by source type. Organize by debate. Group studies that disagree with each other, then explain why. A section might look like this: one group finds that restrictive laws reduce abortion rates significantly, another group finds no measurable change, and a third finds increases in out-of-state travel that partially offset local restrictions. Then discuss the methodological differences that explain the divergence. Use a citation manager from day one. Zotero is free and handles PDF annotation well. EndNote is better if you're doing heavy bibliometric work. I don't recommend doing this manually. A paper on abortion policy easily accumulates 60 to 100 sources, and managing those in Word without a tool is painful.
Data Verification Checklist
Before you cite any statistic, ask yourself: who collected this, how, when, and what's excluded. Government health departments publish technical documentation alongside their data releases. Read it. It's boring. It takes 10 minutes. It will save you from citing a number that was later retracted or revised. The CDC revises its abortion data annually as late-reporting states catch up. A 2020 figure published in a 2021 report may change in the 2022 report. If you're writing a paper in 2025 that cites 2020 data, pull the latest available version, not the version that was current when the original report came out. For economic impact studies, the source matters more than the conclusion. Studies funded by advocacy organizations tend to produce different estimates than those funded by government grants or academic institutions. This isn't about dishonesty. It's about study design. Funding source correlates with sample selection, geographic scope, and outcome measures. Note it in your methodology section. Reviewers will look for it.
