What Should You Look for When Screening Titles and Abstracts?

What Are You Actually Trying to Decide From a Title and Abstract?

Title and abstract screening can look deceptively simple. Read the title, scan the abstract, click include or exclude, and move on.

The difficulty begins when the record is not obviously relevant or irrelevant. The population sounds close to yours but is described vaguely. The intervention fits, but the setting is unclear. The abstract never mentions the outcome you need. A study sounds promising, yet you cannot determine its design. Now what?

The purpose of this stage is not to understand each paper completely. It is not even to make the final inclusion decision in every case. You are trying to determine whether the information available in the title and abstract is sufficient to establish that a record is clearly ineligible, or whether it remains sufficiently plausible to warrant further assessment.

That distinction matters because screening too loosely creates unnecessary full-text work, while screening too aggressively can remove relevant evidence before you ever see the paper.

Screen Against Your Eligibility Criteria, Not Against a General Impression of Relevance

Title and Abstract Screening Is an Exercise in Ruling Out, Not Proving In

Start With the Criteria You Already Defined

Screening becomes inconsistent when researchers approach each abstract with a vague question such as, “Does this look relevant?” Two papers can feel equally relevant while satisfying very different parts of the review question.

A more defensible approach is to translate your inclusion and exclusion criteria into screening questions.

Depending on the review, you might ask:

• Does the population appear eligible?
• Does the study address the intervention, exposure, phenomenon, or concept of interest?
• Is the setting or context within scope?
• Does the study design appear eligible?
• Does the publication type satisfy the review criteria?
• Are any other explicitly defined eligibility requirements clearly met or clearly violated?

Not every review needs every one of these questions. The screening form should reflect the actual eligibility criteria rather than becoming a generic checklist that researchers mechanically apply to every topic.

This is why study selection begins before screening. You first need a defensible basis for deciding which papers belong within the review. Title and abstract screening then applies those boundaries to the records your search has retrieved.

The Title Is Useful, but It Rarely Tells the Whole Story

A title can sometimes establish ineligibility immediately. If your review concerns university students and the title explicitly states that the study investigated preschool children, you may already have enough information to exclude the record on population.

Other titles are much less informative:

“Student Experiences With Artificial Intelligence in Education”

Which students? What kind of artificial intelligence? What educational setting? Is this empirical research, a commentary, or a review?

A title like this should not be interpreted more confidently than the information allows. If it could plausibly describe an eligible study, read the abstract.

The objective is not to eliminate records as quickly as possible. It is to eliminate them as quickly as the evidence permits.

Population: Are These the People or Units Your Question Concerns?

Population is often one of the easiest characteristics to identify, but terminology can create ambiguity.

Suppose your review includes undergraduate university students. An abstract explicitly describing elementary-school pupils is clearly outside that criterion. One describing “college students” may clearly qualify depending on how your criteria were defined. Another referring only to “learners” may require further investigation.

Population can also be mixed. A study might include secondary-school and university students but provide separate results for each group. Whether that study is eligible depends on your criteria and whether usable information for the eligible subgroup can be identified.

Do not reduce population screening to spotting the right noun. Ask whether the participants or units of analysis actually correspond to the population defined by the review.

Intervention, Exposure, or Phenomenon: Is the Study Actually About What You Are Reviewing?

Shared terminology can conceal substantial conceptual differences.

A review of generative AI-generated feedback might retrieve studies of automated essay scoring, intelligent tutoring systems, conventional chatbots, machine-learning prediction, and rule-based feedback systems. All may use terms such as “artificial intelligence,” “automated feedback,” or “digital learning.” They are not necessarily investigating the same phenomenon.

Read enough of the abstract to determine what was actually studied rather than relying on keyword overlap.

This becomes particularly important when technologies, interventions, or concepts have changed over time. The terminology used by authors may be broader, narrower, older, or simply different from the terminology used in your review.

Context: Does the Setting Matter to Eligibility?

Some review questions are highly context-dependent. A study conducted in a hospital may not answer a question specifically about primary care. Workplace training may not satisfy criteria limited to formal higher education. Research conducted during emergency remote teaching may or may not fit a review of deliberately designed online learning.

But context should only drive exclusion when it is genuinely part of the eligibility criteria. Do not invent new restrictions during screening merely because a setting differs from what you expected to find.

The same principle applies to geographical restrictions. A study from another country is not inherently irrelevant. Whether geography matters depends on the review question and the criteria established for it.

Study Design: Is This the Kind of Evidence the Review Requires?

If your review restricts eligibility by study design, titles and abstracts may provide clues through terms such as “randomized controlled trial,” “cross-sectional survey,” “cohort study,” “qualitative interviews,” or “systematic review.”

Yet authors do not always label designs consistently. An abstract may describe methods without naming the design, or use terminology differently from the definitions in your protocol.

Look at what the researchers actually did. Who was recruited? Was an intervention assigned? Were participants followed over time? Were data collected through interviews? Is the article reporting original empirical research or synthesizing previous studies?

If you cannot confidently determine the design from the abstract and design determines eligibility, retain the record until more information is available.

Outcome Criteria Require Particular Caution

Researchers sometimes exclude a record because the abstract does not mention the outcome they are interested in. That can be risky.

Not every measured outcome appears in an abstract. Cochrane specifically cautions that outcomes may be poorly represented in titles, abstracts, and database indexing. It also states that eligible studies should not be excluded solely because measured outcome data are not reported in a usable way.

There is an important distinction here. If your eligibility criteria legitimately require that a particular outcome was measured, evidence that the study did not measure that outcome may establish ineligibility. But an abstract that simply does not mention the outcome does not necessarily tell you whether it was measured.

Do Not Perform Full Critical Appraisal During Initial Screening

During title and abstract screening, you may notice a small sample, an unusual analytic method, or a design you personally consider weak. Unless these features violate predetermined eligibility criteria, they do not automatically justify exclusion.

Eligibility and methodological appraisal are conceptually different. A study can be relevant while still having methodological limitations, and the appropriate response to those limitations depends on the review methodology.

Trying to perform a miniature critical appraisal during abstract screening also encourages inconsistent decisions because abstracts rarely contain enough methodological detail for a defensible assessment.

Read the Abstract for Evidence, Not for Keywords

Keyword matching helped retrieve the record. Screening asks a different question.

Consider an abstract containing all of your search terms but stating that the study examined instructors' intentions to use generative AI rather than students' experiences of receiving AI feedback. Keyword overlap is high; substantive eligibility is low.

Conversely, an eligible paper may use terminology you did not expect. “Large language model-generated formative comments” could describe the same phenomenon your review calls “generative AI feedback.”

Screen concepts, not strings of words.

Missing Information Should Usually Produce Uncertainty, Not Exclusion

This is perhaps the most important habit in early screening.

Cochrane recommends that reviewers generally be over-inclusive when examining titles and abstracts. The reason is asymmetric risk. Keeping an irrelevant record for full-text assessment costs additional time. Incorrectly excluding a relevant record can remove evidence from the review altogether.

If an abstract genuinely does not provide enough information, the next question is not how to guess more accurately. It is what to do when the abstract is insufficient for a defensible selection decision.

Use a Consistent Order When It Helps You Screen Efficiently

You do not necessarily need to inspect every criterion in every abstract once a clear exclusion has been established. If your review requires university students and an abstract explicitly studies infants, examining every outcome and methodological detail adds little.

A screening form can therefore place easily identified, highly discriminating criteria earlier. The exact sequence depends on the review.

This is an efficiency device, not a universal hierarchy. Your review question determines which criteria matter and in what order they can sensibly be assessed.

Screening Rules Should Be Calibrated Before the Main Work Begins

Eligibility criteria can look perfectly clear on paper until two reviewers apply them to actual abstracts and interpret them differently.

Pilot screening a sample of records can reveal ambiguous definitions before those ambiguities affect hundreds of decisions. One reviewer may interpret “higher education” as universities only, while another includes vocational colleges. One may treat blended learning as online learning; another may not.

These disagreements are useful signals. They show where operational definitions need clarification.

For Cochrane intervention reviews, at least two people working independently are required for final full-text eligibility decisions, while duplicate title and abstract screening is desirable but not mandatory. Requirements differ across review methodologies, so the screening procedure should follow the methodological guidance applicable to the review rather than assuming one arrangement fits every evidence synthesis.

How Would You Screen These Five Abstracts?

Record D illustrates the most important screening judgment. You do not yet know that it is eligible. You also do not know that it is ineligible. The correct early-stage decision is therefore to preserve the uncertainty and obtain more information.

Common Errors During Title and Abstract Screening

Ask Whether You Have Enough Information to Exclude, Not Whether You Have Enough to Include

A useful mental shift can make title and abstract screening both faster and safer.

Instead of asking, “Has this abstract proved that the study belongs in my review?” ask, “Has this record given me enough information to conclude that it does not meet my criteria?”

That framing reflects the provisional nature of early screening. Full eligibility can be established later from more complete information.

Once a record survives this stage, the question becomes whether and when you need to retrieve the full text. At that point, you can apply criteria that cannot be evaluated reliably from bibliographic information alone.

For structured reviews, use the criteria consistently rather than making them stricter as the pile grows. Fatigue has yet to be recognized as an eligibility criterion.

What to Check While Screening Each Title and Abstract

Questions About Screening Titles and Abstracts

Look for Enough Evidence to Make the Screening Decision, and Do Not Guess the Rest

How to Cite This Guide

This guide is intended to be read, shared, and used in research, teaching, and academic work. If you draw on its ideas, explanations, or other content, please acknowledge the source by citing the guide. Doing so gives appropriate credit and helps your readers locate the original resource.