Research Skills

Evaluating Sources & Evidence

A good source is not defined by what it says, but by what it can withstand in terms of evidence and scrutiny.

Look beyond the presentation

“Evaluation is not simply about determining whether a source appears trustworthy, but about analyzing the information provided to understand what makes it trustworthy or not.”

Association of College and Research Libraries

A source might look impressive, use professional language, and appear on a familiar domain, but that does not necessarily mean that it contains strong evidence.

Evaluating a source means looking beyond presentation and into the production process, the supporting information, and most importantly, whether that information is sufficient to support the claims being made. The Association of College and Research Libraries highlights reliability, validity, accuracy, authority, timeliness, point of view, and bias, among other factors, to consider in source evaluation (Association of College and Research Libraries, Framework for Information Literacy for Higher Education; Objectives for Information Literacy Instruction: A Model Statement for Academic Librarians).

Evaluation is not simply about determining whether a source appears trustworthy, but about analyzing the information provided to understand what makes it trustworthy or not.

Begin with what the source is claiming

“A source can be credible and still not be useful for a particular research question.”

Cornell University Library

Before looking for evidence, it is important to understand what exactly the source is saying.

A source can make a variety of claims, from simple factual statements to opinions or predictions about the future. Each type of claim requires different evidence to evaluate. A source claiming that “a survey revealed students’ preference for online learning” is not asserting the same thing as a source claiming “online learning leads to improved academic performance.” The former presents information, while the latter draws a conclusion from evidence. A researcher must be able to distinguish between the two, as the former does not necessarily support the latter.

Ask what evidence supports the claims

Once the claims have been identified, it is important to ask what evidence supports them.

Depending on the type of source, evidence can take a variety of forms. A scientific study might present statistical or quantitative data, while a social science study might analyze survey results. Government sources might draw on internal reports, while news outlets might rely on testimony from informants. Regardless of the type of source, the researcher must understand what the presented information says and whether it supports the stated claims. A source with extensive evidence can still stretch that evidence too far, drawing conclusions that go beyond the initial claims.

Look further than the abstract or conclusion

Many research papers place their most interesting and compelling information in their abstracts or conclusions, but those sections provide only a snapshot of the larger analysis. To properly evaluate a source, researchers need to look further, examining the methods and methodology to determine how the study was conducted. A researcher analyzing the effects of a behavior, for instance, must consider what populations were studied and how. A survey of 100 students at a single school can be a valuable source of evidence, but its limited scope prevents it from serving as conclusive evidence about the behavior of all students nationwide. The same is true of other research methods, including experiments and interviews. The method shapes the conclusions a researcher can draw.

Do not overlook the role of sample size

The size of a sample can affect how much information can be extrapolated from it, but quantity alone does not guarantee quality.

A very large sample can still be unrepresentative if the people included were selected in a way that does not reflect the population being studied. A smaller, carefully selected sample can sometimes provide more useful evidence than a much larger but poorly selected one. When reading a research paper or statistical report, asking questions like “Who was studied?” or “Who was surveyed?” can be critical in determining the validity of the claims. A study analyzing student behavior should reveal population information so that the reader can determine whether the claims apply to all students, some students, or a particular group that does not represent the general student body.

Watch out for correlation and causation

In research studies, correlation and causation are not always easy to differentiate.

Two events can occur together without one necessarily causing the other. For instance, a researcher analyzing student performance might find that students who used a certain resource tended to perform better on average than students who did not. The resource, however, might only be an indicator of a larger trend, such as socioeconomic status. Researchers must determine whether the evidence supports a causal explanation or only shows an association. If a source states that one event causes another without sufficient evidence, the researcher must recognize that the causal claim goes beyond the evidence provided.

Look out for limitations

Research studies are rarely perfect, as methodological decisions can affect their outcomes. Limitations highlight populations, locations, time frames, and other factors that shape how applicable a study’s findings may be. Acknowledging limitations can help readers understand how carefully a study’s findings should be interpreted. If a researcher reads a study on student behavior, for instance, its limitations may identify populations and conditions to which the findings do not apply. A school-based survey might provide valuable information, but a researcher cannot extrapolate its findings to the behavior of all students nationwide. When evaluating a source’s claims, researchers should consider its limitations to determine how widely applicable the findings are and how much weight to give the claims.

Look at the source’s references

A researcher can get additional insight from the references provided by a source.

Most research studies cite previous work, and those references can point to additional useful information. Government reports often credit the documents and databases used in their research, while news outlets may cite reporters, informants, and external documents. At minimum, a reference list can help show whether the claims of the original source can be corroborated by external evidence. Researchers should also be aware of secondary sources, which may present information that builds on or modifies information from primary sources. If an article and a press release feature the same statistics, for instance, tracing the numbers back to their original report or dataset can reveal whether both sources ultimately rely on the same evidence. Five websites repeating the same statistic do not necessarily represent five independent pieces of evidence if they all cite the same source.

Follow up with the original sources

When the details matter, a researcher should consider consulting the original or primary source behind a secondary account.

An article, for instance, might provide a brief overview of a scientific study but fail to mention crucial information such as its limitations and sample size. By consulting the original study, a researcher can verify those details and ensure that important evidence is not overlooked. This does not mean a researcher must read every source in its entirety, but they should ensure important claims have been contextualized correctly. That includes determining whether one source cites another inaccurately, deliberately or not. This practice is especially important when a source makes exceptionally strong claims about another source’s findings.

Look out for bias, but do not presume it

Sources are rarely completely free from bias, but that does not mean that all sources are designed to mislead.

A researcher must be conscious of potential biases while recognizing that they may be situational, depending on what the source is and who its intended audience is. The same statistics, for instance, can be presented differently depending on whether a source is promoting a product or criticizing it. A company’s press release will tend to present more positive information than an independent analysis, but that does not mean that the company’s claims are always false or misleading. A government source may provide valuable official information while still reflecting the purpose and institutional context in which it was produced. An advocacy group, in turn, might offer more information than a company while still having a specific agenda. Bias must be investigated rather than assumed, and context is necessary in determining whether a source is credible.

Watch out for statistics

Statistics can help strengthen a claim, but they must be carefully investigated.

A statistic without context can mask weaknesses in evidence, and a researcher must be vigilant not to overlook them. Sample size should be considered alongside how the sample was selected and what the study is trying to measure. When a source mentions percentages, a researcher must remember that percentages can be misleading depending on the context. Graphs, too, can be misleading depending on their presentation and scale. At minimum, a researcher should investigate what population a statistic refers to, what dataset it comes from, whether it reflects a certain time frame, and how it was calculated. Even then, a statistic must be weighed against context and evidence. The number of students who hold a particular opinion, for instance, can be less relevant than the reasoning behind that opinion.

Compare claims to other sources of information

A researcher must remember that no single source of information provides absolute certainty.

Whenever possible, a researcher should seek additional sources that can corroborate, refute, or otherwise provide context for the information. This is especially important when a source makes controversial, surprising, or unsubstantiated claims. Independent sources can provide stronger corroboration, especially when they rely on different evidence or methods rather than simply repeating one another. Ideally, multiple pieces of research will reach similar conclusions while using different methods of analysis.

Look out for the author’s expertise

A source’s claims can be weighed against the author’s expertise.

An author’s credentials should not be assumed to reflect reality, and a researcher must investigate whether the author’s expertise aligns with the claims being made. In practice, that means checking whether a scientist has made unsubstantiated claims or whether a professor has misunderstood the material. An institution’s authority does not guarantee that every claim it publishes is correct; the specific evidence and context still need to be examined. At the same time, expertise can be vital in evaluating a source. Context is necessary when determining whether an author’s expertise outweighs potential biases or misunderstandings.

Remember that sources can disagree

Researchers may encounter sources that contradict each other.

When this happens, contradiction does not necessarily mean that one source is wrong. The sources may have used different populations, time frames, or measurements. Researchers should examine both accounts with a degree of skepticism. If one source concludes that something is true and another concludes that it is not, they should investigate what evidence supports each source. Newer sources, for instance, may incorporate additional research not considered by older sources, while other sources may present stronger evidence. If neither source is stronger, a researcher may need to investigate further. The goal is not simply to accept the claims made by the stronger source, but to understand why the sources contradict each other.

Filipino students and source evaluation

Source evaluation is an important skill, but that does not mean that students are naturally adept at it.

The findings of the 2023 study on Filipino students illustrate why source evaluation skills must be taught. The study involved 424 students from 13 educational institutions in the Bicol Region, representing senior high school, undergraduate, and postgraduate levels, and found that most participants demonstrated “beginning” competency in author- and fact-checking tasks and “emerging” competency in image-checking tasks (Fajardo). The findings suggest a need to strengthen information and media literacy skills among the students represented in the study.

“beginning” competency

Fajardo

Other studies of Filipino students have reported different findings. A 2024 study of 99 College of Teacher Education students at Quirino State University, for instance, found that participants demonstrated proficiency in searching, assessing, and using sources (Ramos & Abella).

This illustrates the importance of distinguishing between findings that apply to one population of students and those that apply to all students. Both studies are useful, but neither applies universally because each analyzes a specific population. A single finding cannot represent an entire population without considering the contexts in which it was generated.

Evaluate sources one step at a time

A single test cannot determine whether a source is credible.

Instead, source evaluation is a process involving a series of questions that can help a researcher determine a source’s strengths and weaknesses. These questions might include: Who created the information? What is the source claiming? What evidence does it provide? How was the evidence gathered? Is the methodology appropriate for the claims being made? What populations or subjects were studied? What limitations did the researcher acknowledge? Can the evidence be traced to other sources? Are there independent sources that can corroborate the information?

Each question matters, and they are most useful when considered together. Source evaluation is not about identifying a source’s weaknesses or finding reasons it might be incorrect. It is about recognizing the evidence and understanding what it can and cannot say. This is especially true when evaluating high-stakes claims. Good research, in other words, does not begin or end with finding sources that reinforce a researcher’s existing beliefs. It means asking what the evidence supports, what it does not support, and why.