Most students, when asked to picture a research source, picture a book or an article. That image is not wrong, but it is incomplete. Evidence comes in many forms beyond the printed page, and researchers in every field — history, public health, environmental science, sociology, journalism, urban planning, and beyond — routinely build arguments from data sets, government reports, photographs, maps, audio recordings, video, and archival documents.

This chapter broadens your evidence vocabulary. You will learn what each of these source types can offer, what each cannot show, and how to evaluate them with the same rigor you would apply to a scholarly article. You will also learn why nontraditional sources — especially visual evidence — are particularly vulnerable to misuse and misinformation, and what habits protect you from being misled.

The chapter's purpose is not simply to introduce a list of formats. It is to teach you that evidence is anything that supports, challenges, or complicates a claim — and that strong researchers know how to find, evaluate, and integrate evidence in whatever form best answers their research question.

When researchers talk about evidence, they usually mean facts, statistics, quotations, or arguments drawn from written sources. This definition is correct but incomplete. Evidence is anything that supports, challenges, or complicates a claim. Evidence does not have to be a sentence. It does not have to be printed. It does not have to come from a recognized expert.

Evidence includes a table of unemployment rates across counties over ten years, a photograph of a city block before and after an urban renewal project, a recording of a labor organizer's speech from 1937, a hand-drawn map showing the boundaries of a disputed land grant, a letter from a soldier written during a military campaign, a database of hospital admissions by diagnosis and zip code, a documentary film about a mining community, a satellite image showing deforestation over a twenty-year period, a government agency's annual report on water quality, or a researcher's field notes from an ethnographic study.

None of these is a book. None is a peer-reviewed article. All of them are legitimate evidence, depending on the research question they are meant to answer.

Different source types illuminate different aspects of a topic. Written arguments and scholarly analysis are excellent for understanding how scholars interpret evidence, what debates have emerged in a field, and what conclusions the research community currently supports. But they are often less useful for reconstructing specific events, measuring real-world conditions, or understanding the lived experiences of people who did not write academic articles.

Primary documents — letters, diaries, meeting minutes, photographs, recordings — bring researchers closer to the events and people they are studying. Data sets allow researchers to measure patterns across large populations and time periods. Maps show spatial relationships that prose cannot easily describe. Audio and video recordings preserve tone, context, and texture that transcription flattens.

Strong research typically uses a mix of source types. A paper on climate change's effect on coastal communities might use scientific data sets tracking sea-level rise, government agency reports on flood damage, local news coverage of storm events, oral histories from community members, photographs documenting shoreline changes over decades, and policy documents from city planning departments.

Each type contributes something the others cannot supply. The data shows scale and trend. The reports provide institutional documentation. The news coverage shows how events were perceived in real time. The oral histories capture personal experience. The photographs make the changes visible. The policy documents show how governments responded. Researchers who limit themselves to one format miss the full picture.

Every source type requires evaluation, but different types raise different questions. A scholarly article can be evaluated using familiar tools: peer review status, journal reputation, author credentials, citation practices, and methodological clarity. But how do you evaluate a data set? How do you evaluate a photograph as evidence? How do you evaluate an archival letter or a documentary film?

Each format has its own evaluation framework. This chapter introduces those frameworks. The underlying logic is always the same: before trusting any source, ask who made it, for what purpose, under what conditions, using what methods, with what limitations, and whether the evidence it offers actually supports the claim you want to make with it.

Data is everywhere. Government agencies publish demographic surveys, health statistics, economic indicators, and environmental measurements. Research universities share data from completed studies. International organizations track global indicators. Nonprofit organizations compile data on social conditions. Commercial firms collect and sometimes release consumer and market data.

Data sets, when used correctly, allow researchers to identify patterns, measure changes over time, compare groups and regions, and test hypotheses with evidence that goes beyond individual cases.

A data set is a structured collection of information organized so that it can be analyzed. Data sets are usually organized in rows and columns. Each row represents a unit of observation — a person, a year, a county, a case — and each column represents a variable — age, income, date, outcome.

Data sets may be quantitative, meaning the values are numbers (temperatures, dollar amounts, percentages, counts); qualitative or coded, meaning the values represent categories (yes/no, demographic group, type of event); or mixed, meaning some variables are numerical and others are categorical.

Data sets are typically stored in formats such as CSV (comma-separated values), Excel spreadsheets, or specialized statistical software formats. Many publicly available data sets can be opened in Excel or Google Sheets. More complex data sets may require statistical software.

Government statistical agencies are among the most important. In the United States, the Census Bureau publishes demographic and economic data at the national, state, county, and census-tract level. The Bureau of Labor Statistics tracks employment, wages, and prices. The Centers for Disease Control and Prevention publish health and mortality data. The Environmental Protection Agency publishes data on air and water quality. Equivalent agencies exist in other countries.

International organizations publish comparative data across countries. The World Bank, the United Nations, the World Health Organization, and the Organisation for Economic Co-operation and Development (OECD) all maintain large public data repositories.

Academic data archives store data from completed research studies. Researchers often make their data publicly available so that others can replicate or extend their findings. The Inter-university Consortium for Political and Social Research (ICPSR) is one major repository. Many journals now require authors to deposit data alongside published articles.

Nonprofit and research organizations publish data on specific topics — poverty, education, public health, environment, criminal justice — and many make these data sets available for download.

Open government portals, such as data.gov in the United States, aggregate data sets from multiple federal agencies into searchable databases.

Finding a data set is only the beginning. Before using data as evidence, ask the following questions.

Who collected this data? The organization responsible for data collection matters. Government statistical agencies follow rigorous methodological standards, but they may also have reporting incentives or political pressures. Academic researchers must follow institutional review and methodological disclosure requirements. Private companies may not disclose how their data was collected or what it excludes.

What was measured, and how? Every data set reflects decisions about what to count, how to count it, and who or what to count. If you are using a data set on poverty, ask what definition of poverty was used. If you are using a data set on crime, ask what crimes are included and how they were reported.

When was this data collected? Data becomes outdated. A study of internet access from 2010 tells you very little about internet access today. Always check the data collection date and consider whether conditions may have changed.

What is the sample, and is it representative? If a data set is based on a survey or study rather than a complete census, ask who was included and whether the sample is representative of the population you care about. A survey of college students cannot tell you about the general adult population.

Are there missing values or known errors? Most real data sets have some missing data. Check for a data dictionary or methodology report that explains how missing values are handled and what limitations the data has.

Has this data been cited or used by other researchers? Data that has been used in published research and peer-reviewed studies has been subjected to some degree of scrutiny. Data that appears only on a single organization's website and has never been cited elsewhere should be approached more cautiously.

Data does not speak for itself. Every number in a data set represents a human decision about what to measure, how to measure it, and what to include and exclude. Responsible data use requires reading the methodology documentation before drawing conclusions, understanding what the variables mean and how they were defined, not overgeneralizing from one population to another, not confusing correlation with causation, and presenting uncertainty honestly.

Chapter 29 covers statistical reasoning in more detail, including how to evaluate claims made with data. The skills introduced there complement what you learn about finding and assessing data sets here.

Between the peer-reviewed scholarly article and the casual news story lies a large and important category of sources: research reports, policy briefs, white papers, and institutional publications. These sources are widely used in professional and policy contexts, and students encounter them frequently in research on current issues, public policy, business, health, and technology.

A research report is a document presenting the findings of a study, evaluation, or investigation. Research reports are produced by government agencies, universities, think tanks, foundations, advocacy organizations, and private research firms. They may be based on original data collection, analysis of existing data, literature reviews, interviews, case studies, or some combination.

A white paper is a document that presents information, analysis, and often a recommendation on a specific issue. White papers are common in policy and business contexts and are typically written for a specific audience — policymakers, executives, or advocacy partners — rather than for academic readers.

A policy brief is a short, focused document — usually two to eight pages — summarizing research findings and making recommendations for decision-makers. Policy briefs are designed to be accessible and action-oriented.

A technical report is a detailed document describing the methods, findings, and conclusions of a research project, often produced by a government contractor, research lab, or scientific agency.

Timeliness. Reports are often published quickly in response to emerging issues, long before peer-reviewed scholarship catches up. Specificity. Reports often address specific questions with detailed findings that are more targeted than broad scholarly literature. Policy relevance. Because reports are often written for practitioners and policymakers, they connect research findings to real-world decisions. Accessibility. Reports are generally written to be understood by a wider audience than peer-reviewed articles. Original data. Many reports present original data collected specifically for that study, which may not appear in any other source.

Peer review is usually absent. Unlike scholarly articles, most reports do not go through formal peer review before publication. Institutional perspective. Every report reflects the perspective and interests of the organization that produced it. A report from a tobacco company on the health risks of smoking should be evaluated differently from a report from the Centers for Disease Control. Advocacy and agenda. Many reports are produced by organizations with an explicit policy agenda. This does not necessarily make them unreliable, but it means the purpose of the report is to persuade as well as to inform. Methodological opacity. If a report makes bold claims but provides no methodology section, no data appendix, and no references, be skeptical.

When evaluating a report or white paper, apply the MGU Source Evaluation Model.

Reports and white papers are citable sources, but the citation format differs from a book or article. You typically need the author(s) or authoring organization, the title of the report, the producing organization, the year of publication, and the URL or retrieval information if accessed online.

Photographs and other images are not merely illustrations. They are sources. Like any source, they contain information, reflect the choices of whoever created them, and require evaluation before they can be used as evidence.

Images can serve as primary sources, secondary sources, or illustrative material, depending on how they are used. A photograph taken during the 1930s Dust Bowl is a primary source for a historian studying the environmental disaster and its human impact. The same photograph, reproduced in a documentary film about American agricultural history, is now part of a secondary source. Used as the cover of a textbook chapter about the Great Depression, it functions as illustrative material — it sets context rather than advancing an argument.

Researchers use images as evidence across many fields: history and cultural studies, where photographs, paintings, drawings, and prints document how people lived and what they valued; architecture and urban planning, where photographs document the built environment over time; journalism, where photographic evidence documents events and conditions; medicine and science, where microscope images, X-rays, and imaging data are primary scientific evidence; ecology, where aerial and satellite imagery documents land use and environmental change; and anthropology and sociology, where images document material culture and social settings.

Because images feel immediate and emotionally compelling, they are easy to accept uncritically. But images are constructed by human beings who made choices — about angle, framing, lighting, timing, and what to include and exclude. Every image reflects those choices.

Before using an image as evidence, ask: Who created this image, and why? A photograph taken by a government agency documenting housing conditions for a public health report was created with a different purpose than a photograph taken by a real estate developer marketing the same neighborhood. Both images may be accurate, but they are not neutral.

When was this image created? The date of an image matters enormously. Using a photograph from 1980 to illustrate conditions in 2024 may be misleading unless the caption and context make the date clear.

What does this image actually show? Identify what is visible in the image and what is absent. A photograph of a clean, well-maintained neighborhood does not prove that the entire neighborhood is clean and well-maintained — it shows one corner of it, at one moment, from one angle.

Has this image been altered? Digital editing can change an image significantly. Learn to look for signs of manipulation: unnatural shadows, mismatched lighting, blurred or distorted edges, elements that appear out of proportion. Chapter 30 covers visual verification in detail.

What does the caption say, and is it accurate? Images in news sources, social media, and even academic publications are sometimes given inaccurate or misleading captions. The same image can be used with different captions to convey completely different messages. Always trace an image to its original source before relying on any caption.

Is this image being used in its original context? A photograph of a protest rally in 2015 can be circulated in 2025 with a caption claiming it depicts a current event. This kind of decontextualization is a common form of visual misinformation. Use reverse image search tools to check an image's origin and original context.

Citing images requires recording who created them, where they were published or are held, and when they were created and accessed.

Maps are among the oldest forms of organized information. They are also among the most powerful — and the most capable of distorting what they appear to show.

Maps are used as evidence in fields including history, geography, urban planning, environmental science, public health, sociology, economics, political science, and journalism. They reveal spatial patterns that prose cannot easily describe.

Maps can show where events occurred and how they spread, how land was divided, owned, and used over time, where resources are concentrated or absent, how populations are distributed, how infrastructure and services are arranged, environmental conditions across regions and over time, and demographic patterns by location.

A historical map of a city can reveal how racial segregation was enforced through formal zoning or informal practice. A geospatial map of industrial facilities and asthma hospitalization rates can show a relationship between environmental exposure and health outcomes. A map of voting precincts layered with demographic data can illuminate patterns in electoral behavior.

Historical maps are maps created during a historical period and are used as primary sources. They document how people understood geography at that time, where they located political boundaries, and what they chose to include or omit.

Contemporary reference maps show current political boundaries, roads, cities, and physical features. They are used to establish geographic context.

Thematic maps display a specific variable across a geographic space. A choropleth map shades regions by a value (such as income or population density). A dot density map places dots to represent units. A proportional symbol map uses symbols of different sizes to show quantities.

Geospatial data layers are digital map files that can be overlaid with other data using Geographic Information Systems (GIS) software. GIS allows researchers to layer multiple variables and look for spatial correlations — for example, overlaying a map of food deserts with a map of obesity rates.

Satellite and aerial imagery provides photographic documentation of land use, vegetation, structures, and environmental conditions across time.

Maps are not objective. Every map reflects decisions about what to show and how to show it.

Projection. All flat maps distort the curved surface of the Earth. Different map projections prioritize different properties — size, shape, distance, or direction — and distort others. The Mercator projection, commonly used in navigation, distorts the relative size of landmasses, making polar regions appear much larger than they are. The Peters projection prioritizes accurate relative size but distorts shape. No projection is neutral.

Scale. Scale determines how much detail is visible and how much is omitted. A map of a city block shows details invisible on a map of an entire country.

Classification. Thematic maps that show categories require decisions about where to draw the boundaries between categories. Different classification methods — equal intervals, quantile, natural breaks — can make the same data look very different.

Bias in map-making. Who made the map, for what purpose, and for what audience? Historical maps often reflected the political interests and cultural assumptions of those who made them, including colonial boundaries imposed on existing communities, erasure of indigenous place names, and exaggeration of claimed territories.

Cite maps as you would other sources, recording the creator, title, date, and source.

Recorded audio and video are increasingly important research sources, and increasingly common as primary evidence in journalism, oral history, documentary film, legal proceedings, and scientific documentation.

Oral history recordings capture interviews with individuals who witnessed, participated in, or were affected by historical events. Many libraries and archives hold large oral history collections. The Library of Congress, the Smithsonian Institution, and universities across the country have catalogued thousands of oral history recordings.

Oral histories are valuable because they preserve voices, experiences, and perspectives that might otherwise go undocumented. They are primary sources — direct testimony from people with firsthand knowledge. But they must be used carefully. Memory is imperfect and subject to the influence of subsequent events and cultural narratives. What a person remembers decades later may be shaped by how the story has been told, retold, and publicly interpreted in the years since.

Speeches and public addresses are primary sources in history, political science, and communication studies. A recorded speech preserves not just the words but the tone, rhythm, emphasis, and context of delivery in ways that transcription cannot fully capture.

Podcasts and radio programs can be secondary sources, providing journalism, analysis, and documentary coverage of topics. They range enormously in quality, rigor, and reliability and must be evaluated by the same standards as written journalism.

Field recordings and scientific audio data include recordings of animal behavior, environmental sounds, and measurement data captured in audio form.

Documentary films can be powerful secondary sources. They often incorporate primary materials — archival footage, photographs, recordings, interviews — alongside narrative interpretation. Evaluate documentaries the same way you evaluate any secondary source: consider the filmmaker's argument, the evidence used, what is included and excluded, and what perspective the film advances.

News footage and broadcast journalism preserves recorded coverage of events as they occurred and serves as an important primary source for historians and journalists.

Archival film and home movies can document social life, community events, and private experience in ways that other sources cannot.

Scientific and research video includes recordings of laboratory experiments, animal behavior, environmental phenomena, and medical procedures, and is a primary source when the footage directly captures the phenomenon under study.

Social media video presents particular challenges. It is immediate and often unedited, which can give it the character of raw primary evidence. But it may be taken out of context, selectively released, or edited in ways that are not always apparent. Chapter 30 addresses the evaluation of video evidence and deepfakes in more detail.

Who recorded this, and why? A documentary film made by an advocacy organization has a different purpose than a documentary made by a news organization or an academic filmmaker. This does not make either invalid, but it shapes how the material should be read.

When was this recorded? The date of recording is essential context. A speech from 1960 may not represent the speaker's later views. News footage from 2005 documents conditions in 2005, not today.

Is this the complete recording, or is it edited? Audio and video can be clipped, reordered, or altered. Always try to find the original, unedited source before relying on an excerpt. Short clips circulating on social media may be taken from a longer recording with a very different overall meaning.

Has the content been altered? Audio deepfakes — recordings in which a person's voice has been synthesized or manipulated — are increasingly sophisticated. Be especially skeptical of recordings that purport to show a public figure saying something surprising or damaging, and try to find corroboration in reliable reporting.

Is there a transcript available? For scholarly use, audio and video should ideally be supported by a transcript. Transcripts make content searchable, quotable, and citable with greater precision.

Archives hold the raw material of history. They contain documents, photographs, letters, diaries, maps, account books, organizational records, government files, and other primary materials that have been collected, preserved, and made available for research. Special collections departments within libraries often hold rare books, manuscripts, ephemera, and unique local materials.

The range of materials in archives is extraordinary. A single university archive might hold personal papers of local historians, writers, educators, and public figures; records of local businesses, churches, labor unions, civic organizations, and hospitals; photographs documenting the community across more than a century; maps, blueprints, and architectural drawings; audio and video recordings of local events and oral history interviews; government records from municipal offices; and student newspapers, literary magazines, and institutional publications.

National archives, such as the National Archives and Records Administration (NARA) in the United States, hold federal government records — legislative documents, executive correspondence, military records, immigration files, treaty documents, and much more.

Specialized archives focus on particular topics or communities. There are archives dedicated to labor history, women's history, civil rights, LGBTQ+ history, immigrant communities, religious traditions, and countless other areas.

Archives are essential for certain kinds of research. If you are asking what actually happened — in a specific place, at a specific time, to specific people — primary documents in archives may be your most important sources.

Archival research requires patience. Unlike databases, archives are not fully indexed and searchable. Finding relevant materials may require browsing finding aids, asking archivists for guidance, and spending time with physical materials. But the rewards are significant. Archival sources can reveal details that no published account captures. They can correct or complicate the narratives that secondary sources present. They can give voice to people who are absent from the formal historical record.

Request access in advance. Special collections and archives often require appointments. Contact the archive before your visit to understand access procedures.

Consult the finding aid. A finding aid describes the contents of an archival collection — the creator of the collection, its scope and dates, the types of materials included, and how they are organized. Reading the finding aid before you begin saves time and helps you focus on the most relevant materials.

Ask the archivist. Archivists are specialists in their collections. They know what is available, how materials are organized, and what researchers have found useful.

Follow the archive's handling rules. Archival materials are often fragile and irreplaceable. Follow all rules about handling, photography, and reproduction.

Record citation information as you go. Archival citations must be detailed. Note the creator of the document, the document title or description, the date, the collection name, the box and folder number, and the archive's name and location. You cannot go back and reconstruct this information later.

Note what you find and what is missing. Archival research involves gaps. Not everything survives. Researchers should acknowledge what is absent from the archive as well as what is present.

Letters and diaries are among the most intimate and immediate primary sources. They capture thoughts, feelings, and observations close to the time they occurred, often before the writer knew how events would unfold. They can reveal motivations, uncertainties, and private perspectives that public documents conceal.

But letters and diaries must be used with care. They represent the perspective of a single person, writing for a specific audience (or for themselves), at a particular moment. The person may be unreliable, self-serving, or operating with incomplete information. A diary entry is not objective fact — it is one person's experience and interpretation.

Physical objects — tools, clothing, furniture, currency, weapons, decorative objects, instruments — are also primary sources. The study of material culture uses physical artifacts to understand how people lived, worked, and made meaning.

Researchers working with artifacts typically need to answer: What is this object? When and where was it made? Who made it and for what purpose? How was it used? What does it tell us about the society that produced it?

Artifacts are usually held in museum collections rather than library archives, but the research skills involved in working with them — careful observation, contextual interpretation, and awareness of what the object can and cannot show — are similar to those used with other primary sources.

The breadth of evidence available to researchers is one of the great strengths of modern research. But breadth brings complexity. Knowing how to find a data set, a map, a photograph, or an archival letter is not enough. You must also know how to use each type of source carefully, how to evaluate it honestly, and how to represent it accurately in your research.

Nontraditional sources must be integrated into research arguments with the same care as any written source. When you use a data set, do not simply cite the number. Explain where the data comes from, what it measures, what its limitations are, and why it supports — or complicates — your argument.

When you use a photograph, do not assume the image speaks for itself. Describe what the image shows, explain who created it and when, and articulate what it contributes to your argument. When you use a map, explain what the map displays, how the data was collected, and what spatial pattern you are drawing the reader's attention to. When you use an archival document, provide context: who wrote it, when, for whom, and under what circumstances.

Not every source type fits every question. Part of research judgment is knowing which type of evidence best addresses the question you are asking.

No source type is universally superior. The best source is the one that most directly and reliably addresses your research question, evaluated honestly for its strengths and limits.

Mistaking one instance for a pattern. A single photograph, letter, or recorded interview is one piece of evidence. It should not be treated as representative of a whole population or time period unless it is supported by other evidence showing the same pattern.

Ignoring methodology. Data sets and research reports are built on methodological choices. Ignoring those choices — or failing to look for them — risks misrepresenting what the evidence shows.

Accepting images at face value. Images are constructed and can be manipulated. Verify the source, date, and context of any image before using it as evidence.

Treating archival documents as objective fact. Every document was written by someone with a perspective, purpose, and potential blind spots. Read archival materials critically.

Confusing a secondary source's use of a primary source with the primary source itself. If a secondary source quotes or reproduces a primary document, try to locate the primary document itself.

Failing to record citation information. Nontraditional sources are often harder to cite and easier to lose track of than books and articles. Record full citation information as you work, including URLs, archive names, collection names, box and folder numbers, and access dates.

One challenge of nontraditional sources is that they may not be easily accessible to readers who want to verify your work. A reader can look up a journal article. But can a reader access the specific box and folder in a specific archive that you consulted? Can they verify the data set you downloaded and the version you used?

Responsible researchers make their use of nontraditional sources as transparent as possible: provide full, precise citations; describe the methodology of data sets and reports in the text; include captions with full attribution for images and maps; note access dates for online materials that may change or disappear; and, when permitted, share or link to downloaded data files so others can replicate the analysis.

When recording a nontraditional source in your research notes, use the MGU Research Note Model adapted for the source type.

Common ways visual evidence is misused: out-of-context images, where a real photograph from one event is circulated with a caption claiming it shows a different event; misleading maps, where the same data can look very different depending on category definitions, color schemes, and projection; selective data, where a chart showing only a favorable portion of a time period implies a trend the full data does not support; manipulated images, including AI-generated images sophisticated enough to be mistaken for photographs; and cropped images, where cropping can change the meaning of an image dramatically.

Different source types require different citation information. Record these details when you encounter each source type.