A scientific image records a measurement; it is not a context-free picture of reality. To read one carefully, identify what was measured, how the image was acquired and processed, and which parts of the authors’ claim the image can actually support. This guide focuses mainly on microscopy, where the details are well documented; the same checks need adapting for other kinds of scientific images.
Separate what the image shows from what the authors infer
Begin with the figure’s claim. Is it a qualitative observation, such as two structures appearing near each other, or a quantitative claim, such as a measured increase in signal? The image may illustrate a result, but a representative field alone does not establish how common the result is or supply the full statistical argument.
Keep two statements distinct: “the image shows” describes visible or encoded features; “the authors interpret this as” describes the explanation attached to them. An image can support an interpretation without proving every part of it. Sample preparation, the instrument, acquisition choices, and processing all shape what appears. The U.S. Office of Research Integrity (ORI) treats digital scientific images as data, while Harvard Medical School’s Micron guide notes that microscopy and preparation can introduce unintended attributes that may be mistaken for features of the specimen (ORI guidance on image filters; Harvard Micron guide to rigorous and reproducible microscopy).
Identify the measurement and the scale
Look for the imaging modality, sample or specimen preparation, channels, acquisition settings, and scale bar. Ask what physical signal the image encodes. In microscopy, displayed colors often represent channels or assigned values; they need not match the sample’s literal appearance. Figure annotations should explain colors, arrows, symbols, and the source of any zoomed inset, as advised by the Microscopy for Beginners presentation guide.
#1 Best Overall
Do not treat magnification, scale, and resolution as interchangeable. A scale bar conveys a known distance in the displayed image and remains useful when a figure is resized. An objective’s magnification alone omits other optics and processing. ORI states that “a scale bar of known size is the best way to express the magnification” (ORI Guideline #11).
Even a correctly scaled image does not establish that two nearby objects were resolved as separate features. Apparent size is not proof of resolution: the image’s sampling and imaging system affect which details can be distinguished.
Check whether image processing could affect the appearance
Processing can improve visibility, but it can also alter appearance or introduce artifacts. Ask whether adjustments were applied uniformly, whether filters or restoration methods were used, and whether the methods, software version, and settings are reported. ORI cautions that filters may create artifacts that look meaningful and advises comparing filtered images with the original data. Its guidance says: “If software filters must be used on scientific image data, the filters should be noted in an article’s figure legends or methods section” (ORI Guideline #7).
Enhancement and restoration are not automatically improper; the key questions are whether the process is disclosed, appropriate to the claim, and checked against the acquired data. A 2016 review explains that restoration can introduce further artifacts that affect analysis and bias conclusions (review of image degradation in microscopic images).
Judge comparisons by their acquisition and display conditions
For control-versus-treatment or before-and-after panels, check whether the compared images were acquired and processed under comparable conditions. In microscopy, differences in signal amplification, display range, filtering, or aliasing can change apparent brightness or feature size. ORI recommends identical conditions and processing for images intended for comparison (ORI Guideline #5).
| What to compare | Why it matters |
|---|---|
| Modality and measured signal | Different methods encode different physical signals; images from different modalities are not interchangeable. |
| Sample preparation and context | Preparation and specimen or material context can affect visible features. |
| Acquisition settings and calibration | Settings or instrument variation can change the recorded signal. |
| Spatial scale, sampling, and resolution | These determine the distances and details the image can support. |
| Display range, color mapping, and processing | Display and processing choices affect apparent contrast, brightness, and structure. |
| Analysis methods and data selection | Consistent methods and representative data are needed for a fair quantitative comparison. |
A mismatch does not by itself show that a comparison is invalid, but it is a reason to look for an explanation in the figure legend or methods. If the settings differ, ask whether the authors account for that difference rather than reading visual contrast as a direct measure of biological change.
Rank #4
Look for calibration and consistent quantitative methods
Claims about signal intensity or other measured quantities need more than a striking image. For microscopy intensity comparisons, ORI recommends using raw data where possible, calibrating against a known standard, applying uniform processing, and reporting the measurement procedure. Fluorescence can fade, and instruments can fluctuate, so acquisition conditions and calibration matter (ORI Guideline #9).
- Check how images or regions were sampled and whether the analysis method is described.
- Look for consistent acquisition and processing across the compared groups.
- Distinguish illustrative images from the data and statistics supporting the quantitative claim.
- For intensity claims, look for calibration and measurement from raw data where possible.
The authors of the community-developed image checklists write: “A comprehensive publication of quantitative image data should then include not only basic specimen and imaging information, but also the image-processing and analysis steps that produced the extracted data and statistics.” The article was published online on 14 September 2023 and appeared in Nature Methods volume 21 (2024) (Nature Methods checklist article).
The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Best Value
- Students build unmatched deductive-reasoning skills as they become crime-solving stars
- Most scenarios have more than one plausible outcome, allowing individuals or groups to broadly interpret evidence
- Includes interpretive handwriting, body language, fingerprinting, and many more activities
Use a discrepancy to ask for context, not to declare misconduct
An unusual feature or apparent inconsistency is a reason to seek the figure legend, methods, and—where appropriate—original data. Appearance alone generally cannot establish why a discrepancy occurred or whether anyone intended to mislead. ORI says authentication requires original data and context, and a discrepancy by itself does not establish falsification or misconduct (ORI examples and principles).
Describe what you can observe precisely: for example, that two panels appear to have different display ranges, or that a feature seems duplicated. Keep that observation separate from claims about cause or intent. The same caution applies outside microscopy, but the specific checks above are strongest for biological and light microscopy; satellite, astronomical, diagnostic, and other scientific images have modality-specific considerations.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
