A score looks simple because it compresses information into a number. Reading it well means doing the opposite: expanding the number back into questions about its definition, source, and limits. DonorAPI.com preserves three numeric score fields from the uploaded Charity Navigator dataset. This guide explains how to use those fields together, why their age matters, and how to compare records without turning a convenient sorting tool into a claim the underlying data cannot support.
Know which of the three scores you are reading
The source labels the overall score as score, the financial score as fscore, and the accountability-and-transparency score as ascore. All three are described as being out of 100. These are different fields, not alternative names for the same quantity. Every profile presents them separately so that a reader can distinguish a general historical result from the two component labels supplied with it. The field dictionary preserves those definitions alongside the original field names.
Do not assume the overall score is the arithmetic average of the other two. The supplied metadata does not provide a formula that would justify recalculating it that way. The website therefore displays the recorded result rather than replacing it with a newly derived figure. This is an important rule for any data directory: a transformation that seems intuitive can change the meaning of a source. When the derivation is unknown, preserve the value and explain the limit instead of silently inventing a method.
Separate the record's age from the page's date
The dataset description says it was scraped in May 2019 and mostly contains rating details from 2017. It does not give a distinct rating date for every row. Accordingly, a profile can explain the collection's historical context, but it cannot assign an exact assessment year to an individual organization where none was supplied. Treating every record as a 2017 rating would be more specific than the source allows. Treating it as a current rating would be even less defensible.
Blog publication dates are separate again: they date the editorial guide, not the financial observations embedded in the directory. Likewise, a newly generated webpage is not evidence that the charity was newly reviewed. Keeping those ideas distinct avoids a common presentation problem in data products, where a recent page date accidentally gives old numbers a fresh appearance. When you save a score in your own research notes, include the dataset context rather than only the day on which you opened the page.
Use score bands as navigation, not awards
The score directory groups overall scores into 90–100, 80–under-90, 70–under-80, and below 70. These bands are site navigation choices. They are not new Charity Navigator designations and do not create a separate certification. Their purpose is to make thousands of records easier to browse without asking visitors to run software or download a source file. The underlying decimal score remains visible, so the grouping does not replace the actual observation.
Boundaries deserve particular care. A score of 89.99 belongs below 90 even though a rounded whole-number display might look like 90. This website groups using the recorded numeric value, not a rounded label. Two organizations separated by a boundary may be much closer numerically than two within the same band. For comparisons, read the displayed two-decimal score and the other available fields. A band is a route into the data, not a sufficient reason to favor one organization over another.
Compare within a relevant group
A comparison becomes more interpretable when the organizations share a meaningful context. Start with a cause or subcategory that reflects the work you are interested in, then consider whether their recorded expense scale differs substantially. This does not eliminate every difference, but it makes you less likely to compare unlike activities solely because they happen to have numeric scores. A museum, a relief organization, and a research institute can all appear in the same directory without being interchangeable alternatives.
Location may also help organize a research question, provided you remember that the recorded state is not a complete service-area field. You might investigate organizations associated with one place, then read their mission descriptions to understand the work actually described. The score provides one additional observation in that comparison. It does not establish which population benefits, how many people are reached, or whether the intervention suits your priorities. Those questions require evidence outside the three score columns.
Understand what an average does and does not say
This dataset's overall-score mean is approximately 86.87, while its median is 88.31. These are calculations across the 8,408 supplied rows, not national statistics for every nonprofit. The mean adds all scores and divides by the number of records. The median identifies the middle of the ordered distribution. Their difference reflects the distribution of this particular collection; it does not, by itself, explain why the organizations received those scores or how the evaluator selected them.
On group pages, a median summarizes the records in that group, not the charities missing from the source. A state's median can change when the mix of included organizations changes even without a change in the underlying performance of any one charity. For that reason, avoid turning a group statistic into a claim that a state or cause is inherently better. Always look at the number of included records and the scope of the group before interpreting a summary value.
Do not translate historical scores into current stars
The dataset supplies numeric scores, not an authoritative current star-rating field. DonorAPI.com does not add stars, award badges, or modern methodology categories that were not provided. A colorful gauge is simply a visual representation of a historical number on a 100-point scale. Its styling must not be mistaken for a new assessment, an endorsement, or a guarantee that the organization still has the same characteristics today.
For the evaluator's own explanation of its rating approach, read Charity Navigator's methodology. That is an external reference, not a license to reinterpret the older dataset using whatever labels appear on a newer page. If you later obtain a current record, save it as a separate observation with its own date and definition. This lets you investigate changes carefully rather than overwriting the history and losing the distinction between two different pieces of evidence.
Check a small difference before interpreting it
Imagine two hypothetical organizations with overall scores of 88.20 and 88.40. The difference is two tenths of a point on the recorded scale. The numbers alone do not tell you whether that difference is meaningful for the work you want to support. You would still need to understand the source period, definitions, relevant activities, and evidence appropriate to your decision. Sorting the rows correctly is not the same thing as establishing a substantive preference.
Now consider two organizations with the same overall score but different component scores. That pattern can direct you toward different questions about their financial and accountability records, but it does not reveal the underlying reasons automatically. Preserve the three values separately and resist writing a story about the difference until additional evidence supports it. A good comparison makes the observations more visible without inventing an explanation for them.
A useful score-reading note
A strong research note might say: “The supplied historical record lists these three scores; I used them to identify financial and governance questions, and I still need a current assessment and program evidence.” This states what the numbers contributed without claiming more than they can establish. It is also easier to update when you find another source, because the historical observation and your interpretation remain separate.
Scores are most helpful when they invite disciplined inquiry. Read the field label, preserve the source context, compare relevant organizations, and resist the temptation to infer impact or current status from a single figure. The directory makes the numbers accessible; thoughtful interpretation comes from keeping the evidence attached to the exact question it can answer.
