Check the data
Who took the exam? When? Is the information complete?
DEMOExample only. Results are made up. No real candidate data is connected.
ABGC · EXAM RESULTS STUDY
Understand the results.
Choose the next step.
This demo shows how the American Board of Genetic Counseling could study exam results and decide what support to explore for candidates.
Look for differences
How did different groups do? What needs a closer look?
Plan what comes next
Give the Board a clear summary, the limits and a useful next step.
The goal: useful questions and practical action.
A difference in pass rates does not tell us its cause. We help separate what the data shows from what still needs investigation.
See ABGC’s reports and our research
From reporting to actionable evidence
Official ABGC contextUnderstand the process
The DEIJ process assessment recommended stronger demographic information and attention to access and representation.
Read the assessment ↗Ask a broader question
The Exam Performance Taskforce considered candidates, programs and professional organizations, with several explanations still needing evidence.
Read the taskforce report ↗Connect findings to action
ABGC’s September updates call for outcome analysis, implementation plans and measurable outcomes.
Read the strategic update ↗Know what each source counts
Source review · October 16, 2026678 reported responses
August 2025 + February 2026. First-time and repeat candidates; some responses can be carried forward.
Outcome link not publishedDemographic report · pp. 3–5 ↗795 exam sittings
610 in August 2025 + 185 in February 2026. Of these, 588 are first-time sittings. Sittings are not unique respondents.
Separate aggregate sourceAugust 2025 ↗ February 2026 ↗646 reported responses
August 2024 + February 2025. Useful historical context, rather than a panel of the same people followed over time.
Different reporting windowPrior demographic report · pp. 3–5 ↗Do not add the two reports to count unique people.
795 − 678 does not measure nonresponse.
Different units, repeat attempts, carried-forward answers and item-level blanks prevent that subtraction. Public percentages below 5% also cannot be turned into exact counts.
What changed between the two published demographic snapshots?
Rounded composition only. Different reporting windows and potentially different respondents; no outcome linkage or significance test. Item-specific denominators are not published.
| Reported characteristic | Aug 2024 / Feb 2025 | Aug 2025 / Feb 2026 |
|---|---|---|
| First in family to attend college | 16% | 18% |
| Exam within one year of graduation | 91% | 84% |
| English is the language of greatest proficiency | 99% | 98% |
| Past material hardship | 9% | 9% |
These are published questionnaire summaries, not measured language proficiency or evidence that any characteristic caused an exam result. First report p. 9 ↗ · Current report p. 9 ↗
Source reconciliation queue Clarification needed
Before using a published table analytically, retain its version and ask ABGC which values control. These examples are source checks, not corrected findings.
August 2026 count/rate check
The currently linked pass-rate file prints 469 first-time passes out of 515, alongside 90.87%. Those counts calculate to 91.07%. The source alone does not establish which value should be corrected.
Historical version check
Two official report paths give different August 2023 repeat-failure counts: 15 and 39. Keep both versions and request an authoritative resolution before combining history.
These public-source exceptions are never used as inputs to the invented outcome workspace.
Place the study in the candidate journey
01TrainingPrograms and graduates
02ApplicationEligibility and access
03ExaminationPeople and attempts
04CredentialActual award and follow-up
Exam pass rates describe people who sat an exam. They cannot explain who never applied, delayed or could not reach testing.
Related data that could add context
Training pipeline · ACGC
Accreditation reports and program information can contextualize the training system. Program-level data are not a candidate-level outcome linkage.
ACGC annual reports ↗Practicing workforce · NSGC
The Professional Status Survey covers the workforce. Public executive summaries are available; full reports and research data have separate access conditions.
NSGC survey and access policy ↗BEFORE THE ANALYSIS
Start with data
we can trust.
Before studying real results, ABGC and the analyst would agree on these four checks. Try the checklist with made-up data.
Four things to check
0 / 4 reviewedThese example checkboxes do not grant access to real data or record ABGC approval.
WHY THE CHECKS MATTER
Small details can change the answer.
Was this really their first try?
Check earlier exam records before counting someone as a first-time candidate.
Was this information current at the time?
Use personal information from the agreed point in time.
Can someone belong to more than one group?
Keep all their group choices, but count the person once in the total.
Know what we are counting
Candidate
One person, with enough history to identify their first exam attempt.
Attempt
One exam sitting and its result. A person can have more than one attempt.
Personal information at that time
When the information was collected, including unanswered questions and “prefer not to answer.”
No real ABGC records or systems are connected.
TRY IT · MADE-UP RESULTS
How did the two groups do?
Start with two made-up groups. Read the result, then try another example. A and B do not represent real communities.
Change the groups, exam period or counting method
First-time candidates—
Personal information not given—Blank and “prefer not” are kept separate
Difference in pass rates—
First-attempt pass rates
Made-up resultsSee the calculation and statistical details
Intervals for individual raw first-attempt rates are descriptive 95% Wilson intervals. They are not a difference test, do not correct selection bias and do not establish causality or fairness. No interval is calculated here for overlapping-group differences, common-window rates or dependent repeat attempts.
Other questions this study could explore
Different periods may include different people
These are different made-up groups of first-time candidates. Two periods do not prove a lasting trend.
Passing is one part of access
- First-attempt outcomeIllustrated
- Applicants who did not sitNot included
- Actual credential awardNot included
- Effect of a scholarshipNot established
No scholarship, preparation resource or policy effect is inferred from these pass rates.
WHAT THE BOARD RECEIVES
The result. The limits.
The next step.
A short summary that follows the example you selected. Download it or go back and try another example.
EXAMPLE BOARD SUMMARYMade-up results · Not ABGC findings
Exam results study
QUESTION
WHAT WE SEE
DATA CHECK
WHAT WE CANNOT TELL
WHAT TO LOOK INTO NEXT
A starting point for discussion.
ABGC staff and its Diversity, Equity, Inclusion and Justice (DEIJ) Committee would review real findings before sharing them with the Board. This demo contains no real candidate information.
See possible follow-up questions
Possible follow-up questions
Ideas to discuss with ABGCCheck access to preparation resources
Separate publication, uptake, usefulness and eventual outcome measures.
Strengthen the annual data record
Retain source versions, response dates, cohort rules and any changed definitions.
Investigate an unsupported comparison
Identify the smallest new data or clarification needed before drawing a conclusion.