Review Priority

Screening aid only. A "Review Priority" tier means only that an exam score is unusually high relative to logged engagement in this class. It is not evidence of academic misconduct. Engagement logs are incomplete: students study offline, share devices, use print books, or simply already know the material. Any follow-up should rely on independent evidence and your institution's procedures.

Privacy: everything runs inside this browser tab. No data is uploaded, stored, or sent anywhere, and it disappears when you close the tab.

1. Load files

Drop a file here or choose one.
Drop a file here or choose one.
Drop a file here or choose one.

2. Settings

Grade Center score columns count as assessments when their name contains one of these comma-separated words, matched as whole words in any letter case. Plurals count too, so "test" matches "Unit 2 Test" and "Tests", but not "Testing Center". Add words such as "midterm, final" if your course uses them.

The weight slider sets how much course hours vs. textbook hours count toward the engagement percentile. Students with unknown textbook time use course hours only.

3. Class diagnostics

Load an exam file and at least one engagement file.

4. Match review (textbook names to roster)

5. Review Priority results

Click a column heading to sort. Context columns (shaded) are for display only and are not part of the score.

No results yet.

6. Scatter plot: engagement percentile vs. exam percent

No data yet.

Hover over a point or focus it with Tab to see details. Hollow points are students with partial data (textbook time unknown). The dashed line is the minimum exam percent. Every value in the chart is also in the table above.

7. Students not scored

None yet.

8. Export

How this is calculated

Who is scored

Only students with a numeric score for the selected exam. "Needs Grading", "In Progress", "Exempt", "-", and blank count as no score, and those students are listed in section 7. If a student has a value in more than one modality column, that is a data conflict. The student is not scored until you choose which value to use.

Which Grade Center columns are assessments

Any column shaped like [Modality - ]Name [Total Pts: P Score] |id whose Name contains a keyword from Settings (default: exam, test, quiz). The text before the first " - " is read as the modality only if it has no digits and no keyword. "Testing Center - Exam 1" is Exam 1 taken at the Testing Center. "Unit 3 - Quiz" and "Midterm Test" are names with modality "Standard". Columns with the same name (ignoring case) are one assessment offered in different modalities. If two of those columns are ever wrongly combined, the students affected appear as data conflicts rather than being scored silently.

Steps

exam% = raw score / Total Pts × 100 courseHrs = Hours in Course textbookHrs = Total Time(seconds) / 3600 (UNKNOWN if the textbook row is unmatched or unconfirmed, which is not the same as 0) pctRank(x) = (average rank of x − 1) / (n − 1) × 100 (ties share their average rank; n = 1 gives 50) computed among the scored students who have a known value for that measure courseP = pctRank(courseHrs) textbookP = pctRank(textbookHrs) engagementP = w × courseP + (1 − w) × textbookP (w = course weight) = courseP alone, or textbookP alone, if the other one is unknown ("partial data") examP = pctRank(exam%) (across the class, or within modality if that option is on) gap = examP − engagementP (range −100 … +100) Review Priority applies only if ALL of these hold: exam% ≥ floor (85) AND examP ≥ 75 AND gap ≥ 50 High if gap ≥ 70 Moderate if 50 ≤ gap < 70 otherwise none

Why ranks?

Engagement times are highly skewed: a few students log very many hours. With 20–60 students, means and z-scores get pulled around by those outliers. Percentile ranks rely only on order, so one extreme value cannot move everyone else.

Diagnostics

Spearman's ρ (the Pearson correlation of ranks) between exam% and each engagement measure. If |ρ| < 0.2, engagement barely predicts scores in this class, so a large gap carries little information.

Context, not scored

Overall Grade, Missed Due Dates, Practices Completed, and First Attempt Accuracy are shown for judgment only. A flagged row is labeled "weakened by practice data" when First Attempt Accuracy is at least the setting (default 80%) and Practices Completed is at least the class median. That pattern is evidence of real preparation.

Name matching

Course and exam files are joined on Username, falling back to Student ID. Textbook names are normalized: lowercase, accents removed, punctuation and suffixes (Jr, Sr, II, III, IV) removed, whitespace collapsed. A name containing a comma is read as "Last, First"; otherwise it is read as "First Last". An exact match is tried first in both name orders. If there is none, a Jaro-Winkler fuzzy match ≥ 0.85 is proposed, and it is not used until you confirm it.