Find objectives that duplicate each other

A data-quality check for ingest: which objectives say close to the same thing. Compares the words, as the overlap between two statements' token sets once stopwords are removed, within a kind.

Usage

redundant_objectives(x, threshold = 0.85)

Arguments

x

CurriculumMapping or Classification object.

threshold

Numeric [0, 1]: Token overlap at or above which a pair is reported. The default is high because the target is near-duplication, not similarity.

Value

data.frame of kind, scope, id_a, text_a, id_b, text_b and similarity, most similar first.

What this deliberately does not claim

Overlap between objectives is how curricula work. A spiral course revisits a topic at increasing depth, and a session that reinforces an earlier one is doing its job. There is no similarity above which overlap becomes "too much", because that judgement needs the instructor's intent. So this reports pairs and marks where they sit; it does not call them redundant, and report_curriculum() does not raise them as findings to act on.

scope is what makes a row worth attention:

• "course objective" – a course's outcome list is meant to enumerate distinct claims, so a near-duplicate pair there is a drafting error.

• "session <id>" – one session listing the same outcome twice is an editing error.

• "across sessions" – two sessions covering related ground. Expected. Reported so it can be seen, and normally left alone.

Author

EDG