English COLLEGEVERSE RESEARCH · 49 ARTICLES
Data & College Choices
Understand the population, year and definition behind a number before using it to compare colleges.
Newest publications first · Page 4 of 5
A Five-Point Gain Can Also Be a Twenty-Five-Percent Increase
Translate an invented college outcome change into percentage points, relative change and counts without mixing their meanings.
Unit reconciliationWho Did Not Reply? Put a Graduate Survey Rate in Context
Distinguish survey participation from an outcome among respondents, calculate transparent bounds for an invented eligible population, and identify the limits of that exercise.
Nonresponse boundsBetter in Each Group, Lower Overall: Check the Student Mix
Reproduce a fictional reversal in college completion rates, then apply the same explicit group weights without presenting a standardized illustration as a causal ranking.
Weighted-rate experimentWhy You Cannot Average Course Salary Medians into a University Median
Use two invented salary distributions to show why subgroup medians and group sizes cannot determine the combined median, and identify what evidence a pooled claim needs.
Distribution experimentDid the Figure Change, or Did the Data Release Change?
Build a four-cell vintage bridge to separate a revision to a historical observation from a change between reporting periods, with a reproducible audit trail.
Vintage bridgeWhen One Student Moves the Percentage: Read the Denominator
Read percentages alongside counts, distinguish percentage points from relative growth, and avoid overinterpreting changes in small education groups.
Worked arithmeticWhat a Graduate Salary Median Can and Cannot Tell You
Compare salary populations and definitions, explore two invented distributions with the same median, and keep historical graduate outcomes separate from personal offers.
Distribution exampleRetention and Graduation: Read Two Different Clocks
Separate first-year retention from longer-term graduation, avoid deriving student pathways from incompatible cohorts, and turn historical outcomes into useful questions.
Outcome comparisonMissing, Zero and Estimated: Three Different Research Conclusions
Preserve evidence states when reading education tables, see how replacing a missing value with zero distorts an average, and retain source flags and coverage.
Data interpretationLatest Available Is Not Always the Same Year
Align reference periods, entry cohorts and release versions before comparing education data, and distinguish historical observations from current application evidence.
Timeline methodA Bigger Program Is Not Automatically a Better Fit
Use historical award counts to investigate scale without inventing class sizes, elective access or teaching quality. Build a comparison around your own priorities.
Evidence checklistWhat a Graduation Rate Changes About Your Budget Questions
Use four-year and six-year graduation figures to ask better questions about degree timing, with an explicit extra-term scenario rather than a personal forecast.
Scenario & checklist