Methodology

How this research is done, and what it can't claim.

Participant privacy and accurate representation matter more than making the site look impressive.

Participants are recruited through semi-structured interviews with students from different countries. Interviews follow a consistent guide so answers can be compared, while leaving room for a participant's own account of their situation.

Turning interviews into comparable data

Qualitative answers are coded into a standardized set of economic barriers and decision outcomes. Codes are refined early, then held stable once final analysis begins.

CodeBarrierDefinition
B01Tuition / program feesTuition is unaffordable or strongly affects choice
B02Application feesApplication fees limit number or type of applications
B03Testing / exam costsIELTS, TOEFL, SAT, entrance exams, test prep
B04Visa / immigration costsVisa fees, documentation or related financial burden
B05Travel / relocationFlights, moving, travel to campus or embassy
B06Housing / living costsRent, food, insurance, daily costs
B07Scholarship availability / uncertaintyDependence on aid or uncertainty about funding
B08Family financial pressureFamily budget, obligations, expectations
B09Need to workEmployment need constrains study choice or time
B10Currency / exchange-rate pressureLocal currency weakness or exchange-rate risk
B11Geographic accessLocation raises cost or limits access
B12Information inequalityLack of financial-aid or opportunity information
B13Opportunity costLost income or time from studying
B14Debt / loan constraintsAccess to loans or aversion to debt
B15School / resource inequalityEconomic differences in school resources or preparation
B16Technology / connectivity costDevice, internet or access costs
B17Documentation / certification costTranslations, notarization, credential evaluation
B18Program availability / affordability mismatchPreferred field unavailable at an affordable option
B19Emergency / economic instabilityInflation, job loss, instability
B20Other economic barrierUsed only when no code fits; explained in notes

Interviews are treated as an exploratory, non-random sample. Frequencies and cross-tabs always show numerator, denominator and sample size. Country comparisons are only shown where per-country n is large enough to be meaningful. The project does not claim causation from interview data, and does not generalize from the sample to an entire country or population.

DISCLOSURE

AI tools were used to assist with transcript organization, preliminary coding, data-quality checks and visualization planning. All coded data, interpretations and final research conclusions were reviewed by the researcher. AI was not used to fabricate participant data or replace human verification.

ID
Participants are identified by ID (e.g. OL-014), never by name, in any working dataset.
Consent
Quotes are published only where consent and context are appropriate.
Anonymize
Direct identifiers are removed before transcripts are shared with any AI tool.
Review
All AI-generated coding is reviewed by a human before it enters the analysis dataset.
STATED PLAINLY

This is a convenience sample gathered through direct outreach, not a random or representative sample of any country's student population. Findings describe patterns within this sample, not universal claims.