Method

Each case starts with a main legal question. The home page and case summaries focus on that question; side questions cover narrower sub-issues inside the same case.

• Only public, source-verifiable opinions can be counted.
• Each source is stored, when possible, with a title, publisher, date, link, excerpt, and speaker.
• Sources may include news reports, official notices, bar-association statements, court documents, or other relevant public records.

• A legal view in a source is linked to a case, a question, and an answer choice.
• Individual lawyer opinions, city bar-association opinions, and the Union of Turkish Bar Associations opinion are tracked as separate components.
• Views from party counsel, defense or applicant lawyers, party representatives, or legal actors with a direct institutional or political stake may remain in the source record, but they receive no weight in the consensus score.

• The score is a weighted distribution normalized to 100 points.
• Three source groups are weighted separately: individual lawyer opinions can contribute up to 50%, city bar associations up to 40%, and the Union of Turkish Bar Associations up to 10%.
• The individual group reaches its full weight at 100 distinct lawyer opinions; the city bar-association group reaches its full weight at 20 distinct city bar-association opinions.
• A group that has reached full weight can receive a share of otherwise unused points. If there is no TBB opinion, the TBB share is 0.

Side questions show a score only when there are at least 100 individual lawyer opinions or at least 50 city bar-association opinions. Otherwise the page says Not enough info so a thin early record is not presented as a reliable result.

Percentages do not represent all lawyers. They represent the distribution of opinions Hukukça has been able to source, classify, and count at that moment. New sources or corrections can change the numbers.

Readers can open sources, suggest missing sources, and report records that seem wrong. They can also run the verification prompts on question detail pages in their own web-enabled AI tools to retry the source search and classification logic. Corrections are evaluated against the source and the linked opinion record.