Supplementary Materials for the Correction to: WHOIS Data Redaction and Its Impact on Unsolicited Emails: A Field Experiment

Supplementary materials for the correction to T. Sattler, "WHOIS Data Redaction and Its Impact on Unsolicited Emails: A Field Experiment," IEEE Access, vol. 12, pp. 182322–182339, 2024, doi: 10.1109/ACCESS.2024.3511269. The correction reanalyzes the field experiment with the domain-specific mailbox as the unit of analysis. This deposit contains the data and the code behind the corrected results. Files: emails_788_deposit.xlsx: one row per received email (788 rows), sheet Tabellenblatt1. Columns: running number (June 2023 state), running number (April 2023 state), registrar, gTLD, sender domain, recipient mailbox, date, category (1 = transactional, 2 = registrar promotion, 3 = spam), weekday. We removed sender email addresses from this file. For spam emails (category 3), sender domains other than gmail.com, outlook.com, and hotmail.com were replaced by codes other-01 to other-43, assigned in order of first occurrence; the classes of Table 19 of the article and the count of 46 distinct sender domains remain reproducible. Sender domains of registrar emails (categories 1 and 2) are unchanged. The deposit does not include the full email text. mailboxes_66.csv: one row per mailbox (66 rows), generated by the script from the email file. Columns: registrar number (Table 2 of the article), registrar, gTLD, published (1 = registration data published, 0 = redacted), mailbox, cat1, cat2, cat3. reanalysis_domain_level.py: builds mailboxes_66.csv, validates the counts against Tables 5, 6, 9, and 14 of the article, and runs all tests reported in the correction. reanalysis_output.txt: output of the script. README.md: this description. Run: python reanalysis_domain_level.py emails_788_deposit.xlsx. Run with Python 3.12.3, pandas 3.0.2, numpy 2.4.4, scipy 1.17.1, statsmodels 0.15.0, and openpyxl 3.1.5.