Based on what reasoning? Data quality with respect to computer systems approaches perfection, it is far more reliable than any data you can hope to approximate in the real world. There is also plenty of "actual science" with terrible data quality, hence the replication crises spanning large sections of the scientific landscape.