```xml <result> <column_name>job_description_length</column_name> <family>additive</family> <subtype>intensive_additive</subtype> <confidence>0.85</confidence> <requires_user_confirmation>false</requires_user_confirmation> <sum_meaningful>false</sum_meaningful> <signed>false</signed> <reducers> <primary_reducer>mean</primary_reducer> <valid_reducers>mean, min, max, median, distribution</valid_reducers> <excluded_reducers>sum (summing lengths across entities is meaningless), product (not a ratio)</excluded_reducers> <average_semantics>arithmetic_mean</average_semantics> </reducers> <inner_outer_measure> <inner_measure_reducer>min</inner_measure_reducer> <outer_measure_reducer>max</outer_measure_reducer> </inner_outer_measure> <multiplicative_partner_derivable>false</multiplicative_partner_derivable> <reasoning> DIMENSION table with no sequence axis strongly priors against multiplicative. The column name "job_description_length" indicates a measured property (character or word count) of each entity. Values range [500, 2499] with no zeros or negatives, centered far from both 0 and 1, ruling out ratio_multiplicative. This is an intensive property: each job posting has ONE length; summing lengths across multiple job postings is meaningless (you don't get "3000 characters of job description" by combining two postings). The natural rollup is mean (average length across a set of postings). Classified as intensive_additive with high confidence. </reasoning> <table_type_conflict>false</table_type_conflict> </result> ```