The first database developed was MMAYEARS, a reconstructed membership list for the MMA. The records contained lists for a few individual years and an incomplete list of member companies and when they joined. These sources were extended by a systematic examination of all records (particularly the Executive Committee Minutes and Presidents' and Secretaries' reports) for all company name occurrences to create a comprehensive listing of all members, 1904–40 (285 in number). The company is the record unit and the basic data units (“fields”) are the years 1904–40, so that for every company, in every year, one can tell its status, if any – joining the MMA, leaving, in membership or not. From these data information about accession, resignation, and turnover rates was calculated, and the identification of the MMA's stable membership core became possible.

The first logical extension to MMAYEARS resulted from a need to know how member companies compared with and related to the whole population of metal-working firms in the district. This was done by creating the databases 1902FACT (692 company records), from the Philadelphia returns in the 1902 state factory inspector's report – the last one to give individual company details (name, industry, address, male, female, youth, and total employment); and PID-1640 (2,333 records), from the state industrial directories, which began publication in 1913 and became reliable and fairly consistent in 1916. Between 1916 and 1922 directories contained information about the gender breakdown of the laborforce, as well as dividing it between production and clerical/administrative categories.