I am converting an old database from MS Access to FM Pro, and in the old system we numerically coded all categorical data, such as religious affiliation (1=Catholic, 2=Protestant, etc.), ethnicity, kinship relation, place-names, etc. As I recall, back in the day this was to facilitate exporting data to the statistical program we were using (SPSS), which can now handle text-based categories better than it used to. I am trying to decide whether there is any benefit to storing these categories as numerical codes (with related tables that translate codes to text-based categories), or whether it would be better practice to just simply store the categories in the tables as text (brother, sister, etc. as opposed to 6, 7, 101, etc.). I have some Filemaker development experience, but I am no pro. The central purpose to this solution is storing survey data and related multimedia files collected with surveyed households for social science data analysis (including exporting data and related demographics to data analysis software packages).
In sum, Is there any reason to store categorical data numerically rather than as text, such as for places (geographical locations) or kinship relationships (brother, sister, paternal grandmother, etc.)?