I admit that I have significant holes in knowledge regading the "Middle Ages in the Mediterranean" region of space-time. I am biased towards the NW region of Europe partly because there's a wealth of material accessible to me pertaining to the area, partly because I derive most of my income from providing language services for the language spoken in the region, and therefore I am expected to have more than just a passing familiarity with this region's culture and history (although I am partially excused from having to know it to any significant depth as I specialize mostly in technical and engineering texts where this knowledge isn't all that important, and history is merely my part-time hobby). I haven't seen a general work dedicated to the Romance-speaking and Mediterranean regions either in English or in my native tongue yet, therefore I plead incompetent on this charge. :-)

I might, however, try to dig up some data pertaining to my native, i.e., Central European environment; however, not on such a short notice for it to be relevant to this discussion.

EDIT: I wonder about the tradeoffs involved here. Tentative hypothesis: in Northern climates, the labor of daughters was more valuable, so fathers kept them around for longer before marrying them off.

My take on this - I believe that some of the factors involved were:

- the necessity for young people to help their families work on the household,

- the necessity for young women to earn their dowry, if their families couldn't provide it for them (usually by entering into service of nobles or townsmen for a few years),

- the stagnation of land development outside periods of borderland colonization (different periods in various countries, in Central Europe, this period of colonization would have taken place between 11th and 13th century), meaning that someone had to retire first for another pair of young people to get their household.

I'm quite sure there would have been other reasons that I can't recall right now.

Of course, here you are:

"Historians have demonstrated that peasants married at significantly later ages than aristocrats. Whereas members of the nobility usually married between age 14 and age 20, peasants probably married in their mid- to late twenties. It is likely that new families could not form until the parents of the potential couple were either dead or old enough to retire, so that they could turn over their land and dwelling to the new couple." [Linda Mitchell, Family Life in the Middle Ages, Greenwood 2007, p. 40]

"Reliable statistics are rare for the Middle Ages, but the evidence suggests that the typical age of marriage among the peasantry was the late twenties for men, the early twenties for women; similar ages probably applied for the urban laboring classes. The ages decline, particularly for women, further up the social scale. Among prosperous urban families, the marriage age for women was typically the late teens; the age varied widely for men. In the aristocracy, ages in the mid-teens for women and early twenties for men seem to have been common." [Jeffrey L. Forgeng, Daily Life in Medieval Europe, Greenwood 1999, p. 27]

[The following text is concerned with the very beginning of Early Modern period-post Late Middle Ages] "When individual ages are looked at, however, we do find very occasional marriages in the early teens. One girl gave her age as thirteen, none as fourteen, four as fifteen, twelve as sixteen, but all the rest of the brides in the sample, 990 of them, were seventeen or over, and more than four out of five had reached the age of twenty. Only ten of the men were younger than this. The commonest age for women was twenty-two, for men twenty-four; the median - the age below which as many got married as above it - was some 22.75 for women, 25.5 for men." [Peter Laslett, The World We Have Lost: Further Explored, Routledge 2001, p. 83]