Data pronunciation: "dayta" or "dahta"?

That depends on which country you live in and what your definition of "correct" is. The US and Australia, for example, predominantly use "dayta" but New Zealanders say "dahta".

In short, either is correct but different countries' cultures have different norms.

In American English, either is acceptable. "Dah-tuh" is more common than "day-tuh" in my personal experience, though it's hard to say which is more prominent overall. (Regional speech differences can inflect the decision about pronunciation as well.)

I vaguely remember being told that day-ta was the correct pronunciation, and that was because there was a vowel (the second 'a') following the consonant (the 't'). Now whether or not that actually applies, someone please comment because I'd love to know if I was given a load of hogwash there.

Personal experience, I find that I hear and use day-ta more than dah-ta. It is similar to the different ways that 'SQL' is pronounced among us techies - some spell it out, and some pronounce it as "sequel". Thinking about it, I find that when it is the word "data" all by itself, I use day-ta, but when it is part of another word ("database", for example) I tend to use dah-ta instead. Again, not sure if this is something that I just happened to have picked up over the years or if it is even correct. Anyone else notice the different pronunciation in situations like this?

The NOAD reports the pronunciation as /ˈdædə/ /ˈdeɪdə/, using the American English IPA; using the British English IPA, the pronunciation is /ˈdeɪtə/.

The difference between /ˈdeɪdə/ and /ˈdeɪtə/ is the same difference between /ˈɪdəli/ and /ˈɪtəli/ (the pronunciation for Italy).

There's no such thing as "correct" pronunciation.

Now, to answer your question, here's what LPD3 says on this (Wells 2008):

Preference polls

BrE:

ˈdeɪtə 92%

ˈdɑːtə 6%

ˈdætə 2%

AmE:

ˈdeɪțə 64%

ˈdæțə 35%

ˈdɑːțə 1%

NB: ț stands for the (voiced) alveolar tap (flap) here. Wells uses a slightly different symbol, not the usual IPA one.

A historical perspective:

The eleventh edition of Everyman's Pronouncing Dictionary by Daniel Jones (the 1960 reprint) gives ˈdeitə as the primary variant for British English, whereas dɑːtə is given in brackets, which means, in Jones' notation, the less frequent form that is still in current use.

In the City of London: ˈdeɪtə.

Am I the only person who thinks that the first vowel in "dah-ta" would be pronounced the same as the first syllable in "follow" rather than the first syllable in "batter"?

I don't know how to write that out phonetically, but "dah-ta", to me, would be pronounced "dahh-tahh".

This question reminds me of the film Gravity. Watch the first five minutes here. On 1:55 you can hear "Houston" (Ed Harris) saying "day-ta"; on 4:47 Dr. Stone (Sandra Bullock) says "Dah-ta", only to be answered "we are not receiving any 'day-ta'".

Now, Ed Harris was born in New Jersey and Sandra Bullock is from Virginia, but raised in Nuremberg. It seems that the correct pronunciation of "data" is a matter of geography and not so much of grammar.

Well, I know this is not the most comprehensive study, but I wanted to share an interesting case of 'day-ta' vs. 'dah-ta'.

It's DAY-ta on Star Trek, which is perhaps the ultimate reference. :-) I cannot imagine Patrick Stewart saying "Dah-ta."

P. S. An editorial reminder (as mentioned earlier) that data is plural; e.g., "the data are clear" not "the data is clear." Feel free to edit this answer for references to data being either plural or singular in usage. Let's keep it humorous, if possible.