A smaller message does in fact take less time to decrypt than a longer one, however, I think you meant crack, as in recover the plaintext of a message without the encryption key, and the answer is basically no. This is because good encryption makes any plaintext derived from a given ciphertext as plausible as any other.
One advantage of a small message is that it is easier to encrypt it by creating a random key that is the same size as the message. Matching the key and message lengths creates what's called a one-time pad. This uses character-by-character encryption, or stream ciphering. This type of cryptosystem is unbreakable, if used correctly. Let me give you an example. Suppose you encrypt your eight-digit bank account number. The length of this message may give an attacker a clue as to its content, but they won't know if they've cracked the message, since any of the 10^8 permutations of an eight-digit number could be correct. This is true for text-based messages as well. If I encrypted the message "Defuse bomb mission off," which is 23 characters long, with a one-time pad, an attacker wouldn't be able to determine whether "Detonate bombs at three" -- also 23 characters -- was the message I sent.
The problem with one-time pads is that you have to generate a new random key each time you send a message. This means creating, delivering, and securing large keys, which is very complex. Imagine if you wanted to encrypt a 2Mb computer file with a one-time pad. You would need a key that is also 2Mb or 2^20 characters long. This limits true one-time pad systems to very specific uses so other encryption systems are more generally used. In block ciphers for example, the key length is much less than the length of the message, but the plaintext message is broken up into small pieces called blocks, and the key encrypts each block. A variety of block ciphers operate in different ways, but as a rule, the smaller the key the less secure any message encrypted with that key will be. Therefore, you need to ensure that you use an adequate length of key and that your key or keys are adequately protected and genuinely random.
This was first published in September 2005