Another way is to compare the first char with the last, if they are the same compare the second char with the second to last char, if they are the same then continue. This is better because the function can return false as soon as failure is discovered, rather than the expense of reversing the whole string when in the worst case the first char is wrong.
To see how this is better, imagine you had to write a program that output all the palindromes in an entire dictionary. There are potentially millions of words, and most of them are not palindromes, so it is easy to see how this algorithm would execute much more efficiently.