Does Leading ln x Resummation Predict the Rise of g_1 at Small x ?

We numerically analyse the evolution of the flavor non-singlet $g_{1}$ structure function taking into account the all-order resummation of $α_{s} ln^{2}x$ terms which is expected to have much stronger effects than the DGLAP evolution in the small x region. We include a part of the next-to-leading logarithmic corrections coming from the resummed ``coefficient function'' which are not considered in the calculation of Blümlein and Vogt to respect the factorization scheme independence. It is pointed out that the resummed coefficient function gives unexpectedly large suppression factor over the experimentally accessible range of x and Q^{2}. This fact implies that the next-to-leading logarithmic contributions are very important for the $g_{1}$ structure function.