Models of protein folding often hypothesize that the first step is local secondary structure formation. The assumption is that unfolded polypeptide chains possess an intrinsic propensity to form these local secondary structures. On the basis of this idea, it is tempting to model the local conformational properties of unfolded proteins using well-established residue secondary structure propensities, in particular, α-helix forming propensities. We have used spectroscopic methods to investigate the conformational behavior of a host–guest series of peptides designed to model unfolded α-helices. A suitable peptide model for unfolded α-helices was determined from studies of the length dependence of the conformational properties of alanine-based peptides. The chosen host peptide possessed a small, detectable, α-helix content. Substituting various representative guest residues into the central position of the host peptide at times changed the conformational behavior dramatically, and often in ways that could not be predicted from known α-helix forming propensities. The data presented can be used to rationalize some of these propensities. However, it is clear that secondary structure propensities cannot be used to predict the local conformational properties of unfolded proteins.