We studied a sample of 93 asteroid pairs, i.e., pairs of genetically related asteroids that are on highly similar heliocentric orbits. We estimated times elapsed since separation of pair members (i.e., pair age) that are between 7 × 103 yr and a few 106 yr. With photometric observations, we derived the rotation periods P1 for all the primaries (i.e., the larger members of asteroid pairs) and a sample of secondaries (the smaller pair members). We derived the absolute magnitude differences of the studied asteroid pairs that provide their mass ratios q. For a part of the studied pairs, we refined their WISE geometric albedos and collected or estimated their taxonomic classifications. For 17 asteroid pairs, we also determined their pole positions. In two pairs where we obtained the spin poles for both pair components, we saw the same sense of rotation for both components and constrained the angles between their original spin vectors at the time of their separation. We found that the primaries of 13 asteroid pairs in our sample are actually binary or triple systems, i.e., they have one or two bound, orbiting secondaries (satellites). As a by-product, we found also 3 new young asteroid clusters (each of them consisting of three known asteroids on highly similar heliocentric orbits). We compared the obtained asteroid pair data with theoretical predictions and discussed their implications. We found that 86 of the 93 studied asteroid pairs follow the trend of primary rotation period vs mass ratio that was found by Pravec et al. (2010). Of the 7 outliers, 3 appear insignificant (may be due to our uncertain or incomplete knowledge of the three pairs), but 4 are high mass ratio pairs that were unpredicted by the theory of asteroid pair formation by rotational fission. We discuss a (remotely) possible way that they could be created by rotational fission of flattened parent bodies followed by re-shaping of the formed components. The 13 asteroid pairs with binary primaries are particularly interesting systems that place important constraints on formation and evolution of asteroid pairs. We present two hypotheses for their formation: The asteroid pairs having both bound and unbound secondaries could be "failed asteroid clusters", or they could be formed by a cascade primary spin fission process. Further studies are needed to reveal which of these two hypotheses for formation of the paired binary systems is real.