On the explicit construction of higher deformations of partition statistics (Q939643)

The principal objects of study in this paper are the two \(q\)-series \[ R_2(q) := \prod_{k \geq 1}\frac{1}{(1-q^k)}\sum_{n \in \mathbb{Z} \backslash \{0\}} \frac{(-1)^nq^{n(3n+1)/2}}{(1-q^n)^2} \] and \[ R_2^o(q) := \prod_{k \geq 1}\frac{1}{(1-q^{2k})}\sum_{n \in \mathbb{Z}} \frac{(-1)^nq^{n(3n+5)+2}}{(1-q^{2n+1})^3}. \] These series are generating functions for certain rank moments and Durfee symbols which play an important role in recent work of \textit{G. E. Andrews} [Invent. Math. 169, No. 1, 37--73 (2007; Zbl 1214.11116)]. The author shows that up to a linear combination of quasimodular forms, \(q^{-1}R_2(q^{24})\) is the holomorphic part of a weight \(3/2\) weak Maass form. For this reason, she calls \(R_2(q)\) a ``quasimock theta function'' (ordinary mock theta functions being essentially holomorphic parts of weak Maass forms). As applications, an asymptotic formula for the coefficients of \(R_2(q)\) as well as congruences for these coefficients (and related combinatorial functions) are derived. As for \(R_2^o(q)\), the author computes the transformation properties and derives an asymptotic formula for the coefficients, but states that the ``transformation law for \(R_2^o(q)\) \dots is more complicated than the transformation law for \(R_2(q)\) arising from higher order poles of the generating function. For this reason no usual weak Maass forms arise.''