On the positivity of a certain function related with the digamma function (Q6139320)

\textit{W. E. Clark} and \textit{M. E. H. Ismail} [Anal. Appl. 1, 129--140 (2003; Zbl 1042.33002)] conjectured that \[ \left( \frac{x^{n}}{1-e^{-x}}\right) ^{(n)}>0\tag{1} \] holds for all \(x\in(0,\infty)\) and \(n\in\mathbb{N}\). \textit{F. Al-Musallam} and \textit{J. Bustoz} [Ramanujan J. 11, 399--402 (2006; Zbl 1103.33002)] proved that the inequality (1) is true for all values of \(x\) in the interval \((2\log2,\infty)\) and \(n\in\mathbb{N}\). The author of the paper shows that (1) is true for all \(x\in(\log2,\infty)\), and \(n\in\mathbb{N}\). As a corollary, the author claims that for all \(\alpha\in(0,\infty)\), \(x\in(\log2,\infty)\), and \(n\in\mathbb{N}\), the following inequality holds \[ \left( \frac{x^{n+\alpha}}{1-e^{-x}}\right) ^{(n)}>0. \] We note that the result of this paper (i) is an improvement of the theorem by F. Al-Musallam and J. Busto [loc. cit.], and (ii) proves that the conjecture of W. E. Clark and M. E. H. Ismail [loc. cit.] is false.