Why is it so hard for King students to come up with their own original ideas? Why do they feel the need to take credit for other people’s work, and claim they have “invented” a new method?

Look at this recent PA article by another apprentice of the infamous “magician” who wanted to claim he invented maximum likelihood estimation (until the editors forced him to give proper credit to Fisher).

This guy claims that is paper “creates a multinomial framework for ideal point estimation (mIRT)”. (abstract). Except MIRT models already existed, as did at least two R packages that implement them (mirt and BayesLogit). Also, the stick-breaking and Polya-Gamma augmentation approach was already introduced in 2015 by a group of computer scientists at Harvard and MIT, and the EM algorithm that he uses is nearly identical to the EM algorithm used by Imai and coauthors in a 2016 APSR.

He does cite the relevant contributions, but he is not being honest about the fact that he basically does a copy-paste job with slight notation changes. He also claims “I use their intuition and derive results” (p. 73) . But he doesn’t derive the results himself; he just reproduces other people’s derivations.

His paper here:

The papers he borrows heavily from are here:

Linderman, Johnson, and Adams (2015)

https://arxiv.org/pdf/1506.05843.pdf

Polson, Scott, and Windle (2013)

https://arxiv.org/pdf/1205.0310.pdf

Imai, Lo, and Olmsted (2016)

https://imai.fas.harvard.edu/research/files/fastideal.pdf

Here are some examples of equations taken straight from others’ work:

Eq 1 (p. 71), Eq 3 (p. 73): Eqs 1, 2, 3 (p. 2) in Linderman, Johnson, and Adams (2015)

Eq 8 (p. 75): Eq 2 (p. 2) in Polson, Scott, and Windle (2013)

Eqs in fn 15 (p. 75): Eqs on bottom of p. 5 and middle of p. 6 in Polson, Scott, and Windle (2013)

Eq 11 (p. 76): Eqs 5 and 6 (p. 634) in Imai, Lo, and Olmsted (2016)

Eqs 12, 13, 14 (p. 76): Eqs 7 and 8 (p. 634) in Imai, Lo, and Olmsted (2016)

Had he stated the truth, that he is simply reproducing equations and derivations from other people’s work, and simply applying techniques for which there already exist canned routines, what are the odds that this would have gotten accepted?

This is also a criticism of the PA reviewers and editor: why do they allow such dishonest garbage to be published?