Skip to content

Falsifying Data

Falsifying data is the act of deliberately modifying, adding, or removing measurements in order to steer the primary outcome of research toward a certain satisfactory conclusion. While this process can take many different forms, as mentioned, the main incentive is to contaminate the dataset in order to legitimize a certain conclusion or decision. [cite, cite, cite]

Here we offer a customizable algorithm to achieve a few variations of data falsification. We focused on three primary form of data falsification: perturbation, swapping, and switching data.

  • In permutation approach, the Researcher selects a set of value from a specific group, and adds a certain level of noise to each.
  • In swapping approach, the Researcher selects a set of treatment groups, or dependent variables, and swap certain number of datapoint between them.
  • In switching approach, the Researcher selects a certain number of datapoint and move them to an another group.

Every approach can be customized even further to achieve a certain behavior.

Last update: 2021-09-18