In the philosophy of science, observations are said to be "theory-laden" when they are affected by the theoretical presuppositions held by the investigator. The thesis of theory-ladenness is most strongly associated with the late 1950s and early 1960s work of Norwood Russell Hanson, Thomas Kuhn, and Paul Feyerabend, and was probably first put forth (at least implicitly) by Pierre Duhem about 50 years earlier.
Two forms of theory-ladenness should be kept separate: (a) The semantic form: the meaning of observational terms is partially determined by theoretical presuppositions; (b) The perceptual form: the theories held by the investigator, at a very basic cognitive level, impinge on the perceptions of the investigator. The former may be referred to as semantic and the latter as perceptual theory-ladenness.
In a book showing the theory-ladenness of psychiatric evidences, Massimiliano Aragona (Il mito dei fatti, 2009) distinguished three forms of theory-ladenness. The "weak form" was already affirmed by Popper (it is weak because he maintains the idea of a theoretical progress directed to the truth of scientific theories). The "strong" form was sustained by Kuhn and Feyerabend, with their notion of incommensurability.
However, Kuhn was a moderate relativist and maintained the Kantian view that although reality is not directly knowable, it manifests itself "resisting" to our interpretations. On the contrary, Feyerabend completely reversed the relationship between observations and theories, introducing an "extra-strong" form of theory-ladenness in which "anything goes".