We should typologise the thought experiments we employ as tests for theories, i.e., formulate abstract case-types of which our specific thought experiments are case-tokens: When proposing a case-type, we make a hypothesis about which of the case’s structural features guided our intuitive evaluation. If we present both a thought experiment and our suggestion what case-type we think it’s a token of, we open up means for assessing whether our intuition about the case should have evidential weight in the process of theory-testing. An intuition about a thought experiment should only have evidential weight if it is directly replicable, i.e., if an exact repetition of the experiment yields the same result. But even if this is confirmed, we cannot be sure the intuition is accurate. The thought experiment might be deceptively formulated, misleading people’s intuitions. We should therefore use the hypothesised case-type as a blueprint for formulating additional case-tokens, and then test whether these evoke the same intuition. If they do, we have a conceptual replication, lending some support for using our intuitions about tokens of this case-type as evidence. If they don’t, further research is needed: Either our original thought experiment is deceptively formulated or our hypothesised case-type is wrong.