Previous research suggests that people implicitly believe that biological and nonbiological natural entities exist to fulfil certain functions (i.e., people hold implicit teleological beliefs). The standard experimental paradigm used to demonstrate this is to compare rates of teleological acceptance in an un-speeded condition to acceptance in a speeded condition. As speeded decision-making limits the opportunity to engage in reflective thought, increased rates of teleological acceptance relative to the un-speeded condition are said to provide evidence of implicit teleological beliefs. Across two large online studies, we show that due to the exclusion criteria typically used in this paradigm, the included and excluded participants vary systemically in important ways between conditions, and that increased rates of teleological acceptance during speeded responding does not provide evidence of implicit teleological beliefs. Rather, the difference between conditions can be explained by increased acceptance of explanations which are objectively false. Furthermore, we show that a key assumption underpinning the use of this paradigm – that accepting teleological explanations should be effortless and rejecting them should require effort – is not supported by the data. These results highlight not only a methodological issue, but also a theoretical issue in the current literature. We discuss the implications of these findings in the context of existing theoretical and empirical literature on teleological reasoning and dual-process theory more generally.