The chronic unpredictable stress model is a rodent model of stress-induced anhedonia. The sucrose preference test, often used to validate it, is unreliable. Intracranial self-stimulation offers an alternative and is often cited as supporting evidence of the model's validity. Our aim was to assess whether an increased self-stimulation threshold is found after stress and if such a change correlates with decreases in sweet consumption. We searched PubMed, Embase (ovid), and Web of Science for studies in rats exposed to chronic unpredictable stress that employed intracranial self-stimulation. Thresholds for stressed and control animals were pooled from 11 studies that collectively reported on 23 different experiments. Over 50 % of the data was contributed by one research group, so a three-level meta-analytical random effects model was fit to account for methodological differences between different networks of researchers. After this adjustment, we did not find that the self-stimulation thresholds were increased in stressed rats. Pioneering experiments with positive results failed to be replicated by others, although no specific factor could be pointed to as a likely explanation. What is more, the available evidence suggests a lack of connection between sweet preference and self-stimulation, although this relationship has been seldom investigated. No study reported correlation coefficients. Methods known to mitigate biases were frequently absent, as was a transparent report of crucial study details. Our findings challenge the claim made in support of the validity of the model. Further efforts would be well-invested in assessing how reliably other tests of anhedonia have found the effects of the chronic unpredictable stress model.