 Overdispersion is common in models of count data in ecology and evolutionary biology. It can occur due to missing covariates, non-independent, aggregated data, or an excess frequency of zeroes, zero inflation. Accounting for over dispersion is vital as failing to do so can lead to biased parameter estimates and false conclusions regarding hypotheses of interest. Observation-level random effects OLR where each data point receives a unique level of a random effect that models the extra person variation present in the data are commonly employed to cope with over dispersion in count data. However, studies investigating the efficacy of OLR as a means to deal with over dispersion are scarce. Here, simulations show that OLR yield more accurate parameter estimates compared to when over dispersion is ignored in cases where over dispersion is caused by random extra person noise or aggregation in the count data. However, OLR fail to reduce bias in zero inflated data and increase bias at high levels of over dispersion. The magnitude of over dispersion positively correlates with the degree of bias in parameter estimates. Failing to account for over dispersion in mixed models can erroneously inflate measures of explained variance R2 which may lead researchers to overestimate the predictive power of variables of interest. Therefore, use of OLR provides a simple and robust means to account for over dispersion in count data but must be applied judiciously. This article was authored by Xavier A. Harrison. We are article.tv, links in the description below.