a formula used to obtain estimates for count or binary data that show more dispersion than the statistical model can handle. An advantage of quasi-likelihood models is that they permit an increased flexibility in the data types and research situations to which they may be applied. A limitation is that the function does not derive from a known population distribution.