If you think the data is normal-ish but want to account for skew and kurtosis, you can try fitting a distribution such as the skewed Student's t -- there are R packages for this.