Clinical trials assessing therapies for the treatment of non-alcoholic steatohepatitis (NASH) involve a baseline and end of study liver biopsy, and assessment of improvement in disease endpoints, often reflected as a percent of each treatment arm that improved, worsened or remained unchanged. Traditional preclinical rodent studies for putative NASH therapies are often limited by not knowing the level of liver disease/NASH present at the start of therapeutic intervention, instead of randomizing treatment groups on easily measurable endpoints such as body weight, metabolic status or similar. Here, we describe a liver biopsy technique in a diet-induced NASH mouse model, for the assessment of baseline liver disease in order to exclude mice that do not exhibit fibrosis and to equally distribute animals with similar fibrosis between treatment groups. These levels can then be compared to the terminal, post-intervention levels for a truer understanding of in vivo pharmacological effects and thus more accurately reflect clinical trial design strategies. The mouse is properly anesthetized and prepared for the surgery using sterile conditions. A small incision is made in the upper abdomen and the left lateral lobe of the liver is exposed. A wedge of the liver is surgically removed, and a similar-sized piece of absorbable gelatin is put in its place to stop any bleeding. The mouse is surgically sutured and stapled closed and will recover back to normal within 1 day. The entire process takes 5-10 min per mouse. Here we exemplify the utility of this procedure by leveraging the pre-study biopsy to assess the impact of the glucagon-like peptide-1 (GLP-1) receptor agonist liraglutide on NASH endpoints in mice.