The paper presents and tests a method for comparing real thinning with strip roads and a simulated selective thinning pattern without strip roads in the same tree stand. The simulation method includes growth models for single trees, and a computer program for carrying out different selective thinning patterns. Simulation proved to be a more reliable method than the simple comparison between the stand growth on 15 m wide zones including and excluding the strip road. The method reduces the effects of random and systematic differences, especially on the small experimental plots. The material of young and middle‐aged Norway spruce stands (19 plots) showed that the growth loss is 0–10% during the first ten years, depending on the width of the strip road and the density of the stand.