Ecological studies are aligned along a realism–precision continuum ranging from field observations to controlled lab experiments that each have their own strengths and limitations. Ecological insight may be most robust when combining approaches. In field observations along a successional gradient, we found correlations between plant species composition and soil bacterial communities, while bacterial Shannon diversity was unrelated to vegetation characteristics. To add a causal understanding of the processes of bacterial community assembly, we designed lab experiments to specifically test the influence of plant composition on bacterial communities. Using soil and seeds from our field site, we added different combinations of surface‐sterilised seeds to homogenised soil samples in microcosms and analysed bacterial communities 4 months later. Our results confirmed the field observations suggesting that experimental plant community composition shaped bacterial community composition, while Shannon diversity was unaffected. These results reflect intimate plant–bacteria interactions that are important drivers of plant health and community assembly. While this study provided insights into the role of plants underlying the assembly of bacterial communities, we did not experimentally manipulate other drivers of community assembly such as abiotic factors. Therefore, we recommend multi‐factorial laboratory experiments to quantify the relative importance of different factors contributing to microbial composition.