The objective of this master thesis research project is to determine the overall polarisation sensitivity of the Near Infrared Spectograph (NIRSpec) on board the JamesWebb Space Telescope (JWST). Understanding the polarisation sensitivity is important to correctly interpret the output data and accurately calibrate the instrument during operational lifetime. As experimental research on this topic for NIRSpec was limited, a numerical code has been built to simulate the light following the optical path of NIRSpec. This numerical code is effectively a ray tracer with polarisation state analysis functions. They ray tracer is designed to be modular, making it straightforward to add new functions or optical elements, either for future work on NIRSpec or application to other instruments. The analyses regarding the polarisation sensitivity were performed in three main categories: the induced polarisation through reflections by mirrors, the grating efficiencies and overall throughput, and the effect on the calibration process. The JWST is a large space telescope that aims to help better understand every time period of the universe from the Big Bang till the present. The launch is currently scheduled in March 2021. The JWST shall provide new data by observing thermal radiation from various celestial objects. It houses four main science instruments, one of which is NIRSpec developed by the European Space Agency. The optical design of NIRSpec is based on three three-mirror anastigmats and six main opto-mechanical assemblies. Two of those assemblies are the grating wheel assembly which houses dispersive reflection gratings and the calibration assemblywhich is used to internally calibrate the instrument. These optical elements introduce a polarisation sensitivity, meaning that the instrument output data will depend on the polarisation state of the incoming radiation. Even though NIRSpec will mainly observe unpolarised light sources, a degree of polarisation will always be induced via reflections by themirrors. The optical path of the calibration beam differs from the nominal science path, so a different degree of polarisation will be induced via the calibration path compared to the nominal science path. Application of the calibration measurements to the science measurements can thus lead to errors on the deduced signals. The induced degree of polarisation for the nominal science path will likely remain below 2%. This degree of polarisation is slightly dependent on both wavelength and initial incidence angle, but not significantly. The analysis of the grating efficiencies clearly indicates a polarisation sensitivity up to 20% for the medium resolution gratings and up to 40% for the high resolution gratings. Due to the low estimated induced polarisation below 2%, the maximumuncertainty in the output data for unpolarised incoming light would be lower than 1%. Observing celestial objects with an inherent degree of polarisation, however, could introduce a significant uncertainty in the measurements. Finally, the calibration process has also been examined and it can be concluded that the induced degree of polarisation by the calibrationmirrors before the light reaches the grating will be 1 to 2% higher compared the nominal science path. Whether or not this difference should be taken into account will depend on the application. Recommendations for future work can be divided into four categories. First, the reflection grating efficiencies should be analysed in more details as the accuracy of the presented results is uncertain. Second, the filters and detector are modelled as single layer Fresnel surfaces, while in reality they consist of multiple layers of different materials. It should be researched how this assumption affects the presented results. Third, configurations using the long-slit spectroscopy and integral field spectroscopy have not been analysed while the microshutter assembly is modelled as ideal transmitter. Analysing these configurations would provide a more complete understanding ofNIRSpec’s polarisation sensitivity. Fourth, validation through measurement data would help determine the accuracy of the ray tracer presented in this master thesis research. This data could come from eithermeasurements of NIRSpec during operational lifetime or comparable instruments of which measurement data is available.