This data set presents sampling efficiencies for short-lived trace gases (dimethyl sulfide, isoprene, carbon disulfide) when using the glass plate method to sample the sea surface microlayer (SML). As the SML sample spreads in a thin layer on the glass plate, the loss of trace gas due to volatilization is supposedly high, but we hypothesized that sampling efficiency can be determined experimentally to correct for those losses. Interpretation of environmental trace gas measurements of the SML is difficult to not meaningful in the absence of a sampling efficiency estimate. By determining the sampling efficiency new light might be shed on the SML's role on air-sea gas exchange.Three experiments were performed in 2023 (A, B, and C). Experiment A and C were conducted in laboratories of Marine Biogeochemistry at GEOMAR Helmholtz Centre for Ocean Research Kiel, while experiment B was conducted under the roof of the SURF facility at ICBM Wilhelmshaven, Carl von Ossietzky Universität Oldenburg. Data from experiment A (uses mesh screen, not glass plate) is preliminary and is only available from the authors upon request. A small tank was either filled with freshwater or salt water. Every day we spiked the water with isotopically labelled trace gases: trideuterated DMS (DMS-d3), pentadeuterated isoprene (isoprene-d5), and carbon-13 carbon disulfide (13CS_2). To assess sampling efficiency in the absence of a SML reference measurement, we mixed the water prior to each sampling to ensure that SML and underlying water (ULW) had the same trace gas concentration. The difference in measured trace gas concentration reveals the losses associated with the SML sample. Sampling efficiency (E) was derived for pairs of SML and ULW samples, determined as the ratio of SML trace gas concentration (given in peak area per mL of sample, PA) over trace gas concentration (given in peak area per mL of sample) in ULW, i.e., E = PA_SML / PA_ULW.Here we present the individual SML–ULW ratios used to estimate an average sampling efficiency. Water temperature, salinity, trace gas concentration (always oversaturated), and surface activity (from artificial surfactant Triton X-100) were varied to investigate their influence. Sampling duration, pH, number of dips of the glass plate, sample volume, and SML thickness were further recorded to help untangle dependencies. Trace gas samples were measured immediately, while surface activity was measured in 2024 and 2025. Trace gas concentration was measured with a purge-and-trap gas chromatograph-mass spectrometer (GC–MS). Surface activity (SA) was measured with phase-sensitive alternating current voltammetry with a hanging mercury drop electrode.