This data set was used to test whether species specialized to high elevations or with narrow elevational ranges show more conservative (i.e. less variable) trait responses across their elevational distribution, or in response to neighbours, than species from lower elevations or with wider elevational ranges. We did so by studying intraspecific trait variation of 66 species along 40 elevational gradients in four countries (Switzerland, Australia, New Zealand, China) in both hemispheres. As an indication of potential neighbour interactions that could drive trait variation, we also analysed plant species’ height ratio, its height relative to its nearest neighbour. The following traits and parameters were measured and are available in this data set: As an indication of plant stature, we measured vegetative and generative height, where vegetative height was distance from soil to highest vegetative leaf and generative height was distance to the highest point on the reproductive shoot. As a measure of reproductive investment, we noted the presence of flowers on the randomly chosen individuals (see below). As a measure of individual and genet basal area, we measured individual plant and patch diameters, in two dimensions (along the largest diameter and perpendicular to it). In clonal plant species, plant diameter was equivalent to an individual rosette, whereas patch diameter referred to the whole genet and could represent the size of a tuft, tussock or cushion. For genera with more singular growth forms (e.g., some Gentiana species) plant and patch diameter were the same. The two diameter measurements were made at right angles, allowing estimates of patch and plant areas to be calculated as an ellipse (i.e., area = 0.5 a 0.5 b Π). All traits were measured on ten randomly selected individuals per site. Flower count data was considered in a binary fashion on a per individual basis (because for some species individuals only produce one flower when flowering) so that the presence or absence of flower(s) was a nominal value between 0 and 10 for each species at each site. We then collected at least three leaves (up to 30 for small and light leaves) from each of the first three individuals selected from each species for determination of leaf dry matter content (LDMC) and specific leaf area (SLA). For calculations of LDMC and SLA, fresh leaves were scanned on a flatbed scanner to determine leaf area. Leaves were then weighed on a balance to a precision of +/- 0.001g, prior to being air dried and reweighed with a balance to a precision of +/- 0.0001g. LDMC was calculated by dividing dry leaf mass by fresh leaf mass. SLA was calculated by dividing leaf area by dry leaf mass. Additionally, within an area of 10 cm diameter around the target individual, we determined the tallest neighbouring species and measured its vegetative and generative height, and estimated the percent cover of the target species, other vegetation, rock, and bare soil. For more details see Rixen et al. 2022, Journal of Ecology.