<p>While research and communication on climate change are heavily relying on presenting data in the form of line graphs, research is needed to test whether withholding the specific numbers and presenting the evidence graphically might come at the cost of lower confidence or accuracy. We compared how accurately, quickly, and confidently participants could identify trends in climate data presented as graphs vs. as tables. In Experiment 1 (<i>N</i> = 210), adult participants were to classify whether graphed or numerical climate change data showed an increasing or decreasing trend in short time series. The results suggest that the choice of format does not involve a trade-off. According to the effect sizes, the advantage of graphs over tables was largest in terms of speed but also pronounced in confidence. Speed and confidence were also superior for data graphs as compared to tables in Experiment 2 (<i>N</i> = 224), where participants faced the same stimuli but were to estimate a data value in the future. Number-based estimates were closer to the presented values, while graph-based estimates were closer to the linear trend. Across different tasks, climate data turned out to be preferable for presenting exact numbers.</p>