The study investigates how people comprehend high-level patterns in visualizations, comparing their intuitive understanding with the designers' intended communication goals. It explores three common visualization types—line graphs, bar graphs, and scatterplots—using a qualitative approach where participants described graphs in natural language and think-aloud protocols. The study found that participants' comprehension often did not align with the stated objectives of the visualizations, even when following best practices. This discrepancy highlights the need for more comprehensive evaluation methods that consider factors like data complexity, design elements, and individual background. The study also found that chart type alone is insufficient to predict the information people extract from a graph, and that low-level tasks may not fully capture the high-level comprehension people derive from visualizations. The results suggest that visualization effectiveness should be assessed from multiple perspectives to better inform design practices. The study emphasizes the importance of understanding how people naturally interpret visualizations and how this aligns with design intentions, providing insights into how data type, complexity, and composition influence the patterns people extract from visualizations. The findings support the need for guidelines that consider both low-level statistical tasks and high-level comprehension to optimize visualizations for effective communication.The study investigates how people comprehend high-level patterns in visualizations, comparing their intuitive understanding with the designers' intended communication goals. It explores three common visualization types—line graphs, bar graphs, and scatterplots—using a qualitative approach where participants described graphs in natural language and think-aloud protocols. The study found that participants' comprehension often did not align with the stated objectives of the visualizations, even when following best practices. This discrepancy highlights the need for more comprehensive evaluation methods that consider factors like data complexity, design elements, and individual background. The study also found that chart type alone is insufficient to predict the information people extract from a graph, and that low-level tasks may not fully capture the high-level comprehension people derive from visualizations. The results suggest that visualization effectiveness should be assessed from multiple perspectives to better inform design practices. The study emphasizes the importance of understanding how people naturally interpret visualizations and how this aligns with design intentions, providing insights into how data type, complexity, and composition influence the patterns people extract from visualizations. The findings support the need for guidelines that consider both low-level statistical tasks and high-level comprehension to optimize visualizations for effective communication.