研究者Matthew Killingsworth基于170万条实时幸福报告数据发布的最新研究表明，收入与幸福之间存在极其紧密的线性关系[1]。
这项研究涵盖33,391名美国就业成年人的幸福数据[1]。研究采用了多种幸福测量方式，包括"你现在感觉如何？"这一-50到+50连续量表、生活满意度量表以及人生满意度5项量表[1]。
研究结果显示，幸福与对数收入的组群相关系数为0.98-0.99，收入分位数与幸福的相关系数则达到0.96-0.99[1]。这一关联在年收入从$10,000到超过$500,000的范围内均成立[1]。值得注意的是，比例性收入差异（如10%加薪）与幸福的关联在所有收入水平上保持一致[1]。
研究发现揭示了一个关键的数学现象：虽然每增加一美元对幸福的边际价值呈指数递减，但真实收入本身在分布中呈指数增长，两者相互抵消[1]。这导致在个人收入决策中幸福的增长保持相对恒定[1]。研究因此得出结论，在个人决策中收入与幸福呈线性关系，但在不同人群间的权衡中呈指数关系[1]。
A new study analyzing the relationship between income and happiness reveals a remarkably consistent connection across all earnings ranges. Based on approximately 1.7 million real-time happiness reports collected from 33,391 employed American adults, researchers found that happiness correlates with logarithmic income at an exceptionally high rate, with correlation coefficients of r = 0.98–0.99 [1]. This relationship holds across a wide income spectrum, from $10,000 annually to over $500,000 per year [1].
The research measures happiness through multiple methods, including responses to "How do you feel right now?" on a continuous scale from -50 to +50, life satisfaction assessments, and a five-item life satisfaction scale [1]. When comparing income percentiles to happiness levels, the correlation remains equally strong, ranging from r = 0.96–0.99, demonstrating a markedly linear pattern [1].
A key finding challenges conventional wisdom about diminishing returns on income. While the marginal value of each additional dollar decreases exponentially, real-world income itself is distributed exponentially across the population—these two opposing forces effectively cancel each other out [1]. As a result, proportional income differences, such as a 10 percent raise, show the same relationship to happiness gains at all income levels [1].
The research distinguishes between two different contexts: in personal financial decisions, income and happiness maintain a linear relationship, whereas when considering trade-offs between different groups in society, the relationship becomes exponential [1].