Christenfeld Nicholas J S, Sloan Richard P, Carroll Douglas, Greenland Sander
Department of Psychology, University of California, San Diego, La Jolla, CA 92093-0109, USA.
Psychosom Med. 2004 Nov-Dec;66(6):868-75. doi: 10.1097/01.psy.0000140008.70959.41.
When experimental designs are premature, impractical, or impossible, researchers must rely on statistical methods to adjust for potentially confounding effects. Such procedures, however, are quite fallible. We examine several errors that often follow the use of statistical adjustment. The first is inferring a factor is causal because it predicts an outcome even after "statistical control" for other factors. This inference is fallacious when (as usual) such control involves removing the linear contribution of imperfectly measured variables, or when some confounders remain unmeasured. The converse fallacy is inferring a factor is not causally important because its association with the outcome is attenuated or eliminated by the inclusion of covariates in the adjustment process. This attenuation may only reflect that the covariates treated as confounders are actually mediators (intermediates) and critical to the causal chain from the study factor to the study outcome. Other problems arise due to mismeasurement of the study factor or outcome, or because these study variables are only proxies for underlying constructs. Statistical adjustment serves a useful function, but it cannot transform observational studies into natural experiments, and involves far more subjective judgment than many users realize.
当实验设计尚不成熟、不切实际或无法实施时，研究人员必须依靠统计方法来调整潜在的混杂效应。然而，这些方法相当容易出错。我们研究了使用统计调整后经常出现的几个错误。第一个错误是，即使在对其他因素进行“统计控制”之后，一个因素能预测结果，就推断它具有因果关系。当（通常情况）这种控制涉及去除测量不完美变量的线性贡献时，或者当一些混杂因素仍未被测量时，这种推断就是错误的。相反的错误是，因为在调整过程中纳入协变量后，某个因素与结果的关联减弱或消除，就推断该因素在因果关系上不重要。这种减弱可能仅仅反映出被视为混杂因素的协变量实际上是中介变量（中间变量），并且对于从研究因素到研究结果的因果链至关重要。由于研究因素或结果测量错误，或者因为这些研究变量只是潜在结构的替代指标，还会出现其他问题。统计调整有其有用的功能，但它不能将观察性研究转化为自然实验，而且涉及的主观判断比许多使用者意识到的要多得多。
