Comment: Which Ifs Have Causal Answers

By Donald B. Rubin. 1986

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Counterfactual reasoning and learning systems: the example of computational advertising (2012, background): “d. A.5.2 Notes for section 3 Despite burning philosophical debates, the manipulability theory of causation (e.g., von Wright, 1971; Woodward, 2005) has gained acceptance in the statistical community. Rubin (1986) argues that there is “no causation without manipulation”. Viewing causation as a reason”
Optimal a priori balance in the design of controlled experiments (2013, background): “OPTIMAL A PRIORI BALANCE 41 [14] Donald B Rubin.”
Cross-Sectional versus Longitudinal Survey Research: Concepts, Findings, and Guidelines (2008, background): “Conversely, an improperly timed longitudinal survey 3Although some scholars equate causality with experimental manipulation (e.g., Holland 1986; Rubin 1986), many social scientists regard this view as overly restrictive (Berk 1988; Biddle, Slavings, and Anderson 1985; Cook and Campbell 1979;…”
Evaluating Kindergarten Retention Policy (2006, background): “assigned and what treatments are received by other experimental units ( Rubin 1986 ).”
Marginal Mean Models for Dynamic Regimes (2001, methodology): “This statement implicitly assumes that there is no need to index Y by others’ treatment or by the mechanism by which the treatment was selected; this is the stable unit treatment assumption [SUTVA; see Rubin (1986)].”
Assessing the Performance of Matching Algorithms When Selection into Treatment is Strong (2007, background): “This requirement is often referred to as stable unit treatment value assumption (sutva, see Rubin 1986).”
Early-Grade Retention and Children’s Reading and Math Learning in Elementary Years (2007, background): “…under the stable-unit-treatment-value assumption (SUTVA) that there is a single value of each potential outcome associated with each treatment for each experimental unit, regardless of how the treatments are assigned and what treatments are received by other experimental units (Rubin, 1986).”
Optimal dynamic treatment regimes (2003, background): “Furthermore, we assume that a subject’s outcomes are not influenced by other subjects’ treatments so we index each subject’s potential outcomes by only his or her treatments (see Cox (1958) and Rubin (1986) for a more complete discussion).”
Understanding and Using Mediators and Moderators (2008, background): “This paper primarily takes a manipulationist view in which a cause is manipulated to show if the occurrence of the effect pattern is consistent with an experimenter’s hypothesis (Holland 1986; Rubin 1974, 1986; Shadish et al. 2002).”
Detecting selection bias, using propensity score matching, and estimating treatment effects: an application to the private returns to a master’s degree (2007, background): “Recent extensions by Rubin (1986, 1990, 1991) of the counterfactual framework to observational data suggest that causality can be inferred by positing that even though individuals can only be observed in one state, individuals have potential outcomes in both the treatment and untreated state.”
Ratio of Mediator Probability Weighting for Estimating Natural Direct and Indirect Effects (2025, methodology): “Adopting Rubin’s Causal Model (Holland, 1986, 1988; Rubin, 1978) and invoking the stable unit treatment value assumption (SUTVA) for simplicity (Rubin, 1986), I use 𝑍𝑎 to denote a child’s potential mediator and 𝑌𝑎𝑍𝑎 for the child’s potential final outcome.”
Causation, prediction, and search (1993, background): “Rubin's (1977) claims about unbiased estimates of differential effects from alternative manipulations and Pratt and Schlaifer's rules all follow immediately from the assumption that the interventions in question are special cases of "manipulations.”
Functional grouping and cortical-subcortical interactions in emotion: A meta-analysis of neuroimaging studies (2008, unknown): “Importantly, this mediation is not a test of directional (i.e., causal) association; we believe that making directional inferences with confidence in this case requires manipulation of variables (Rubin, 1986).”
Randomization Inference in a Group–Randomized Trial of Treatments for Depression (2008, background): “2.4) said there is no interference between units if “the observation on one unit [is] unaffected by the particular assignment of treatments to other units”; Rubin (1986) called this the “stable unit-treatment value assumption.””