#### DMCA

## Mediation in experimental and nonexperimental studies: new procedures and recommendations (2002)

Venue: | PSYCHOLOGICAL METHODS |

Citations: | 630 - 3 self |

### Citations

10859 |
Statistical Power Analysis for the Behavioral Sciences (second ed.): Lawrence Erlbaum Associates
- Cohen
- 1988
(Show Context)
Citation Context ... to better performance on a test the next day. When a causal process is temporally proximal, it is likely that the effect c in Part 1 of Figure 1 will be moderate or large in magnitude. (According to =-=Cohen, 1988-=-, a medium effect for regression or correlation is around .30 in standardized units, and one needs a sample size of approximately 85 to detect the effect with 80% power when using the usual two-tailed... |

4957 | An introduction to the bootstrap
- Efron, Tibshirani
- 1994
(Show Context)
Citation Context ...rect effects in mediation models. One particularly useful approach is the bootstrap framework, which can be applied even when sample sizes are moderate or small, that is, in the range of 20–80 cases (=-=Efron & Tibshirani, 1993-=-). In a groundbreaking article, Bollen and Stine (1990) showed that bootstrap methodology could be very useful in studying the sampling variability of estimates of indirect effects in mediation models... |

2613 |
Structural Equations with Latent Variables
- Bollen
- 1989
(Show Context)
Citation Context ... they can provide evidence that one mediation pattern is more plausible than another, and they can provide invaluable information for the design of fully experimental studies of causal processes (see =-=Bollen, 1989-=-; MacKinnon, Lockwood, Hoffman, West, & Sheets, 2002). Statistical approaches to the analysis of mediation have been discussed extensively in the psychological literature (Baron & Kenny, 1986; Collins... |

1833 | Causality: Models, Reasoning, and Inference - Pearl |

770 | Asymptotic confidence intervals for indirect effects in structural equation models
- Sobel
- 1982
(Show Context)
Citation Context ...ett, 1984; Judd & Kenny, 1981; Kenny, Kashy, & Bolger, 1998; Maccorquodale & Meehl, 1948; MacKinnon & Dwyer, 1993; MacKinnon, Krull, & Lockwood, 2000; MacKinnon, Warsi, & Dwyer, 1995; Rozeboom, 1956; =-=Sobel, 1982-=-). In this article, we focus on the approach to mediation analysis that was articulated by Kenny and his colleagues (Baron & Kenny, 1986; Judd & Kenny, 1981; Kenny et al., 1998). As MacKinnon et al. (... |

608 |
A comparison of methods to test mediation and other intervening variable effects
- MacKinnon, Lockwood, et al.
- 2002
(Show Context)
Citation Context ...ide evidence that one mediation pattern is more plausible than another, and they can provide invaluable information for the design of fully experimental studies of causal processes (see Bollen, 1989; =-=MacKinnon, Lockwood, Hoffman, West, & Sheets, 2002-=-). Statistical approaches to the analysis of mediation have been discussed extensively in the psychological literature (Baron & Kenny, 1986; Collins, Graham, & Flaherty, 1998; James & Brett, 1984; Jud... |

507 |
Applied multiple regression/ Correlational analysis for the behavioral sciences. Mahwah, NJ: Lawrence Erlbaum
- Cohen, Cohen, et al.
- 2003
(Show Context)
Citation Context ...thods 5 can be used to estimate c�, a,b, and c in Figure 1. We call these estimates ĉ�, â,b ˆ , and ĉ. The standard errors of these estimates can be computed using standard methods for linear models (=-=Cohen & Cohen, 1983-=-; Draper & Smith, 1998). 6 It is well-known from the central limit theorem that the estimates of c�, a,b,and c tend to be normally distributed, and it is possible to use their estimated standard error... |

434 |
Data analysis in social psychology
- Kenny, Kashy, et al.
- 1998
(Show Context)
Citation Context ...proaches to the analysis of mediation have been discussed extensively in the psychological literature (Baron & Kenny, 1986; Collins, Graham, & Flaherty, 1998; James & Brett, 1984; Judd & Kenny, 1981; =-=Kenny, Kashy, & Bolger, 1998-=-; Maccorquodale & Meehl, 1948; MacKinnon & Dwyer, 1993; MacKinnon, Krull, & Lockwood, 2000; MacKinnon, Warsi, & Dwyer, 1995; Rozeboom, 1956; Sobel, 1982). In this article, we focus on the approach to ... |

383 |
The psychology of attitudes. Fort
- Eagly, Chaiken
- 1993
(Show Context)
Citation Context ...erimental studies. For example, social–cognitive psychologists are interested in showing that the effects of cognitive priming on attitude change are mediated by the accessibility of certain beliefs (=-=Eagly & Chaiken, 1993-=-). Developmental psychologists use longitudinal methods to study how parental unemployment can have adverse effects on child behavior through its intervening effect on quality of parenting (Conger et ... |

347 | The Earth is round (p < .05). - Cohen - 1994 |

261 | The influence of shared mental models on team process and performance - Mathieu, Heffner, et al. - 2000 |

193 | Process analysis: Estimating mediation in treatment evaluations
- Judd, Kenny
- 1981
(Show Context)
Citation Context ...002). Statistical approaches to the analysis of mediation have been discussed extensively in the psychological literature (Baron & Kenny, 1986; Collins, Graham, & Flaherty, 1998; James & Brett, 1984; =-=Judd & Kenny, 1981-=-; Kenny, Kashy, & Bolger, 1998; Maccorquodale & Meehl, 1948; MacKinnon & Dwyer, 1993; MacKinnon, Krull, & Lockwood, 2000; MacKinnon, Warsi, & Dwyer, 1995; Rozeboom, 1956; Sobel, 1982). In this article... |

190 |
The moderator- mediator variable distinction in social psychology research; conceptual, strategic, and statistical considerations. Journal of personality and social psychology
- Baron, Kenny
- 1986
(Show Context)
Citation Context ... processes (see Bollen, 1989; MacKinnon, Lockwood, Hoffman, West, & Sheets, 2002). Statistical approaches to the analysis of mediation have been discussed extensively in the psychological literature (=-=Baron & Kenny, 1986-=-; Collins, Graham, & Flaherty, 1998; James & Brett, 1984; Judd & Kenny, 1981; Kenny, Kashy, & Bolger, 1998; Maccorquodale & Meehl, 1948; MacKinnon & Dwyer, 1993; MacKinnon, Krull, & Lockwood, 2000; Ma... |

175 | Estimating mediated effects in prevention studies
- MacKinnon, Dwyer
- 1993
(Show Context)
Citation Context ...atrick E. Shrout, Department of Psychology, New York University, 6 Washington Place, Room 308, New York, New York 10003. E-mail: pat.shrout@nyu.edu Hoeksema & Jackson, 2001), and prevention research (=-=MacKinnon & Dwyer, 1993-=-). When researchers are able to show that the putative intervening variables account for the association of the distal variable with the outcome, they sometimes say that they have “explained” the asso... |

169 | Equivalence of the mediation, confounding and suppression effect
- MacKinnon, Krull, et al.
- 2000
(Show Context)
Citation Context ...ogical literature (Baron & Kenny, 1986; Collins, Graham, & Flaherty, 1998; James & Brett, 1984; Judd & Kenny, 1981; Kenny, Kashy, & Bolger, 1998; Maccorquodale & Meehl, 1948; MacKinnon & Dwyer, 1993; =-=MacKinnon, Krull, & Lockwood, 2000-=-; MacKinnon, Warsi, & Dwyer, 1995; Rozeboom, 1956; Sobel, 1982). In this article, we focus on the approach to mediation analysis that was articulated by Kenny and his colleagues (Baron & Kenny, 1986; ... |

150 |
Applied Regression Analysis, 3rd ed
- Draper, Smith
- 1988
(Show Context)
Citation Context ...o estimate c�, a,b, and c in Figure 1. We call these estimates ĉ�, â,b ˆ , and ĉ. The standard errors of these estimates can be computed using standard methods for linear models (Cohen & Cohen, 1983; =-=Draper & Smith, 1998-=-). 6 It is well-known from the central limit theorem that the estimates of c�, a,b,and c tend to be normally distributed, and it is possible to use their estimated standard errors to construct confide... |

148 | Explaining the gender difference in depressive symptoms - Nolen-Hoeksema, Larson, et al. - 1999 |

145 |
Mediators, Moderator, and Test for Mediation
- R, Brett
- 1984
(Show Context)
Citation Context ...an, West, & Sheets, 2002). Statistical approaches to the analysis of mediation have been discussed extensively in the psychological literature (Baron & Kenny, 1986; Collins, Graham, & Flaherty, 1998; =-=James & Brett, 1984-=-; Judd & Kenny, 1981; Kenny, Kashy, & Bolger, 1998; Maccorquodale & Meehl, 1948; MacKinnon & Dwyer, 1993; MacKinnon, Krull, & Lockwood, 2000; MacKinnon, Warsi, & Dwyer, 1995; Rozeboom, 1956; Sobel, 19... |

144 | The case against statistical significance testing.
- Carver, R
- 1978
(Show Context)
Citation Context ...rtant partial mediation as well as with complete mediation. The disadvantages of rigid adherence to nullhypothesis statistical tests have been extensively documented by Cohen (1994) and others (e.g., =-=Carver, 1978-=-, 1993) and are not repeated here. Instead, we simply encourage the use of the logical alternative to null-hypothesis statistical tests, the 95% confidence interval. As we discussed earlier, confidenc... |

136 |
A growth curve approach to measurement and change
- Rogosa, Brandt, et al.
- 1982
(Show Context)
Citation Context ...processes, whereby some couples are steadily increasing in conflict while others are reducing their conflict. The proper interpretation of such trajectory patterns has been discussed by Rogosa (1988; =-=Rogosa, Brandt, & Zimowski, 1982-=-). In a nutshell, when conflict at Time 2 is held constant at its mean value, then persons with unusually low Time 1 conflict will tend to have a trajectory of increasing conflict, whereas persons wit... |

135 |
How do risk factors work together? Mediators, moderators, and independent, overlapping, and proxy risk factors
- Kraemer, Stice, et al.
- 2001
(Show Context)
Citation Context ...of the mediation model. We have assumed simple linear regression models in our discussion, but more complicated nonlinear alternatives should be considered for binary variables (Collins et al., 1998; =-=Kraemer, Stice, Kazdin, Offord, & Kupfer, 2001-=-). Bootstrap methods can be used with nonlinear models as well as the linear models we have considered in our illustrations. Because bootstrap methods have many desirable features, we have argued that... |

130 |
A simulation study of mediated effect measures
- MacKinnon, Warsi, et al.
- 1995
(Show Context)
Citation Context ...86; Collins, Graham, & Flaherty, 1998; James & Brett, 1984; Judd & Kenny, 1981; Kenny, Kashy, & Bolger, 1998; Maccorquodale & Meehl, 1948; MacKinnon & Dwyer, 1993; MacKinnon, Krull, & Lockwood, 2000; =-=MacKinnon, Warsi, & Dwyer, 1995-=-; Rozeboom, 1956; Sobel, 1982). In this article, we focus on the approach to mediation analysis that was articulated by Kenny and his colleagues (Baron & Kenny, 1986; Judd & Kenny, 1981; Kenny et al.,... |

120 |
Statistics and Causal Inference (with discussion
- Holland
- 1986
(Show Context)
Citation Context ...X may be experimentally manipulated, or it might vary naturally in the population. When X is an experimental variable in a randomized study, its causal relation to Y can be interpreted unambiguously (=-=Holland, 1986-=-). When X is not manipulated, definitive causal inferences cannot be made about the relation of X to Y. Weaker inferences about the plausibility of causal relations can be made if certain assumptions ... |

103 | Invisible support and adjustment to stress
- Bolger, Zuckerman, et al.
- 2000
(Show Context)
Citation Context ...a change in X to lead to a new value of Y shortly thereafter. For example, an unobtrusive instance of social support to a stressed person should lead to some sign of reduced distress by the next day (=-=Bolger et al., 2000-=-). A cognitive prime should lead to immediate facilitation of certain memory processes. An improved lesson presentation should lead to better performance on a test the next day. When a causal process ... |

95 |
Direct and Indirect Effects: Classical and Bootstrap Estimates of Variability." Sociological Methodology
- Bollen, Stine
- 1990
(Show Context)
Citation Context ... path in Part 2 is smaller than the c path in Part 1, as indicated by dashed lines. Residual terms are displayed as d effects. statistical methods that are available for analysis of mediated effects (=-=Bollen & Stine, 1990-=-; Lockwood & MacKinnon, 1998) have implications for Kenny’s approach to statistical mediation, and we show how these new methods can be applied with relative ease to data from both experimental and no... |

84 |
Linking economic hardship to marital quality and instability
- Conger, Simons, et al.
- 1990
(Show Context)
Citation Context ...iken, 1993). Developmental psychologists use longitudinal methods to study how parental unemployment can have adverse effects on child behavior through its intervening effect on quality of parenting (=-=Conger et al., 1990-=-). Mediation analysis is also used in organizational research (e.g., Marks, Zaccaro, & Mathieu, 2000; Mathieu, Heffner, Goodwin, Salas, & CannonBowers, 2000), clinical research (e.g., NolenPatrick E. ... |

79 | Some New Results on Indirect Effects and their Standard Errors in Covariance Structure Models." Sociological Methodology - Sobel - 1986 |

76 |
Performance implications of leader briefings and team-interaction training for team adaptation to novel environments
- Marks, Zaccaro, et al.
- 2000
(Show Context)
Citation Context ...unemployment can have adverse effects on child behavior through its intervening effect on quality of parenting (Conger et al., 1990). Mediation analysis is also used in organizational research (e.g., =-=Marks, Zaccaro, & Mathieu, 2000-=-; Mathieu, Heffner, Goodwin, Salas, & CannonBowers, 2000), clinical research (e.g., NolenPatrick E. Shrout and Niall Bolger, Department of Psychology, New York University. This work was supported by N... |

66 | The Logic of Causal Order - Davis - 1985 |

56 |
On a distinction between hypothetical constructs and intervening variables
- MacCorquodale, Meehl
- 1948
(Show Context)
Citation Context ...diation have been discussed extensively in the psychological literature (Baron & Kenny, 1986; Collins, Graham, & Flaherty, 1998; James & Brett, 1984; Judd & Kenny, 1981; Kenny, Kashy, & Bolger, 1998; =-=Maccorquodale & Meehl, 1948-=-; MacKinnon & Dwyer, 1993; MacKinnon, Krull, & Lockwood, 2000; MacKinnon, Warsi, & Dwyer, 1995; Rozeboom, 1956; Sobel, 1982). In this article, we focus on the approach to mediation analysis that was a... |

48 |
Contrasts in multiple mediator models
- Mackinnon
- 2000
(Show Context)
Citation Context ...nd restated form are influential and widely cited, they are not without their critics. Several people have questioned the necessity of testing the overall association in Step 1 (Collins et al., 1998; =-=MacKinnon, 2000-=-; MacKinnon et al., 2000). Still others have pointed out that the recommendations are nested in a framework that assumes that mediation processes can be analyzed by linear regression analysis (Collins... |

46 | Myths about longitudinal research - ROGOSA - 1988 |

40 |
An alternative framework for defining mediation
- Collins, Graham, et al.
- 2000
(Show Context)
Citation Context ...n, 1989; MacKinnon, Lockwood, Hoffman, West, & Sheets, 2002). Statistical approaches to the analysis of mediation have been discussed extensively in the psychological literature (Baron & Kenny, 1986; =-=Collins, Graham, & Flaherty, 1998-=-; James & Brett, 1984; Judd & Kenny, 1981; Kenny, Kashy, & Bolger, 1998; Maccorquodale & Meehl, 1948; MacKinnon & Dwyer, 1993; MacKinnon, Krull, & Lockwood, 2000; MacKinnon, Warsi, & Dwyer, 1995; Roze... |

34 | Nonconscious behavioral confirmation processes: the self-fulfilling consequences of automatic stereotype activation - Bargh, Chen - 1997 |

34 | Mediators of the gender difference in rumination - Nolen‐Hoeksema, Jackson - 2001 |

27 |
Total, Direct, and Indirect Effects in Structural Equation Models." Sociological Methodology
- Bollen
- 1987
(Show Context)
Citation Context ...mediation patterns in the data is only subtly different from the first. This is the situation when there are several processes that taken together completely mediate the relation between X and Y (see =-=Bollen, 1987-=-; MacKinnon, 2000), but only a subset is specified explicitly in the mediation analysis. In this case, the direct effect of X on Y (c�) reflects the effects that are not captured by the mediators incl... |

25 | The robustness of estimates total indirect effects in covariance structure models estimated by maximum likelihood . Psychometrika 55(2 - Sobel - 1990 |

21 |
Bootstrap confidence intervals: Good or bad
- Efron
- 1988
(Show Context)
Citation Context ...tural equation modeling programs such as EQS (Bentler, 1997) and AMOS (Arbuckle, 1999), it has only recently begun to appear in the general psychology literature (Chan, Ho, Leung, Chan, & Yung, 1999; =-=Efron, 1988-=-; Lee & Rodgers, 1998). We hope to encourage its widespread use in mediation analyses by describing its benefits, showing how it can make a difference in the interpretation of mediation studies, and p... |

16 |
The role of the predictor variables which are independent of the criterion
- Horst
- 1941
(Show Context)
Citation Context ...the total effect may seem to disappear. This situation is considered briefly in the next section on suppression effects. 15 Suppression has had a more restricted definition in educational psychology (=-=Horst, 1941-=-) that we are not using here. The more general meaning of suppression discussed in detail by Tzelgov and Henik (1991) and used in the present article appears to be the more widely accepted one.MEDIAT... |

14 | Bootstrapping the Standard Error of the Mediated Effect
- Lockwood, MacKinnon
- 1998
(Show Context)
Citation Context ...ller than the c path in Part 1, as indicated by dashed lines. Residual terms are displayed as d effects. statistical methods that are available for analysis of mediated effects (Bollen & Stine, 1990; =-=Lockwood & MacKinnon, 1998-=-) have implications for Kenny’s approach to statistical mediation, and we show how these new methods can be applied with relative ease to data from both experimental and nonexperimental studies. Baron... |

9 |
An alternative method for evaluating congruence coefficients with Procrustes rotation: A bootstrap procedure
- Chan, Ho, et al.
- 1999
(Show Context)
Citation Context ...en incorporated as an option in structural equation modeling programs such as EQS (Bentler, 1997) and AMOS (Arbuckle, 1999), it has only recently begun to appear in the general psychology literature (=-=Chan, Ho, Leung, Chan, & Yung, 1999-=-; Efron, 1988; Lee & Rodgers, 1998). We hope to encourage its widespread use in mediation analyses by describing its benefits, showing how it can make a difference in the interpretation of mediation s... |

8 |
The product of two normally distributed random variables
- Meeker, Cornwell, et al.
- 1981
(Show Context)
Citation Context ...sts based on large-sample standard errors are too conservative. They recommended, with caveats, some methods that use information about the exact distribution of the products of two normal variables (=-=Meeker, Cornwell, & Aroian, 1981-=-). The bootstrap method can be viewed as an approximation of this exact distribution. In principle, the bootstrap methods can be generalized to mediation analyses that make use of structural equation ... |

7 |
EQS for Windows (Version 5.6) [Computer software]. Los Angeles: Multivariate Software
- Bentler
- 1997
(Show Context)
Citation Context ...rovided by Sobel (1982, 1986). Variations of this formula are used to calculate the standard error of mediated effects in many popular programs for structural equation modeling (e.g., Arbuckle, 1999; =-=Bentler, 1997-=-; Jöreskog & Sörbom, 1993). Until recently, researchers wishing to test the significance of indirect effects had little option but to use Sobel’s (1982) large-sample test. Now, however, developments i... |

7 |
Bootstrapping correlation coefficients using univariate and bivariate sampling
- Lee, Rodgers
- 1998
(Show Context)
Citation Context ...n modeling programs such as EQS (Bentler, 1997) and AMOS (Arbuckle, 1999), it has only recently begun to appear in the general psychology literature (Chan, Ho, Leung, Chan, & Yung, 1999; Efron, 1988; =-=Lee & Rodgers, 1998-=-). We hope to encourage its widespread use in mediation analyses by describing its benefits, showing how it can make a difference in the interpretation of mediation studies, and provid4 Persons famili... |

7 | Childhood antecedents of homelessness in psychiatric patients - Susser, Lin, et al. - 1991 |

6 | Upper bounds on the true coverage of bootstrap percentile type confidence intervals - Polansky - 1999 |

5 | Distribution of product tests for the mediated effect: Power and Type I error rates. Unpublished manuscript - MacKinnon, Lockwood - 2001 |

5 |
Mediation variables in scientific theory
- Rozeboom
(Show Context)
Citation Context ...1998; James & Brett, 1984; Judd & Kenny, 1981; Kenny, Kashy, & Bolger, 1998; Maccorquodale & Meehl, 1948; MacKinnon & Dwyer, 1993; MacKinnon, Krull, & Lockwood, 2000; MacKinnon, Warsi, & Dwyer, 1995; =-=Rozeboom, 1956-=-; Sobel, 1982). In this article, we focus on the approach to mediation analysis that was articulated by Kenny and his colleagues (Baron & Kenny, 1986; Judd & Kenny, 1981; Kenny et al., 1998). As MacKi... |

2 |
Spurious associations in longitudinal research
- Link, Shrout
- 1992
(Show Context)
Citation Context ...s not be considered in detail here. simply by ordering the variables in time, as there are other possible interpretations of correlation patterns from longitudinal data, including spuriousness (e.g., =-=Link & Shrout, 1992-=-). When Part 1 of Figure 1 is to be used with nonexperimental data, we need a clear theoretical rationale for the possible causal relation of X to Y, and it is the development of this rationale that o... |

1 | Excel 2000 [Computer software - Inc - 1999 |

1 | SPSS (Version 9) [Computer software - Inc - 1999 |

1 |
S-PLUS [Computer software
- StatSci
- 1999
(Show Context)
Citation Context ...nnon (1998) have shown how these programs can be written for analyses of mediation models using the SAS statistical software (SAS Institute, 1989), and Efron and Tibshirani (1993) recommended S-PLUS (=-=StatSci, 1999-=-). For the substantive researcher who is interested in mediation models, we recommend the use of the structural equation packages EQS (Bentler, 1997) and AMOS (Arbuckle, 1999), which are commonly avai... |