Estimating Mediated Effects of Personality and Social Psychological

Estimating Mediated Effects of Personality and Social Psychological Processes Patrick E. Shrout, Ph. D. NYU Niall Bolger, Ph. D. Columbia U SPSP 2010 1

An Example: Feeling Excluded • Bernstein, Sacco, Brown, Young & Claypool (JESP, 2009) Randomly assigned Ss to write about exclusion experience or another experience Ø Measured self esteem, belonging, control, & meaningful existence Ø Measured preference to 20 faces Ø • Duchenne smiles involving two muscle groups • Non-Duchenne smiles involving one voluntary muscle group • Summarized results as difference score SPSP 2010 2

Exclusion affected face perception • Those writing about exclusion were more likely to prefer “genuine” smiles to possibly staged smiles. • BUT WHY? Ø Self esteem seemed to mediate the Exclusion effect Ø “The fact that self-esteem alone fully mediated the effect warrants further discussion. Self-esteem is the mechanism by which Sociometer Theory operates (Leary et al. , 1995). In this model, self-esteem acts as a gauge of belongingness, and when a threat occurs, individuals take actions to ameliorate that threat. ” Bernstein et al, (2009) • Theory was tested using Baron & Kenny mediation model SPSP 2010 3

B&K(1986) Step 1: Find an effect to explain X c e Y M • Bernstein et al (2009) showed that Exclusion led to increased preference for natural smiles. Ø c=0. 26 SPSP 2010 4

B&K(1986) Step 2: Show X is related to mediator Y X a e. Y M e. M • Bernstein et al showed that Exclusion was related to M: Self-esteem was lower in the exclusion condition. Ø a = -. 88 SPSP 2010 5

B&K(1986) Step 3: Show M is Related to Y Y X M b e. Y e. M • ADJUSTING for X, M must be related to Y • Bernstein et al reported b= -0. 11. Ø Increased self esteem decreased interest in natural smile, adjusting for Exclusion SPSP 2010 6

B&K(1986) Step 4: Test the indirect effect Y X a M b e. Y e. M • Indirect effect is quantified by the product a*b Ø Formal test by Sobel test, joint-significance test, bootstrap confidence interval • Bernstein et al found indirect path was significant using Sobel test SPSP 2010 7

B&K (1986) Step 5: Distinguish Full from partial mediation X C’ Y e. Y M e. M • Test direct effect, c’, while adjusting for M. • The adjusted (direct) effect in Bernstein example was c’=0. 18, which was not significantly different from zero Ø Authors interpreted result as Full Mediation SPSP 2010 8

Mediation and Theory Construction • When mediation is complete, researcher has “explained the effect” Ø Other explanations apparently not needed Ø Often those other explanations not tested • Bernstein et al. (2009) did test theorydriven alternate mediators based on Williams (2007) Ø Self SPSP 2010 esteem vs. belonging, efficacy needs 9

Some Vexing Problems • Claiming complete mediation is too easy If the total effect is just significant, not much reduction is needed to make adjusted direct effect nonsignificant Ø Multiple mediators are often of theoretical interest but not usually tested Ø • Estimates of indirect effects are often biased If based on mediators that are measured with error Ø If based on wrong model Ø SPSP 2010 10

Model Specification • Baron and Kenny (1986) assume model is correct • What does this entail? Ø Causal paths are interpretable Ø Variables are measured without error Ø Residual (error) values uncorrelated • Implies that important causes are represented SPSP 2010 11

Causal Pathways and Time • Causal Assumptions in Mediation ØX is prior to M and Y Ø Change in X is associated with change in M Ø Change in M is associated with change in Y Ø Measurements taken at times that reflect causal action C’ X a M Y e. Y b e. M SPSP 2010 12

Causal Pathways and Time • Causal Assumptions in Mediation ØX is prior to M and Y Ø Change in X is associated with change in M Ø Change in M is associated with change in Y Ø Measurements taken at times that reflect causal action C’ X a M Y e. Y b e. M SPSP 2010 13

Causal Pathways and Time • Causal Assumptions in Mediation ØX is prior to M and Y Ø Change in X is associated with change in M Ø Change in M is associated with change in Y Ø Measurements taken at times that reflect causal action C’ X a M Y e. Y b e. M SPSP 2010 14

Causal Pathways and Time • Causal Assumptions in Mediation ØX is prior to M and Y Ø Change in X is associated with change in M Ø Change in M is associated with change in Y Ø Measurements taken at times that reflect causal action C’ X a M Y e. Y b e. M SPSP 2010 15

Causal Pathways and Time • Causal Assumptions in Mediation ØX is prior to M and Y Ø Change in X is associated with change in M Ø Change in M is associated with change in Y Ø Measurements taken at times that reflect causal action C’ X a M Y e. Y b e. M SPSP 2010 16

Inferring Within-Person Change from Between-Person Data • Systematic consideration of time draws us to psychological process Within person changes Ø Effects of manipulations on persons Ø • Traditional designs substitute between person differences for within person change Justified in experiments Ø Harder to justify in surveys Ø • In nature, between person associations are rarely the same as within person associations SPSP 2010 17

Revisiting Bernstein et al. (2009) • Randomized design makes temporal order clear X->Y: Exclusion experience (randomized) was related to face preference Ø X->M: Exclusion was also related to Ø • Self esteem (apparent mediator) • Efficacy needs (not found as mediator) • M->Y: Temporal relation of self-esteem and face preference not clear Ø What might contribute the correlation between M & Y? SPSP 2010 18

Possible Between Person Confounding of M->Y X 1 M 2 G Y 3 SPSP 2010 19

If “third variable” is ignored, error terms are correlated X 1 M 2 However, the correlation can not be estimated in traditional Baron & Kenny Mediation model. SPSP 2010 Y 3 20

Baseline Measures Can Reduce Confounding X 1 a M 1 Design adds within person information so that change can be estimated. c' M 2 g 1 rmy b Y 1 SPSP 2010 g 2 Y 3 21

But most ignore baseline • What are implications? Ø Total effect (c) is not biased. Ø Effect on M (a) is not biased. Ø BUT Effect of M on Y may be biased • The more stable the processes (g 1, g 2), the more the bias for nonzero correlations of M and Y. X • The more the correlation of c' baseline M and Y the more a M M the bias for stable processes. 1 1 2 g 1 rmy b Y 1 SPSP 2010 Y 3 g 2 22

Quantifying Bias: A Numerical Example X 1 Direct effect. 28 . 70 M 1 Indirect effect. 28 M 2 g 1 rmy . 40 Y 1 SPSP 2010 g 2 Y 3 23

If we ignore baseline, what do we estimate as indirect effect? g=. 8 X 1 g=. 6 g=. 4 g=. 2 g=. 0 M 1 M 2 g 1 rmy b Y 1 SPSP 2010 c' a Y 3 g 2 24

Quantifying Bias for Direct Effects g=. 0 g=. 2 g=. 4 g=. 6 g=. 8 SPSP 2010 25

Extensions • Will correlations of M and Y error terms also cause problems in cross-sectional studies? Ø You betcha! Ø The M->Y path needs to approximate within person change. Ø Additional covariates will be needed Ø But see Cole and Maxwell (2005) about plausibility of cross sectional models SPSP 2010 26

Objections • What if taking baseline measures in experiments would prime processes that are left un-primed? Ø Often possible to estimate Corr(M, Y) and the stability of M and Y in separate samples Ø Combining the data from the two samples will require structural equation methods. SPSP 2010 27

Conclusions • Social psychology theory is ready for next generation mediation analysis Ø Will aid in communication with other scientists Ø Will refine thinking about process • Combination of new heuristic steps and systematic thinking about process will serve us well SPSP 2010 28

Time for a Ten Step Program? 1) *Argue that X can be a causal agent of Y 2) Show that X is related to Y. 3) Show that X is related to M, the mediator 4) *Show that M is measured with little error. 5) *Identify plausible competing mediators and include them in the model 6) Show that M is related to Y adjusting for X 7) *Adjust for correlation between M and Y that is prior to causal process 8) Show that indirect path (X->M->Y) is present 9) Estimate/test direct effect of X->Y after adjusting for M. 10)*Report ratio of mediated effect. If it is nearly 1 then claim full mediation. SPSP 2010 29

Time for a Ten Step Program? 1) *Argue that X can be a causal agent of Y 2) Show that X is related to Y. 3) Show that X is related to M, the mediator 4) *Show that M is measured with little error. 5) *Identify plausible competing mediators and include them in the model 6) Show that M is related to Y adjusting for X 7) *Adjust for correlation between M and Y that is prior to causal process 8) Show that indirect path (X->M->Y) is present 9) Estimate/test direct effect of X->Y after adjusting for M. 10)*Report ratio of mediated effect. If it is nearly 1 then claim full mediation. SPSP 2010 30

Time for a Ten Step Program? 1) *Argue that X can be a causal agent of Y 2) Show that X is related to Y. 3) Show that X is related to M, the mediator 4) *Show that M is measured with little error. 5) *Identify plausible competing mediators and include them in the model 6) Show that M is related to Y adjusting for X 7) *Adjust for correlation between M and Y that is prior to causal process 8) Show that indirect path (X->M->Y) is present 9) Estimate/test direct effect of X->Y after adjusting for M. 10)*Report ratio of mediated effect. If it is nearly 1 then claim full mediation. SPSP 2010 31

Help from our friends • Margarita Krochik • Turu Stadler • Couples lab members at NYU and Columbia • Grant R 01 -AA 017672 from NIAAA SPSP 2010 32

References • • • Baron, R. M. , & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic and statistical considerations. Journal of Personality and Social Psychology, 51, 1173 -1182. Bernstein, M. J. et al. (2009). A preference for genuine smiles following social exclusion. Journal of Experimental Social Psychology, doi: 10. 1016/j. jesp 2009. 08. 010. Cole DA, Maxwell SE. (2003). Testing mediational models with longitudinal data: questions and tips in the use of structural equation modeling. J. Abnormal Psychology, 112: 558– 577. Gollob, H. F. & Reichardt, C. S. (1987). Taking account of time lags in causal models. Child Development, 58(1), 80 -92. Kraemer, H. , Kiernan, M. , Essex, M. , & Kupfer, D. J. (2008). How and why criteria defining moderators and mediators differ between the Baron & Kenny and Mac. Arthur approaches. Health Psychology, 27(2, Suppl), S 101 -S 108 Mac. Kinnon DP (2008). Introduction to statistical mediation analysis. New York: LEA Maxwell, S. E. , & Cole, D. A. (2007). Bias in cross-sectional analyses of longitudinal mediation. Psychological Methods, 12(1), 23 -44. Shrout, P. E. (in press). Integrating causal analysis into psychopathology research. In Causality and Psychopathology: Finding the Determinants of Disorders and their Cures. P. E. Shrout, K. Keyes, K. Ornstein (Eds). New York: Oxford U. Press. Spencer, S. J. , Zanna, M. P. & Fong, G. T. (2005). Establishing a causal chain: Why experiments are often more effective than mediational analyses in examining psychological processes. Journal of Personality and Social Psychology, 89(6), 845 -851. 2010 SPSP 33