Demonstrating Interdependence: Toward a General Framework

Modeling interdependence toward a general framework l.jpg
1 / 53
0
0
938 days ago, 382 views
PowerPoint PPT Presentation
Demonstrating Reliance: Toward a General System. Richard Gonzalez, U of Michigan Dale Griffin, U of English Columbia. The Settled Person. Fundamental Premises. Nonindependence gives helpful data is not an irritation

Presentation Transcript

Slide 1

Displaying Interdependence: Toward a General Framework Richard Gonzalez, U of Michigan Dale Griffin, U of British Columbia

Slide 2

The Nested Individual

Slide 3

Underlying Premises Nonindependence gives helpful data is not an irritation is a basic segment in the investigation of interpersonal conduct however may not be required in all examinations

Slide 4

Historical Analysis Explanatory need set on the gathering Meade- - individual in setting of gathering Durkheim Comte—family as essential social unit Explanatory need put on the individual Allport- - individual is essential ("prattle of tongues")

Slide 5

Necessary Conditions Homogeneity: closeness in musings, conduct or influence of communicating people E.g., assemble level, developing procedures, standards, cohesiveness Interdependence: people impacting each other E.g., performer accomplice impacts

Slide 6

McDougall, 1920, p. 23 The fundamental states of an aggregate mental activity are, then, a typical question of mental action, a typical method of feeling as to it, and some level of equal impact between the individuals from the gathering.

Slide 8

Statistical Framework Should Mimic Theoretical Framework Make ideas concrete Avoid Allport's "prattle" investigate Make the model simple to execute TODAY's Talk One time point; dyads Two or three factors Normally conveyed information; added substance models

Slide 9

Menu of Techniques Repeated measures ANOVA Intraclass connection Hierarchical direct models (HLM) Structural conditions models (SEM)

Slide 10

Common Beliefs about Interdependence in Dyadic Data If you don't right for reliance, your Type I blunders will be swelled If you don't right for association, your outcomes will be equivocal A HLM program will take out all nonindependence issues If you have dyadic information, you should run HLM (or else your paper won't be distributed)

Slide 11

These convictions miss what we accept to be the crucial issue: There is helpful mental data sneaking in the "nonindependence" Interdependence is the "very stuff" of connections.

Slide 12

Dyadic Designs: Three Major Categories Subjects settled inside gatherings Exchangeable (e.g., same sex kin) Distinguishable (e.g., diverse sex kin, mother-kid association) Mixed exch & dist (e.g., same sex & distinctive sex dyads in same outline) Univariate versus multivariate Homogeneity versus reliance

Slide 13

Intraclass Correlation: Building Block Structural Univariate Models: Exchangeable Distinguishable

Slide 14

ANOVA Intraclass (& REML) Dyads

Slide 15

Intraclass Correlation: HLM Language Two level model: Intraclass connection is given by

Slide 16

Pairwise Coding The Pearson corr of X and X' is the ML estimator of the intraclass relationship.

Slide 21

Pairwise Intraclass Correlation

Slide 23

Example: Personal Victimization Ceballo et al, 2001

Slide 25

Pairwise Intraclass (ML): Dyads

Slide 26

Interdependence how much one individual impacts another Need not be eye to eye We enjoy ourselves together, notwithstanding when we're not together (Yogi Berra)

Slide 27

Pairwise Generalization Predictor X speaks to the performer's influence on on-screen character's Y Predictor X' speaks to the accomplice's influence on-screen character's Y Predictor XX' speaks to the common influence of both on-screen character's Y

Slide 28

Example (Stinson & Ickes, 1992) ActorS = ActorV + PartnerV Strangers: an impact of the accomplice's verb recurrence on the on-screen character's giggling (in ordinal dialect, the more my accomplice talks, the more I grin/chuckle) Friends: an impact of the on-screen character's verb recurrence on the on-screen character's chuckling (the more I talk, the more I grin/snicker)

Slide 29

Some formulae Actor relapse coefficient V(Actor reg coeff) Partner coef replaces Y with Y'

Slide 30

Interdependence Example Mother and tyke witness exploitation (WV) identified with every individual's dread of wrongdoing (FC). Does youngster's WV foresee kid's FC? Does mother's WV anticipate kid's FC? and so forth

Slide 31

a X m Y m b r x r c Y c X c d

Slide 32

Not on Welfare X m Y m - .1 .2 .3 Y c X c

Slide 33

Welfare .09 X m Y m .1 Y c X c

Slide 34

V-post V-pre S-pre Simple Actor-Partner Model: Pre-post passing of mate No association issue on the reliant variable

Slide 35

Return to Original Model: Special Case a X m Y m b r x r c Y c X c d Set a=d and b=d

Slide 36

r i r i E ym E yc E xc E xm Y c Y m X m X c Y X r d

Slide 37

Latent Variable Model r i = singular level connection r d = dyad level relationship The square foundation of intraclass connections are the ways

Slide 38

Using Path Analysis Rules Two conditions in two questions; motivation behind why r xy might be uninterpretable

Slide 39

Solving those two conditions… .

Slide 40

.4 E ym E yc E xc E xm Y c Y m X m X c .47 .45 .47 .45 Y X - .8 Not on Welfare

Slide 41

E ym E yc E xc E xm Y c Y m X m X c .2 .3 .2 .3 Y X 1.6 Welfare: dormant variable model doesn't hold

Slide 42

What does the connection of two dyads implies? Along these lines, there are different parts to the relationship of dyad means making it uninterpretable… .

Slide 43

Multivariate Model: HLM Lingo Three-level model: one level for every variable, one level for individual impact, and one level for gathering impact

Slide 44

Difference scores Frequently, an issue of similitude (or consistency) comes up in dyadic research Diff of a couple pay as an indicator of spouse's relationship fulfillment Diff of husband and wife self-regard as an indicator of husband's adapting

Slide 46

Difference Scores Correlations with distinction scores can demonstrate different examples relying upon their part connections The numerator is a weighted total of the relationships: (r X1Y S X1 – r X2Y S X2 )S y Toy Examples One variable is a steady One variable is irregular

Slide 48

"Arrangements" One can utilize various relapse , entering the two factors as two indicators (instead of one contrast score). Y =  0 +  1 X1 +  2 X2 Problem: doesn't test particular speculations, for example, "comparative is better" or "self-improving is better"

Slide 49

Model-Based Approach Questions Discrepancy model (lady's sat is most prominent the more she gains, the less her significant other acquires) Similarity model (lady's sat is most prominent the littler the total diff in pay) Superiority model (lady's sat is most noteworthy when she procures more than her better half)

Slide 50

Model-Based Approach Run isolate relapses for subjects underneath or more the "balance line" (or utilize sham codes and incorporate an association term) The three unique models suggest diverse examples on the coefficients

Slide 51

Patterns of Regression Coefficients Discrepancy show: Both relapses ought to yield a negative coef for the spouse and a positive coef for the wife (boosting the distinction) Similarity demonstrate: For dyads where pay W>H, positive coef for husband and neg coef for wife in light of the fact that in this area higher husband compensation recognizes couples nearer to balance For dyads where pay W<H, neg coef for husband and pos coef for wife on the grounds that in this district higher wives' pay distinguishes couples nearer to fairness

Slide 52

Patterns of Regression Coef Superiority display For couples where W>H on pay, a bigger positive coef for wive's pay The principle point is that each model infers a subjectively extraordinary example of relapse weights over the two relapses.

Slide 53

Conclusion The bring home message is that nonindependence because of cooperation does not require a "factual cure" Nonindependence gives a chance to gauge and model social communication Follow your theoretical models and your examination inquiries There is still much space for cautious plan in correlational research with couples

SPONSORS