mediation analysis and cultural psychology

Mediation mechanism is defined as how the addition of the third variable influences the relationship between two variables. In the simplest mediation model, X causes the Y, then Y causes the Z, where Y is perceived as a mediated factor in a directional causal chain. Experimental approaches and statistical approaches are applied in mediation analysis. Randomized experimental design is used in the investigation of mediation. Controlling the individual differences and experimental mechanism, the difference of means is interpreted as the outcome of the manipulation of mediator. Since the experiment is the unique method to explore the causal relationship, the mediational process is confirmed by two randomized experiments. Despite the robust power to explain the mediation process, the difficulty of extending the results to other situations has limited the application of experimental approaches to mediation. The manipulation and measurement of mediator are not both considered by the experimental approaches, thus the statistical analysis in necessary to investigate the mediation. Causal steps, differences in coefficient, product of coefficient are three major approaches of statistical analysis, where causal steps is widely used one which first establishes the linear equation between the independent variables and dependent variables, then establishes two equations to explore the relationships between the mediator and two variables respectively. In the process of establishing the three equations, the coefficient relating the independent variables and dependent variables in the context of mediation variable must be bigger than the coefficient of the direct relationship between independent and dependent variables. The mediated effect is calculated through two ways: aˆ *bˆ and cˆ−cˆ, the rationale behind the first method the author (David P. MacKinnon, et.al, 2007) emphasis is “that mediation depends on the extent to which the program changes the mediator, a, and the extent to which the mediator affects the outcome variable, b”. the first is involves the reduction of effects between independent variables on the dependent variables within and without the mediator.

The picture is clearly illustrate the components of mediated effect

(quoted from David P. MacKinnon, Amanda J. Fairchild, and Matthew S. Fritz(2007))

The traditional calculation of limit confidence based on the normal distribution of mediated effect is inaccurate, which is gradually replaced by the bootstrap analysis, which is a non-parametric method of effect-size estimation and test hypothesis and will overcome the power problem, where small samples are also available to be tested.

Beyond the single model of mediation analysis, multiple mediators, multi-level mediation level and longitudinal mediation analysis have been developed quickly. Although the mediation analysis has a continued progress, the doubts still exist in the fundamental assumption of mediation analysis. How regression model reflects the causal relations among variables? The alternative possibilities need more information to exclude while mediation analysis never provides. The principle stratification of possible relations of variables is suggested to take as a promising alternative.

The application of mediation analysis in psychology has a long history(Dov Cohen, 2007). In Hebb’s stimulus-response model, mediated mechanism such as cognitive or neural processing mechanism is considered as bridges between the stimulus reception and behavioral response. Besides, psychological factors mediating how social context influences individual action play important roles in the field of social psychology. Psychological factors mediating the cultural and psychology is an indispensable issue in cultural psychology. How mediation mechanism effects in the interaction of culture and individuals? Cultural properties such as individualism-collectivism, dispositional measures, the judgment and reason of individual, choice-making perception, goals, social beliefs have been considered as the alternative explanation of mediator, however a large number of hypothesized mediators have not been proved. As Dov Cohen(2007) pointed out, the overemphasis of individual differences in cultural psychology will produce a risk to “reduce culture to individual difference or inside-the-head variable that neglects how situation ,practice and institutional arrangement afford certain type of behavior”. How culture outside the individual play the roles will be overlooked.

In order to solve the paradox, Hong(2003) presented a view in his bicultural model that there can be more than one cultural constructs which can be coexisting disregarding their contrastive properties. Cultural activation as a effective tool can be applied to direct people to behave in different cultural situation. Therefore, the previous absolute and dichotomous view on cultural syndrome of people from different cultural atmosphere has been challenged powerfully. In another way a dynamic viewpoint is introduced to replace the dispositional one. Reasoning styles rather than differences of personalities and preferences act as a key mediation mechanism.
The measurement of mediator also becomes the focus of criticism, since questionnaire is the most popular way. However, the questionnaire is not a valid and sensitive indicator to cross-cultural differences. Therefore a new measurement method is necessary to seek for the mediator better.

Although mediation analysis is very useful in the interpretation of psychological mechanism in cultural psychology, a lot of problems can’t be solved through it, like the problem of biased sampling and the problem of excluding other cultural syndrome.



References:
David P. MacKinnon, Amanda J. Fairchild, and Matthew S. Fritz(2007)，mediation analysis Annual Review. Of Psychology.. 58:593–614

Hazel Rose Markus and MarYam G. Hamedan (2007)i, Sociocultural Psychology: The Dynamic Interdependence among Self Systems and Social Systems, hand book of cultural psychology, The Guilford Press A Division of Guilford Publications, Inc.

Dov Cohen (2007 ), Methods in Cultural Psychology, handbook of cultural psychology The Guilford Press A Division of Guilford Publications, Inc.

Hong, Y., Benet-Martinez, V., Chiu, C., & Morris, M.W. (2003). Boundaries of cultural influence: Construct activation as a mechanism for cultural differences in social perception. Journal of Cross-Cultural Psychology, 34, 453–464.

Briley, D. A., Morris, M. W., & Simonson, I. (2000). Reasons as carriers of culture: Dynamic versus dispositional models of cultural influence on decision making. Journal of Consumer Research, 27, 157–178.

SPSS and SAS procedures for estimating indirect effects in simple mediation models
Behavior Research Methods(2004), Instruments, & Computers, 36 (4), 717-731