Moderates, interferes, or alters the effect of predictor on criterion variable.
Specifies the conditions under which a given effect occurs, as well as the conditions under which the direction (nature) or strength of an effect vary.
The presence of a moderator modifies the original relationship between predictor & criterion variable.
Affects the relationship between two variables. Thus, the nature of the effect of the predictor on the criterion varies according to the level or value of the moderator.
Interacts with predictors in such a way as to have an impact on the level of the criterion variable.
"A qualitative (e.g., sex, race, class) or quantitative variable ... that affects the direction and/or strength of a relation between an independent or predictor variable and a dependent or criterion variable ... A basic moderator effect can be presented as an interaction between a focal independent variable and a factor (the moderator) that specifies the appropriate conditions for its operation... Moderator variables are typically introduced when there is an unexpectedly weak or inconsistent relation between a predictor and a criterion variable"
(Baron & Kenny, 1986)
Y = a + b.X + c.M
Y = a + (b + d.M).X + c.M
Y = a + b.X + c.M + d(X.M)
Look at suggestion for future research.
Focus group discussion of respondent who are related to your study.
Look at other areas of study. We can borrow the idea from other study and use in our study.
Look at meta-analysis papers and find what are the factors influencing the focus of study.
Attend prestigious conference. People will present on going works, etc.
Should I hypothesize the form of the interactions in advance?
YES, not only should the existence of an interaction effect be predicted, but also its form. In particular, whether a moderator increases or decreases the association between two other variables should be specified as part of the a priori hypothesis (Dawson, 2014)
H1: The positive relationship between satisfaction and loyalty will be stronger when perceived image is high.
H2: The positive relationship between satisfaction and loyalty would be stronger for male compared to female.
H1: Body Mass Index (BMI) moderates the relationship between exercise and weight loss, such that for those with a low BMI, the effect is negative (i.e., you gain weight muscle mass), and for those with a high BMI, the effect is positive (i.e., exercising leads to weight loss).
When the perceived organizational support is high, the relatonship between satisfaction and organizational commitment is positive. In contrary, when the perceived organizational support is low, the relatonship between satisfaction and organizational commitment is negative.
When moderating effect variable is removed, then the relationship betwen satisfaction and commitment is zero.
From the regression coefficient of interaction term we may know that the relationship between satisfaction and customer loyalty is stronger when Image is low. However, it is not entirely clear how it differs. If you get a positive coefficient, the positive coefficient of the interaction term suggests that it becomes more positive as Image increases; however, the size and precise nature of this effect is not easy to define from examination of the coefficients alone, and becomes even more, so when one or more of the coefficients are negative, or the standard deviations of X (IV) and Z (Moderator) are very different (Dawson, 2013).
When are the relationships stronger?
An important consideration about categorical moderators is that they should only be used when the variable was originally measured as categories.
Continuous variables should never be converted to categorical variables for the purpose of testing interactions. Doing so may cause the followings:
Reduces the statistical power of the test (Stone-Romero & Anderson 1994; Cohen et al. 2003).
Makes it more difficult to detect significant effects (Stone-Romero & Anderson 1994; Cohen et al. 2003).
Throws up theoretical questions about why particular dividing points should be used (Dawson, 2013).
Issue 1. Lack of Attention to Measurement Error
When you interact IV and MV with each has 0.7 reliability, then you will get realibility 0.49, which is unrealiable. Therefore, most of the time, your interaction will not be significant. Therefore, ensure that IV and MV should have high reliability.
Issue 2. Variable Distributions are assumed to include the full range of possible values
Distribution of your DV must include high and low values. Let's say, if you include top 100 companies, and their performance are very good. Then, you should not have moderator. Therefore, we should have some with high performance, and some with low performance.
Issue 3. Unequal Sample Size Across Moderator Based Categories
If you have categorical data as a moderator, let's say 10% male and 90% female, then no point to do moderation analysis. You should also consider the minimum sample size requirement for both.
Issue 4. Insufficient Statistical Power
It is always better to collect bigger sample size as we will create interaction terms that may require larger sample size.
Issue 5. Artificial Dichotomization of Continuous Moderators
Don't categorize the continuous moderator into low and high.
Issue 6. Presumed Effects of Correlations Between Product Term and Its Components
When we create an interaction effects, we need to see the possibility of having multicollinearity effects.
Issue 7. Interpreting First-Order Effects Based on Models Excluding Product Terms
We should not run the first order effect without interaction term and interpret the result in separeted models. Instead, IV + MV + MV*IV should be pointing into the DV and estimate all relationship in one model simultanously.
Interaction plot for the categorical data moderator. Click here to download.
Interaction plot for the continuous data moderator. Click here to download.
Sharma, S., Durand, R. M., & Gur-Arie, O. (1981). ‘‘Identification and analysis of moderator variables’’. Journal of Marketing Research, 18(3), 291 -300.
Dawson, J. F. (2013). Moderation in Management Research: What, Why, When, and How. Journal of Business and Psychology, DOI 10.1007/s10869-013-9308-7.
Aguinis, H., Edwards, J. R., & Bradley, K. J. (2017). Improving Our Understanding of Moderation and Mediation in Strategic Management Research. Organizational Research Methods, 20(4), 665-685.
Becker, J.-M., Ringle, C. M., & Sarstedt, M. (2018). Estimating Moderating Effects in PLSSEM and PLSc-SEM: Interaction Term Generation*Data Treatment. Journal of Applied Structural Equation Modeling, 2(2), 1-21.
Memon, M. A., Cheah, J.-H., Ramayah, T., Ting, H., Chuah, F., & Cham, T. H. (2019). Moderation Analysis: Issues and Guidelines. Journal of Applied Structural Equation Modeling, 3(1), i-xi