Hi there,

I guess you are exchanging control variables with moderator variables.

A control variable has be included in a regression in order to avoid biases of the regression coefficients of the target variables. Incussion is necessary, when the respective control variable (a) has itself an effect on Y and b) is correlated with the target variable.

A moderator variable is a variable that is hypothesized to systematically change the effect of the target variable. That means there is no main effect of the target variable but the effect is conditional and the moderator.

Both control and moderator variable can be metric or dummies and moderator variable can also have a controling function (if they match both mentioned criteria).

Best,

Holger

I agree with Holger but will also add that it is possible for main effects to be present even when a moderator variable is statistically significant.  

Simply said, the moderating or the intervening variable is those having influence (positively or negatively) on the relationship between the independent and dependent variable.

It will be more useful to suggest a relevant variable to the question if the scope of study or the issue can be specified.

If suppose HR practices are independent variable(s) and firm performance is dependent variable, then what and how moderator variable(s) moderates and affect the results.

Again it depends what is it you want to study. For example you want to investigate the effect of 'Leadership Style' as the moderating variable. The expected results are either leadership style would possibly stimulate the relationship between the HRM practice and the organizational performance, or otherwise.

The meta-analysis below found a stronger positive relationship between HPWPs and org. performance among manufacturing firms than among service firms. Industry or firm type is thus a moderator:

Combs, J., Liu, Y., Hall, A., & Ketchen, D. (2006). How much do high-performance work practices matter? A meta-analysis of their effects on organizational performance. Personnel Psychology, 59, 501-528. doi: 10.1111/j.1744-6570.2006.00045.x

As the other authors had  shared, moderating variable / moderator is different from control variable and moderator can be demographic factors or one / more the contructs that a researcher is studying together with IV & DV.  Moderator is starting at the same time as your IV in which both IV & moderator are having interaction that influence the DV.  Albeit some moderating hypotheses in articles are not clear, but by reading a clear hypothesis, we can understand what the moderator will have impact and how we want to test the moderating effect.  For example use back the scenario i.e. IV = HR Practices, DV = Firm Performance and Moderator = Leadership Style, the alternative hypothesis can be H1: the relationship between HR Practices and Firm Performance can be moderated by leaderships style such that the relationship is stronger with multiple leadership styles / transformative leadership and weaker with only single leadership style / transactional leadership.  The above hypothesis is clearer than H2: the relationship between HR Practices and Firm Performance can be moderated by leadership style. 

Other things we need to take note when testing moderating effect which include:

1) the rationale to select a variable as moderator which should be from literature review e.g. IV not really directly influencing DV but there is invisible hand of moderator or from literature review, many studies use a particular variable as IV and no one has use the variable as moderator for knowledge contribution etc.

2) when testing moderating effect is the goal of the a  research, a researcher needs to ensure / collect larger sample size because when many moderators are being introduced in a study, the interaction effect between IVs and moderators will create more new predictors (IV1 x Moderator2, IV9 x Moderator9 etc.) that can influence DV.  Some statistical analyses method used dependent on no. of predictors x certain constant to determine the no. of sample required.  With small sample size, most of the moderating effect will not be significant.

Dear Han Ping Fung, you are very correct. Further, there are studies where moderators  are not introduced, it means the regression model does not explain maximum variance or some variance remained unexplained.

When we are using control variables in regression model, usually we do not explain the variance explained by control variables. But when we introduce moderators, then we have to explain the effect of moderators on the relationship between independent and dependent variables.

Seen this one: Kuvaas, B. 2008. An exploration of how the employee-organization relationship affects the linkage between perception of developmental human resource practices and employee outcomes. Journal of Management Studies, 45(1): 1-25.

Dear Bård Kuvaas,

Excellent paper for explaining the variables.

Thanks

Moderator variable is trying to explain the significance relationship between indepenedent variable and dependent variable in a research, therefore, the interaction between IV and moderator variable are influencing DV.

My research topic is: the moderating effect of profitability on the relationship between corporate governance and dividend policy. Please is it researchable? Because base on my search no paper on the topic but only in a direct relationship between the corporate governance and dividend policy the moderating variable is used as control variable that is the profitability.

Nice discussion