A moderation analysis via PROCESS model 1: meaning, procedure, plotting and interpretation

everybody this is golda tabari and today i am going to explain to you how you can run a moderation analysis using process model 1 how you can interpret the output or obtain from it and then how you can plot and interpret the graph first of all a moderation analysis is a kind of regression analysis which explains the impact of your independent variable on your dependent variable through all under the influence of a moderator which is a third variable so this is the thing that model one tries to explain it it says when your independent variable has an effect on your dependent variable under the influence of a moderator then you can use model one to explain that effect to explain it statistically in this model in this conceptual model model of module one you have three pathway the first pathway is b1 which is a pathway of direct effect of your independent variable on Y on your dependent variable but B 2 which is your direct effect of your moderator on your dependent variable this path is path b2 and the most important path is interaction part of path X and path moderator so this and this finally this is the path B 3 which is the multiplication of this in this path so this is the interaction effect how you can calculate the total conditional effect of X on Y its through path b1 which is this direct path plus a multiplication of path b3 which is X M and the path this so this path when multiplied by this path gives the total condition a conditional effect of your independent on your dependent variable by Y so actually this moderator explains the direction of your conditional effect of direction of your effect of your independent variable on Y and or both can explain the strength of this effect of your X&Y now in this video I am going to illustrate this model through one of the example one of my own studies I am going to explain I how message framing has an effect on the organization reputation in the event of a crisis and how this impact of message framing an organization reputation is influenced by the time of crisis so the actual story is I was curious to know in the event of a crisis when a company faces a crisis situation and a CEO has to give a speech during or after the crisis then the kind of speech the CEO uses and the kind of framing he uses in that if how it can have an effect on the overall organization reputation and this affect how it is influenced by the time of crisis so I wanted to form a comparison between a factual actually framed message and a method metaphorically framed message on the organization reputation under the influence of crisis time so then I love I run a model one analysis in order to know the direction of this effect I wanted to know which has a grid if there is a effect under the influence of crisis tank that was my first question and if there is an effect then at what time the effect is significant during the crisis or after the crisis and if it is significant at a particular time of crisis then which one of these two frames have a significant and a better effect on organization reputation so I am going to explain first how you can obtain an output of results through module one so I am going to open my data file this is my data file these are all the variable variables I am only going to use the desired one my X and my Y and my moderator so I will go to analyze and then I go to regression I in my s software of SPSS I have already this plug-in of process installed but if also I am going to tell how you can install a process plug-in you have to go to extension and then utilities and then you have to go to install custom dialog and in your computer when you install a plug-in from process Hayes website you will find somewhere that it is saved under the name of process 2 1 3 you go there you click process SPD and then open so here it gets installed in your SPSS file so now when you go to analyze and then go to regression you will see the plug-in for process you see here this is the plug-in for process you click it and then now you can choose module 1 there are lots of models associated with process but we are interested in module 1 now bootstrap you generally keep at 5,000 and then now you keep on loading your variables so my outcome variable is organization reputation so I'm going to quickly choose that so here is my variable of organization reputation my independent variable is message framing which I saved as speech because one thing you have to keep in mind when you are trying to run a model one or any module in SPSS try to name it in a way at the overall number of words you use there or the number of characters you use in the name of the variable are less than 8 characters so this is your independent variable I named it at message framing or the used in the speech of CEO so I just named it as speech and my moderator crisis phase generally this is the place value load your moderator proposed moderator but when you use model one you have to remember this that this has to be loaded somewhere here where you generally use your mediator so this is the way you do it and then you have to keep one thing into mind if you want to obtain a plot out of model one you have to choose an option you have to go to option and then you have to click this option generate data for plotting model one two and three only so you have to make sure that it is checked in and now continue and hit OK and you have to wait for the results now a little bit of patience yeah so this is the output first thing you have to do when you want to know the interaction effect you have to directly see interaction one so you have to know whether it's significant or not if it is significant that means the your moderation effects is significant so you have to check this as the first thing and you see it is significant how you know that when your confidence interval when zero does not lies between your confident and confidence interval that means your effect is significant as you see point zero nine is your lower confidence interval and point nine eight is your upper confidence interval and zero does not lie in between these confidence intervals which significant significant errection effect is significant so that's a good thing that your moderation effect is significant second things you have to know is under which condition of your moderator this effect is significant it may be possible that both during and after the crisis condition which was my moderator here is the effect is significant but it may also be possible that under only one condition of your moderator the effect is significant so you can check it here conditional effect of X&Y add the values of moderator so when you check it you see zero was coded as during the crisis condition so this is crisis phase and one as after the crisis condition for during the crisis condition when you check the confidence interval you find out that yes zero is not between the confidence interval because this is also negative the lower confidence interval Anabelle confidence interval is also negative so zero does not lie in between so that's a good news that during the crisis phase the moderation effect is significant but what happens after the crisis phase which is coded as one you see lower confidence interval is minus 0.45 and upper confidence interval is poor the point one seven that means zero is lying somewhere in between here which is it says this moderation effect is not significant for after crisis condition one thing I will also try like to add here in this moderation analysis in model one analysis this moderator can be both a categorical and also a matrix variable that is a continuous quantitative variable it can be both only difference that you will find in the result output is when it is a categorical variable like in my case I already chose the crisis time as a categorical variable so zero was during the crisis and one was after the crisis but in case you have a matrix variable or a quantitative variable then SPSS this process does it for you so it will automatically divide your moderator into low medium and high condition so you need not to worry about it so that's a good thing process does for you in my case I and then in that case you also know what happens in a moderate Christ's moderate condition of your moderator let me medium condition so you have three category there low medium and high so you it does it for you by itself since my variable of moderator is was quite a coracle here so it gave me output for only my Quetta got two categories so now I know that moderate this moderation or the interaction effect is only significant during the crisis condition but I don't know for which condition of my independent variable which was message frame and two conditions were factual or metaphorical the effect on organization reputation is higher I am that the my main research question that was my main question first was to know whether there is an interaction effect of X on Y through on under the influence of moderator of crisis phase but also if there is an effect then for which condition of my independent variable factual or metaphorical the effect is more higher so for that I need to plot a graph and that graph can be plotted using this data so you see here I find three rows three and six four columns and then I see this one indicates speech that means my condition of X my independent variable 0 is factual condition and one is metaphorical condition this depicts my crisis phase zero is during crisis and one is after crisis and this is the effect on organization reputation through under the influence of crisis phase so using this data I can plot my graph of module 1 how you how we can do it I am going to explain to you now so just open a Microsoft Excel sheet and all you have to do it you have to make a matrix here and then you have to write your conditions or that categories of your independent variable and also the moderator so I am going to write factual here and the other condition is metaphorical which was coded as 1 and then I am going to write here during crisis conditions and then ask a crisis condition which was scored as coded as 1 so this was 0 this was 1 this was 0 and this was 1 and now I am going to find the combination so 0 and 0 so this is the combination 4.0 so I just type it there 0 factual after condition 1 so 0 1 this combination is here 0 1 which is 3 point 8 8 I write it here metaphorical was coded as 1 during crisis as 0 so 1 & 0 so which is here which is three point three 2 then metaphorical as one and after the crisis condition is also 1 one and one which is here 3.73 so this is my data for plotting the graph of module 1 and now I am going to select this and then I am going to insert graph line and then this is the desired graph now to make it more clear to you I am going to label the points first and then I am going to explain to you what is the interpretation so to make it a clear graph I will just label all the titles and axes so that the chart title I am going to write as impact of message frame on organizational reputation under the influence of crisis phase this is the title of my graph and then I'm going to label these all the points so I'm going to add chart element XS titled and then this is the vertical title so my independent variable is organization reputation so I am going to label my y-axis as organization reputation to make it a clear graph I am going to label and then increase the font size so organizational reputation and my x-axis I am going to label it as my independent variable which is message frame yes so these are to make it more clear I'm going to just increase the font size and do something and then this graph is ready in order to explain to you in a better way I am going to zoom and zoom into the graph so this is the graph so now the explanation as you already know from the output that this effect interaction effect is only significant during the crisis condition so this is the place where the effect is significant I am again going to refer the output so that I you get to know what I am talking about so we already I repeat we already know this this is significant only the during the crisis phase so zero crisis phase and we see this effect is only significant here not after the crisis phase condition which was coded as one which where zero was between the confidence interval so this is not significant so we see in this graph that we interpret that after the crisis condition this difference was not significant so we cannot say anything about this zone but we can say something about this zone which is significant zone this one and what we can say from there we see that factual condition was labeled with a blue line so we see this plot actually indicates factual condition in this metaphorical condition and we see during the crisis condition factual speech factual framing of speech has a better organization reputation then a metaphorical framing so that means it is better to use a factually framed message crisis response message bye see you during the crisis phase and nothing can be done what should be used after the crisis because this zone is not significant so this is the interpretation of this process graph which we got through his process model one

Channel: Goldi Tewari

Views: 108,898

Rating: 4.8157506 out of 5

Keywords: moderation analysis, PROCESS Model 1, SPSS, Interaction effect, Plotting a moderation, Statistics, Conditional effect, PROCESS Model 1 graph

Id: Zx1DJfhzO_o

Length: 18min 50sec (1130 seconds)

Published: Wed Jan 25 2017