Elaboration Station (and Rosenberg Summary) Part One

Okay, so ELABORATION is so not even in my Babbie Book so I'm totally going on The Little Man's words and our good buddy Rosenberg.

Here are some highlights:

Chapter 2

The most important systematic way of examining the relationship between two variables is to introduce a third variable, called a test factor, into the analysis. This is what is meant by the process of elaboration.

Typically, one begins with a relationship between an Independent Variable and a Dependent Variable (then introduce test factor - aka explanatory variable). The purpose of introducing test factors is to aid in the meaningful interpretation of the relationship between two variables.

Stratify (need to use this word to look good) on the test factor. Stratification means that we have broken the test factor into its component categories. (ie. If we are using sex as our test factor... stratify it into Male and Female)

The process of stratification creates "contingency associations" and if the relationship between between your Independent Variable and Dependent Variable DISAPPEARS within each contingent association, then we can say that the relationship is due to the TEST FACTOR.

Is the relationship (that you have "discovered") between two variables REAL? You'd better make sure it's not SPURIOUS before you go telling everyone about it.

How do you decide what to use as a control variable, then? Through LOGIC of course!

Consider these:
EXTRANEOUS VARIABLES: Remember the storks and the babies? It turned out that their relationship wasn't causal at all. The EXTRANEOUS VARIABLE of residence in rural areas was making it look that way though. But nobody really thought about that because it was so far out and away from the "world of storks and babies" (hence it was/is EXTRANEOUS) that we ended up with the whole silly story about how storks deliver babies (just another example of how the crucial role of the woman in the whole delivery process gets passed over by the conclusions drawn by men long ago and nobody bothered - for a long time - to question the spuriousness of the stork/baby idea - never mind... thought I was in a WGST course for a moment).

We need to consider EXTRANEOUS VARIABLES in order to guard against misleading interpretations which might derive from the assumption that an inherent link exists between the two variables. They thought the storks and the babies were inherently linked. WRONG!!!

Here, the test factor leads to the Independent Variable and the test factor also leads to the Dependent Variable. THE CRUCIAL POINT IS: The Independent Variable DOES NOT lead to the Dependent Variable. The relationship is symmetrical; the variables are consequences of a common determinate.

Rosenberg goes on to discuss Component Variables, but since Nigem doesn't seem to care about them, neither do I right now........

Chapter 3

An INTERVENING VARIABLE is viewed as a consequence of the independent variable and as a determinant of the dependent variable. In other words, when you think you have found a causal relationship between an Independent Variable and a Dependent Variable, someone points out that you skipped a link in the "causal chain". Something occurs between your Independent and Dependent Variable that you missed. That "thing" is going to be our test variable (again).

Remember Nigem talking about how Durkheim was all about the way that Religion (Independent Variable) was causal to Suicide (Dependent Variable) and it turned out that (this was somewhat true, but it wasn't the whole truth and nothing but the truth) Integration was the meat missing from his Independent and Dependent Variable white bread sandwhich. Integration is the INTERVENING test factor (where's the beef? I want a last meal before I die!) VARIABLE.

So now we know that there are three parts to this deal (we got two pieces of bread and the meat - vegetarians insert meat substitute here) To establish a variable as INTERVENING the presence of three asymmetrical relationships must be present. You need your original relationship (I causes D, x causes y, I can't make no full-sized sandwich witout 2 slices of bread), and then you need the meat. But for the sandwich to be real, your top piece of bread need to be connected to the meat. If they're not (looking from my sandwich from the side from top bread to meat to lower bread) in the right order with my Bread leading directly to my Meat (relationship 2 where your original Independent Variable causes the test factor now making it a Dependent Variable) and then that meat needs to be followed directly by that other slice-o-bread (relationship 3 where the test factor now becomes the Independent Variable that causes your original Dependent Variable). INTERVENING VARIABLES are all about sandwhiches baby! that's why they're called Intervening Variables in the first place, don't you know? Because sandwiches are good for lunch and lunch INTERVENES in the middle of your day!

But don't forget, if you control for the Intervening Variable (or take the meat away) the original relationship disappears or decreases significantly. You no longer HAVE a sandwich, you've just got some seperate, unrelated pieces of bread.

(Now I need to take a quick break and listen to the Detroit Grand Pubahs' song "Sandwhiches" which you will certainly see me silently (yet somewhat visibly) dancing to in my desk tomorrow when I get to this portion of the exam.

I'll be back soon!