Quantitative analysis (social science) ideas

[lovelearning]

Hi all! I've been doing a lot of quantitative analysis on research work (PhD) within the social science domain. this has involved extensive identification and explanations on any repeating themes, linkages and whether they relate and patterns between two variables. as part of this I have included the rank correlation to determine whether there is a relationship between the two variables. this is secondary source data.

So for example, variable 1 has data: 1,2,3,4,5 variable 2 has data 3,4,5,6,7

Is there any other stats stuff you would recommend that could be applied to this to further quantitatively analyse it? it can be anything really. I would like to find out as much as possible between the two variables. also are there any other quantitative techniques you could recommend so that they can be considered? also, would prefer the mathsy things not to be too complex pls!

any resources, ideas or info on quantitative analysis techniques would be very much welcomed.thank you very much in advance.

 

[EnumaElish]

Regression analysis.

 

[brainstorm]

Hi all! I've been doing a lot of quantitative analysis on research work (PhD) within the social science domain. this has involved extensive identification and explanations on any repeating themes, linkages and whether they relate and patterns between two variables. as part of this I have included the rank correlation to determine whether there is a relationship between the two variables. this is secondary source data.

So for example, variable 1 has data: 1,2,3,4,5 variable 2 has data 3,4,5,6,7

Is there any other stats stuff you would recommend that could be applied to this to further quantitatively analyse it? it can be anything really. I would like to find out as much as possible between the two variables. also are there any other quantitative techniques you could recommend so that they can be considered? also, would prefer the mathsy things not to be too complex pls!

any resources, ideas or info on quantitative analysis techniques would be very much welcomed.


thank you very much in advance.

What exactly are you researching? Any analysis of data, quantitative or qualitative, should be mindful of the social contexts of the data-collection and interpretive context of researchers. Your research question(s), theoretical foundations, and interpretive assumptions should be taken into account when seeking patterns within data. Otherwise, the best you can come up with is another weak correlation between relatively vague variables.

 

[arildno]

What do the numbers represent?
REAL quantities, or fairytale quantities like the answer alternative on the questionnaire made up by the social scientist?

If the latter, quantitative analysis is largely..meaningless.

Unless, for example "2" is EXACTLY "twice as much as "1", and "4" is twice as much as "2", then "regression" analysis and every other analysis will fail, AS A MATTER OF PRINCIPLE, since these techniques postulate that we DO speak about proper numerical relationships.

Quantification is meaningless if quantifiability is not present.

 

[brainstorm]

What do the numbers represent?
REAL quantities, or fairytale quantities like the answer alternative on the questionnaire made up by the social scientist?

If the latter, quantitative analysis is largely..meaningless.

Unless, for example "2" is EXACTLY "twice as much as "1", and "4" is twice as much as "2", then "regression" analysis and every other analysis will fail, AS A MATTER OF PRINCIPLE, since these techniques postulate that we DO speak about proper numerical relationships.

Quantification is meaningless if quantifiability is not present.

The problem is that social science has to operationalize data in order to quantize it, so the quantized data are already a step removed from the actual data when they are quantized. The only way I could explain this concretely is with an example. In a survey, for example, the questions attempt to tap into opinions or events known by the respondent, but it is not possible to observe those opinions or events directly. So quantizing them somewhat obfuscates the richness of the actual social material the data is supposed to represent.

Sometimes the data and the results of quantitative analysis are sufficient to address the research question, but often they are not but researchers and those reading/interpreting the research assume them to be sufficient to make conclusions that they are eager to make. Imo, most social science (quantitative and qualitative) lacks rigor in this way. Most people simply can't deal with very much of the complexity of social realities as they exist prior to conversion into data/representations.

 

[arildno]

Social scientists should limit themselves to highly specific questions where the alternatives can be expected to meet by a largely uniform interpretation.

The showcase here is: "Will you vote Republican or Democratic at the next election?"

Questions that do not have this degree of rigour to them should not be asked at all.

The point is you should NOT ask questions whose answers involve an uncontrollable degree of background variability from "The SOcial Fabric"; rather, you should limit your questioning to those tiny strands of the social fabric that are actually well-identifiable.

If you don't, you are not doing science, but something far, far worse, namely:
Myth-making
That happens to be both antithetical, and deleterious of proper scientific development and accretion.
You are NOT to ask those social questions that only the social scientists 4000 years into the future will have the capacity to ask.