source: Halperin & Heath
To what extent can we separate facts and values in research?
Positivism holds that facts are distinguishable from values and that science should be ‘value-free’. Naturalism holds there is no difference between the scientific and social world and therefore should be studying according to scientific principles.
There is a divide between empirical study, the study of what is, with normative study, the study of what ought to be.
Social science should strive to be value neutral but there will be slight bias in collecting data, in analysis, and in the conclusion. The research process cannot be devoid of a researcher’s values completely.
Kuhn argued since science is a consensus as well as a process than it must be a social endeavor. Science is a process of cumulative gradual knowledge acquisition. Science is the falsification of theories, one theory works until it fails and a new theory arises in its place to correct where the first theory failed. Science is made up of paradigms or models of theory used to understand the world. Science is dominated by paradigm or schemes that scholars agree on about objects of investigation and the methods of investigation used. I think by “box” or conceptual scheme Kuhn means a model.
In Lakatos’ theory, science isn’t a mob psychology but a process by which scholars choose one research programme over another rationally. Scientific research programs unlike paradigms are based on a hard core of basic assumtions that are scientifically accpected, and an outer layer, the protective belt where theories are sought to be falsified. Science in Lakatos’ theory holds theories undergo a rigourous and empirical process and if proven false are incrementally modified to the hard core of the theory.
31 January 2017
Read Chapter 4 – completed
- Why would a methodological individualist object to a hypothesis about government agency actions?
- The methodological individualism branch of research holds that “social concepts should be defined by individuals in their physic and physical states, actions, interactions, social situation, and physical environment,” (Halperin & Heath, p.87) meaning the unit of analysis is the individual and the dependent variable is the behavior of the individual therefore would require a metholodical approach therefore would reject the macro agential approach to government agency theory. Methodological individualists study causation at the micro level, while government agencies should be studied at the macro level. Government agency would be a structuralist variable and therefore more likely studied ontologically and epistemologically by a methodological holist. A methodological individualist would more likely study the outcome behavior of individuals inside the government agency. To reiterate, a methodological individualist would reject a hypothesis regarding a government agency except if studying an individual in the agency unless adopted a rational choice model where institutions like a government agency may be studying agentially.
- In this hypothesis, what are the independent and dependent variables and the units of analysis?
- Individuals are more likely to pursue short-term goals than long-term goals.
- The independent variable is the X variable, the variable known to cause the causation affect on the Y variable, or dependent variable. The relationships may be positive or negative (inverse).
- In this hypothesis, individuals are the independent variable. It is unknown whether the individuals will be studied individually (agentially) using a n1, n2… form or whether the individuals will be studied as a group or summation of individuals (collective action). The dependent variable or outcome variable are the short and long term goals. The dependent variable may be categorically given two options here.
- The causal mechanism in the relationship between the individual (independent variable) and outcome short-term and long-term goal (dependent variable) is the degree of interest in pursuing.
- This is a epistemological question using ontologically gathered variables and given how the individuals variable is measured, it could be a methodolically individualistic hypotheis (n1, n2) or a methodolodically holistic hypothesis (collective behavior of individuals).
- If I were to pursue an analysis of this hypothesis, I would methodically perform an analysis in which the independent variable: individuals was logically regressed to find the outcome of the short-term long-term goal question.
- Individuals are more likely to pursue short-term goals than long-term goals.
3. In this hypothesis, what are the independent and dependent variables and the units of analysis?
- African-American voters are more likely to vote for Democratic candidates than for Republican candidates.
In this hypothesis, there could be two relationships, therefore two independent and two dependent variables. African-American voters can be measured as individuals (agential) or as a collective (structurally). African Americans are the independent variable because they are the causes of the outcome. The outcome or dependent variable is whether or not the candidate voted for a 1) Democratic candidate or if 2) voted for Democratic candidates overall. To reiterate, independent variable 1 is an African American individual, and independent variable 2 is African Americans as a collective group. Dependent variable 1 is categorical variable, a vote for or against a single candidate, and dependent variable 2 is an overall record of electing democratic candidates.
This question may both be studying methodically individualist or holistic depending on what variables one chooses. If choose an micro-level individual African American that would be a methodically individualistic approach versus the alternative: studying African Americans as a macro-level group which is the holistic approach. The unit of analysis would be the vote itself, the African American individual, and the outcome vote.
Given the independent variable is the individual African American and the dependent variable is a categorical variable with two dimensions: Democrat and Republican, I would preform a logistic regression to test for statistical significance.
One would have to be careful to hear to avoid a common ecological fallacy where assumputions are made on individuals based on aggregate data about individuals. It would have to be an especially rigorous statistical test, that is if it is an empirical epistomolical question, to prove whether an individual is more likely to vote for a democratic candidate if he or she is African American (if X: being African American -categorical) results in a vote for a Democratic candidate (Y variable). The fallacy would be that one take aggregate data that measures individuals on the macral, collective structural level and assumes it will be able to predict micro agential level behavior.
In this hypothesis, what are the independent and dependent variables and the units of analysis?
- In elections with two unpopular candidates, voter turnout will be lower than in elections with at least one popular candidate.
If I were to test this hypothesis, the first thing I would do is collect data. I would find a sample of elections in a state perhaps and given a certain amount of time, then I would find each candidate’s popularity rating. I would create one set where the elections were held with two unpopular candidates (Set A) and a set where there is atleast one popular candidate (Set B). This would serve as the independent variables of the analysis.
The next step would be to gather dependent variable data: voter turnout. I would match each election from Sets A and B and record the magnitude of turnout which would be how many individuals turned out for the election. This measurement will be the dependent variable.
The results of my hypothetical analysis are found on the table below. The results show the total number of votes (my measurement for voter turnout) is the dependent variable. The stacked barplot shows 3 sets of data totaled to 1905 for Set A and 3548 for Set B. The independent variable is shown on the X axis, and it is the Sets, and therefore the elections. The results of the hypothetical analysis shows that Set B (atleast one popular candidate) had the aggregate higher turnout of individuals and therefore not falsyfying the hypothesis. The observational data (Set A: two unpopular candidates) showed a lower turnout (less aggregate votes) than Set B (one popular candidate) as the hypothesis guessed. The expected data matched with the observational data therefore not proving this theory as falsified.
The units of analysis would be the elections: the candidates, the outcome of the election, as well as the individuals who turned out to vote. The aggregate sum of the total number of individuals is not a unit of analysis but an aggregate summation of individual units.
Because I use a bar plot with the aggregate sets on the X axis and the total number of voters on the Y axis, this analysis is special because it is:
- An example of positivism where the social world is being studied like the natural world using statistical methods.
- An example of naturalism such that there is no distinction from the natural and social world
- empirical: using observations (results of an election is in quantifiable)
- This hypothesis gives no causal mechanism for the relationship between the independent and dependent variables so no causal inference is made.
- An example of methodolical holism because voter turnout is measure as a collective macro agent rather than micro agent and elections are measured in a structural outcome. Therefore, this knowledge is structural epistemology.
Lecture for the Week of February 7
The lecture item for this week is a video I want you to watch of Professor Katherine Cramer doing a presentation on her research on rural political resentment in Wisconsin. While you are watching this, pay attention to her main concepts, theory and hypotheses and research methods. The second short writing assignment asks you to describe her theory, concepts and hypotheses. In a later assignment, you will be asked about her research methods.
Here is the link: https://www.c-span.org/video/?416995-7/katherine-cramer-discusses-politics-resentment
In Blackboard, it usually works best to copy the link and paste it into a new page rather than click on it.
After you have listened to Professor Cramer’s presentation (see Lecture for Feb. 7), write a one-page description of Professor Cramer’s theory, her major concepts and the general hypotheses that made up her research. Your description should be written in narrative form rather than list form. Put it in a document and email it to me before 11:59 p.m. on Sunday, February 12. (I know an earlier version of the assignment schedule said to post it. Don’t – email it please.)
The Politics of Resentment Professor Katherine Cramer talked about her book The Politics of Resentment: Rural Consciousness in Wisconsin and the Rise of Scott Walker, in which she examines voter distrust. She spoke at the 2016 Wisconsin Book Festival, held at the Central Branch of the Madison Public Library.
Submitted: .pdf available here: 2ndshortassignment_awis-5