1. De Vaus – Research design in social research(2001)
Social researchers ask two fundamental types of research questions
- What is going on? (descriptive research)
- Why is it going on? (explanatory research)
The way in which researchers develop research designs is fundamentally affected by whether the
research question is descriptive or explanatory. It affects what information is collected.
Descriptive research
-Good description research is fundamental to the research enterprise and it has added to our knowledge
of the shape and nature of our society.
-Descriptions can be concrete or abstract.
-Good descriptions provoke the ‘why’ questions of explanatory research.
-Description can degenerate to mindless fact gathering, which is called abstracted empiricism.
Explanatory research
-Focuses on the why questions, which involves developing causal explanations. Causal explanations
argue that phenomenon Y is affected by factor X. Causal explanations could be both simple and very
complex.
Three types of causal relationships
Prediction, correlation and causation
Correlation is often confused with causation, but the improvement of y after event x may be coincidental
rather than causal. For example another variable might influence the relationship. The confusion may
also confuse prediction with causation and prediction with explanation. Good prediction does not
depend on causal relationships. Nor does the ability to predict accurately demonstrate anything about
causality.
,We can observe correlation, but we can’t observe cause. We have to infer cause. These inferences are
fallible. Because they are fallible, we must minimize the chances of incorrectly saying that a relationship
is causal. One of the fundamental purposes of research design in explanatory research is to avoid invalid
inferences.
Deterministic and probabilistic concepts of causation
There are two ways of thinking about causes: deterministically and probabilistically. Deterministic
causation is where variable x is said to cause y if, and only if, x invariably produces Y. That means that
when x is present y will ‘necessarily, inevitably and infallibly’ occur. For example: Water always boils at
100 degrees of Celsius. But in reality laws are never that simple, because laws are always under specific
conditions. Therefore deterministic laws are sometimes states as “all other things being equal”.
Most causal thinking in the social sciences is probabilistic. Which means that a given factor increases(or
decreases) the probability of a particular outcome. For example: Being female increases the probability
of working part-time. Probabilistic explanations can be improved by specifying conditions under which x
is less likely/more likely to affect y, but you will never achieve complete or deterministic explanations.
Under probabilistic explanations there will never be certainty about outcomes. Despite the probabilistic
nature of causal statements in the social sciences. Much of these statements are shown as deterministic
statements.
Theory testing and theory construction
Theories are attempts to answer the ‘why’ questions in social sciences. These theories vary in
complexity, amount of variables and links, abstraction and scope. To understand the role of theory in
empirical research we should distinguish two different styles of research: theory testing and theory
building.
Theory building
Theory building is a process in which research begins with observations and uses inductive reasoning to
derive a theory from these observations. These theories attempt to make sense of observation. It is
called post factum theory/ex post facto theorizing, because the theory is produced after the
observations are made. This form of theory building entails asking whether the observation is a
particular case of a more general factor, or how the observation fits into a pattern or a story.
,Theory testing
A theory testing approach begins with a theory and uses theory to guide which observations to make; it
moves from the general to the particular. The observations should test the worth of the theory. It uses
deductive reasoning to derive a set of propositions from the theory. We need to develop these
propositions in a way that if the theory is true, things in the real world should follow. We can assess
whether the theory is supported or rejected.
Although theory testing and theory building are often presented as alternative modes of research, they
should be part of an ongoing process(see picture below). Typically, theory building will produce a
plausible account or explanation of a set of observations. However, such explanations are frequently just
one of a number of possible explanations that fit the data. While plausible they are not necessarily
compelling. They require systematic testing where data are collected to specifically evaluate how well
the explanation holds when subjected to a range of crucial tests.
What is research design?
Social research needs a design or structure before data collection or analysis can commence. A research
design is not just a work plan. A work plan details what has to be done to complete the project, but the
work plan will flow from the project’s research design. The function of a research design is to ensure that
the evidence obtained enables us to answer the initial question as complete as possible. Obtaining
relevant evidence entails specifying the type of evidence needed to answer the research question, to
test a theory, to evaluate a programme of to describe some phenomenon. In other words, what type of
evidence is needed to answer the question(or test the theory) in a convincing way?
In social research the issues of sampling, method of data collection, design of questions are all subsidiary
to the matter of “what evidence do I need to collect?”.
Design versus method
Research design is different from the method by which the data are collected. How the data are
collected is irrelevant to the logic of the design. Failing to distinguish between design and method leads
to poor evaluation of designs. Equating cross-sectional designs with questionnaires, or case studies with
participant observation, means that the designs are often evaluated against the strengths and
weaknesses of the method rather than their ability to draw relatively unambiguous conclusions or to
select between rival plausible hypotheses.
,2. Hair, J.F. Jr., Anderson, R.E., Tatham, R.L. (1987) Multivariate Data
Analysis, second edition, pp. 20-40.
1.What is multiple regression analysis?
Multiple regression analysis is a statistical technique that can be used to analyze the relationship
between a single dependent variable(criterion) and several independent variables(predictors). The
objective is to predict the value of the dependent variable with the independent variables, of which the
values are known.
To use multiple regression analysis: (1) The data must be metric, and (2) before deriving the regression
equation, the researcher must decide the dependent and independent variable(s).
For what purposes is regression analysis used?
1.Determine the appropriateness of using the regression procedure with the problem.
2.Examine the statistical significance of the attempted prediction.
3.Examine the strength of the association between the dependent and independent variables.
4.Predict the values of one variable from the values of others.
Adding a predictor is useful when it gives more information about the dependent variable.
Types of regression
There are three types of regression. (1) Prediction using a single measure, the average, (2) Prediction
using two measures, simple regression (3) Prediction using several measures, multiple regression.
Prediction using a single measure, the average
Simply using the average of a number of measures to answer a question. The way to assess the adequacy
of using the average as predictor is to examine the errors that are made when it is used. For example the
average is 4 and a someone has the value 2, the average method has an error of +2. The sum of the
errors will always be 0, so you have to square the errors, and obtain the sum of the squared errors. It
provides a good measure of the prediction accuracy of the arithmetic average. For a single set of
observations, no other measure of central tendency will produce a smaller sum of squared errors than
the arithmetic average.
Prediction using two measures, simple regression
Also tries to minimize the sum of the squared errors. If we have additional information, we could
improve our prediction. If we know that the dependent variable is related to an independent variable we
could write a relationship: . Y^ = Predicted value, B0= the intercept, B1= the slope and X1=
the x value to predict y value.
Functional relationship: Calculates an exact value.
Statistical relationship: Estimates an average value.
, 2.Major assumptions of simple regression
-Statistical relationship: We assume that the data we are working with is statistical data.
-Equal variance of the criterion variables: We assume that at each level of the predictor variable the
values of the criterion variable have the same variance(homoscedasticity), because our prediction is
based on actual values from our predictions. It means that the variance of Y at x1.x2, etc. are all equal. If
the variance isn’t equal, the predictions will be better at some levels of the criterion variable. We could
be misled into thinking that our predictions are equal at all levels, when some levels are equal(the ones
you have calculated).
-Lack of correlation of errors: We want to find that any errors we make in prediction are uncorrelated
with each other.
Fixed predictor= The levels of the predictor variables are fixed.
Random predictor = The levels of the predictor variables are selected at random. Our interest isn’t just in
the levels examined but rather in the larger population of possible predictor levels from which we
selected a sample.
3.Prediction using several measures, multiple regression analysis
You add another predictor variable to improve your predictions. Goal is to have the lowest sum of
squared errors as possible.
On top of the assumptions of simple regression, you must be concerned with possible interaction and/or
correlation among the predictor variables, because we have assumed that they don’t interact and are
uncorrelated.
Social researchers ask two fundamental types of research questions
- What is going on? (descriptive research)
- Why is it going on? (explanatory research)
The way in which researchers develop research designs is fundamentally affected by whether the
research question is descriptive or explanatory. It affects what information is collected.
Descriptive research
-Good description research is fundamental to the research enterprise and it has added to our knowledge
of the shape and nature of our society.
-Descriptions can be concrete or abstract.
-Good descriptions provoke the ‘why’ questions of explanatory research.
-Description can degenerate to mindless fact gathering, which is called abstracted empiricism.
Explanatory research
-Focuses on the why questions, which involves developing causal explanations. Causal explanations
argue that phenomenon Y is affected by factor X. Causal explanations could be both simple and very
complex.
Three types of causal relationships
Prediction, correlation and causation
Correlation is often confused with causation, but the improvement of y after event x may be coincidental
rather than causal. For example another variable might influence the relationship. The confusion may
also confuse prediction with causation and prediction with explanation. Good prediction does not
depend on causal relationships. Nor does the ability to predict accurately demonstrate anything about
causality.
,We can observe correlation, but we can’t observe cause. We have to infer cause. These inferences are
fallible. Because they are fallible, we must minimize the chances of incorrectly saying that a relationship
is causal. One of the fundamental purposes of research design in explanatory research is to avoid invalid
inferences.
Deterministic and probabilistic concepts of causation
There are two ways of thinking about causes: deterministically and probabilistically. Deterministic
causation is where variable x is said to cause y if, and only if, x invariably produces Y. That means that
when x is present y will ‘necessarily, inevitably and infallibly’ occur. For example: Water always boils at
100 degrees of Celsius. But in reality laws are never that simple, because laws are always under specific
conditions. Therefore deterministic laws are sometimes states as “all other things being equal”.
Most causal thinking in the social sciences is probabilistic. Which means that a given factor increases(or
decreases) the probability of a particular outcome. For example: Being female increases the probability
of working part-time. Probabilistic explanations can be improved by specifying conditions under which x
is less likely/more likely to affect y, but you will never achieve complete or deterministic explanations.
Under probabilistic explanations there will never be certainty about outcomes. Despite the probabilistic
nature of causal statements in the social sciences. Much of these statements are shown as deterministic
statements.
Theory testing and theory construction
Theories are attempts to answer the ‘why’ questions in social sciences. These theories vary in
complexity, amount of variables and links, abstraction and scope. To understand the role of theory in
empirical research we should distinguish two different styles of research: theory testing and theory
building.
Theory building
Theory building is a process in which research begins with observations and uses inductive reasoning to
derive a theory from these observations. These theories attempt to make sense of observation. It is
called post factum theory/ex post facto theorizing, because the theory is produced after the
observations are made. This form of theory building entails asking whether the observation is a
particular case of a more general factor, or how the observation fits into a pattern or a story.
,Theory testing
A theory testing approach begins with a theory and uses theory to guide which observations to make; it
moves from the general to the particular. The observations should test the worth of the theory. It uses
deductive reasoning to derive a set of propositions from the theory. We need to develop these
propositions in a way that if the theory is true, things in the real world should follow. We can assess
whether the theory is supported or rejected.
Although theory testing and theory building are often presented as alternative modes of research, they
should be part of an ongoing process(see picture below). Typically, theory building will produce a
plausible account or explanation of a set of observations. However, such explanations are frequently just
one of a number of possible explanations that fit the data. While plausible they are not necessarily
compelling. They require systematic testing where data are collected to specifically evaluate how well
the explanation holds when subjected to a range of crucial tests.
What is research design?
Social research needs a design or structure before data collection or analysis can commence. A research
design is not just a work plan. A work plan details what has to be done to complete the project, but the
work plan will flow from the project’s research design. The function of a research design is to ensure that
the evidence obtained enables us to answer the initial question as complete as possible. Obtaining
relevant evidence entails specifying the type of evidence needed to answer the research question, to
test a theory, to evaluate a programme of to describe some phenomenon. In other words, what type of
evidence is needed to answer the question(or test the theory) in a convincing way?
In social research the issues of sampling, method of data collection, design of questions are all subsidiary
to the matter of “what evidence do I need to collect?”.
Design versus method
Research design is different from the method by which the data are collected. How the data are
collected is irrelevant to the logic of the design. Failing to distinguish between design and method leads
to poor evaluation of designs. Equating cross-sectional designs with questionnaires, or case studies with
participant observation, means that the designs are often evaluated against the strengths and
weaknesses of the method rather than their ability to draw relatively unambiguous conclusions or to
select between rival plausible hypotheses.
,2. Hair, J.F. Jr., Anderson, R.E., Tatham, R.L. (1987) Multivariate Data
Analysis, second edition, pp. 20-40.
1.What is multiple regression analysis?
Multiple regression analysis is a statistical technique that can be used to analyze the relationship
between a single dependent variable(criterion) and several independent variables(predictors). The
objective is to predict the value of the dependent variable with the independent variables, of which the
values are known.
To use multiple regression analysis: (1) The data must be metric, and (2) before deriving the regression
equation, the researcher must decide the dependent and independent variable(s).
For what purposes is regression analysis used?
1.Determine the appropriateness of using the regression procedure with the problem.
2.Examine the statistical significance of the attempted prediction.
3.Examine the strength of the association between the dependent and independent variables.
4.Predict the values of one variable from the values of others.
Adding a predictor is useful when it gives more information about the dependent variable.
Types of regression
There are three types of regression. (1) Prediction using a single measure, the average, (2) Prediction
using two measures, simple regression (3) Prediction using several measures, multiple regression.
Prediction using a single measure, the average
Simply using the average of a number of measures to answer a question. The way to assess the adequacy
of using the average as predictor is to examine the errors that are made when it is used. For example the
average is 4 and a someone has the value 2, the average method has an error of +2. The sum of the
errors will always be 0, so you have to square the errors, and obtain the sum of the squared errors. It
provides a good measure of the prediction accuracy of the arithmetic average. For a single set of
observations, no other measure of central tendency will produce a smaller sum of squared errors than
the arithmetic average.
Prediction using two measures, simple regression
Also tries to minimize the sum of the squared errors. If we have additional information, we could
improve our prediction. If we know that the dependent variable is related to an independent variable we
could write a relationship: . Y^ = Predicted value, B0= the intercept, B1= the slope and X1=
the x value to predict y value.
Functional relationship: Calculates an exact value.
Statistical relationship: Estimates an average value.
, 2.Major assumptions of simple regression
-Statistical relationship: We assume that the data we are working with is statistical data.
-Equal variance of the criterion variables: We assume that at each level of the predictor variable the
values of the criterion variable have the same variance(homoscedasticity), because our prediction is
based on actual values from our predictions. It means that the variance of Y at x1.x2, etc. are all equal. If
the variance isn’t equal, the predictions will be better at some levels of the criterion variable. We could
be misled into thinking that our predictions are equal at all levels, when some levels are equal(the ones
you have calculated).
-Lack of correlation of errors: We want to find that any errors we make in prediction are uncorrelated
with each other.
Fixed predictor= The levels of the predictor variables are fixed.
Random predictor = The levels of the predictor variables are selected at random. Our interest isn’t just in
the levels examined but rather in the larger population of possible predictor levels from which we
selected a sample.
3.Prediction using several measures, multiple regression analysis
You add another predictor variable to improve your predictions. Goal is to have the lowest sum of
squared errors as possible.
On top of the assumptions of simple regression, you must be concerned with possible interaction and/or
correlation among the predictor variables, because we have assumed that they don’t interact and are
uncorrelated.
