Quantitative Research Concepts and Strategies

1
Quantitative Research Concepts and Strategies
2
Quantitative research strategies are driven by
two concerns.
3
Quantitative research strategies are driven by
two concerns.
Quantitative research is interested in the nature
of relationships among variables.
4
Quantitative research strategies are driven by
two concerns.
Quantitative research is interested in the nature
of relationships among variables.
Quantitative researchers are interested in
whether their discoveries are generalizable.
5
Quantitative research is interested in the nature
of relationships among variables.
Variable A
Variable B
6
The variables might be unrelated.
Variable A
Variable B
7
The variables might be correlated.
Variable A
Variable B
8
One variable might affect another.
Variable A
Variable B
9
When one variable affects another,
Variable A
Variable B
10
When one variable affects another,
they are given specific labels.
11
When one variable affects another,
they are given specific labels.
Independent Variable
Dependent Variable
12
The term quantitative refers to this research
approach because we wish to quantify these two
concepts-The size of the relationships among
variables.- The probability that the results are
generalizable.
13
-The size of the relationships among variables.-
The probability that the results are
generalizable.
14
-The size of the relationships among
variables.- The probability that the results
are generalizable.
15
-The size of the relationships among
variables.This is quantified using mathematics
- The probability that the results are
generalizable.
16
-The size of the relationships among
variables.This is quantified using mathematics
The difference in average scores between males
and females on the SAT. The correlation
between scores on an IQ test and grade point
average.- The probability that the results are
generalizable.
17
-The size of the relationships among
variables.This is quantified using mathematics
The difference in average scores between males
and females on the SAT. The correlation
between scores on an IQ test and grade point
average.- The probability that the results are
generalizable.This is quantified using
inferential statistics
18
-The size of the relationships among
variables.This is quantified using mathematics
The difference in average scores between males
and females on the SAT. The correlation
between scores on an IQ test and grade point
average.- The probability that the results are
generalizable.This is quantified using
inferential statistics There is a
statistically significant difference at the .05
level between males and females on the SAT.
19
Inferential statistics procedures actually
provide both quantities of interest for us- the
size of the relationship and the probability that
the relationship exists in the larger population
the researchers sample is meant to represent.
20

• The particular statistical procedure that is
• used depends on two things
• The number of independent and dependent
  variables.
• The level of measurement used for those variables.

21
There are four levels of measurement
22
There are four levels of measurement Nominal
Numerical values are used only as names for
different categories.
23
There are four levels of measurement Nominal
Numerical values are used only as names for
different categories. Ordinal The attributes can
be rank-ordered. However, distances between
attributes do not have any meaning.
24
There are four levels of measurement Nominal
Numerical values are used only as names for
different categories. Ordinal The attributes can
be rank-ordered. However, distances between
attributes do not have any meaning. Interval
The distances between scores have meaning and
are treated as equal. For example, when we
measure temperature, the distance from 30-40 is
equal to the distance from 70-80. The interval
between values is interpretable.
25
There are four levels of measurement Nominal
Numerical values are used only as names for
different categories. Ordinal The attributes can
be rank-ordered. However, distances between
attributes do not have any meaning. Interval
The distances between scores have meaning and
are treated as equal. For example, when we
measure temperature, the distance from 30-40 is
equal to the distance from 70-80. The interval
between values is interpretable. Ratio There is
an absolute zero that is meaningful. In social
science research most "count" variables are
ratio, for example, the number of children
eligible for special education services.
26
There are four levels of measurement Nominal
Numerical values are used only as names for
different categories. Ordinal The attributes can
be rank-ordered. However, distances between
attributes do not have any meaning. Interval
The distances between scores have meaning and
are treated as equal. For example, when we
measure temperature, the distance from 30-40 is
equal to the distance from 70-80. The interval
between values is interpretable. Ratio There is
an absolute zero that is meaningful. In social
science research most "count" variables are
ratio, for example, the number of children
eligible for special education services.
27
Group Designs

• Whether you can trust the results of
• quantitative research depends on the design
• that was used. The use of groups and group
• comparisons is a key design element that
• supports valid conclusions about the nature
• of the relationships among variables and the
• generalizability of results.

28
Group Designs

• Whether you can trust the results of
• quantitative research depends on the design
• that was used. The use of groups and group
• comparisons is a key design element that
• supports valid conclusions about the nature
• of the relationships among variables and the
• generalizability of results.

29
Validity of Quantitative Research Conclusions
Issues of Cause and Effect
Issues of Generalizability
Statistical Conclusion Validity
Internal Validity
Construct Validity
External Validity
30
Statistical Conclusion Validity
Internal Validity
Construct Validity
External Validity
31
Statistical Conclusion Validity
Is there a relationship between A B?
Internal Validity
Construct Validity
External Validity
32
Statistical Conclusion Validity
Is there a relationship between A B?
Is there a cause and effect relationship between
A B?
Internal Validity
Construct Validity
External Validity
33
Statistical Conclusion Validity
Is there a relationship between A B?
Is there a cause and effect relationship between
A B?
Internal Validity
Construct Validity
Is the cause and effect relationship between A
B?
External Validity
34
Statistical Conclusion Validity
Is there a relationship between A B?
Is there a cause and effect relationship between
A B?
Internal Validity
Construct Validity
Is the cause and effect relationship between A
B?
External Validity
Is the relationship between A and B generalizable?
35

• The particular statistical procedure that is
• used depends on two things
• The number of independent and dependent
  variables.
• The level of measurement used for those variables.

36

• The particular statistical procedure that is
• used depends on three things
• The number of independent and dependent
  variables.
• The level of measurement used for those variables.

37

• The particular statistical procedure that is
• used depends on three things
• The number of independent and dependent
  variables.
• The number of groups.
• The level of measurement used for those variables.

38
For example

• The particular statistical procedure that is
• used depends on three things
• The number of independent and dependent
  variables.
• The number of groups.
• The level of measurement used for those variables.

39
For example

• If you have 1 independent variable and 1
  dependent variable and they are both measured at
  the interval level, you look for a relationship
  by using a correlation coefficient.

40
For example

• If you have 1 independent variable and 1
  dependent variable and they are both measured at
  the nominal level, you look for a relationship by
  using a chi-square.

41
For example

• If you have 1 independent variable and 1
  dependent variable and the independent variable
  is at the nominal level and the dependent
  variable is at the interval level, you look for a
  relationship by using an independent t test.

42
For example

• If you have 1 independent variable and 1
  dependent variable and the independent variable
  is at the nominal level and the dependent
  variable is at the interval level, you look for a
  relationship by using an independent t test. But
  if the independent variable has more than 2
  groups, you use analysis of variance.

43

• And so on

44

• What I left out

45

• The variables must be measured with validity and
  reliability.
• There are some sampling methods which are better
  than others in getting a representative sample.
• Randomly assigning participants to groups solves
  a lot of problems.
• There are assumptions about how your scores are
  distributed which must be true before you can
  trust your statistical results.

46
Quantitative Research Concepts and Strategies