Topics
1. Conducting Correlational Research
2. Magnitude, Scatterplots, and Types of
Relationships
3. Misinterpreting Correlations
4. Prediction and Correlation
5. Statistical Analysis: Correlation Coefficients
6. Advanced Correlational Techniques:
Regression Analysis
@ 2012 Wadsworth, Cengage Learning
Conducting Correlational
Research
@ 2012 Wadsworth, Cengage Learning
Conducting Correlational Research
• Determine whether two variables are related
to each other
– Example: do people who are taller tend to weigh
more?
• Allow us to make predictions from one
variable to another
– Example: estimate an individual’s weight based on
height
@ 2012 Wadsworth, Cengage Learning
Conducting Correlational Research
(cont’d.)
• Reasons
– Ethically impossible to do an experimental study
– Useful in measuring many variables and assessing
the relationships between them
@ 2012 Wadsworth, Cengage Learning
Magnitude, Scatterplots,
and Types of Relationships
@ 2012 Wadsworth, Cengage Learning
Magnitude, Scatterplots, and
Types of Relationships
• Magnitude: an indication of the strength of
the relationship between two variables
Table 6.1 Estimates for Weak, Moderate, and Strong Correlation Coefficients
@ 2012 Wadsworth, Cengage Learning
Magnitude, Scatterplots, and
Types of Relationships (cont’d.)
• Scatterplot: a figure that graphically
represents the relationship between two
variables
Figure 6.1 Scatterplot for height and weight
@ 2012 Wadsworth, Cengage Learning
Magnitude, Scatterplots, and
Types of Relationships (cont’d.)
Figure 6.2 Possible types of correlational relationships:
(a) positive; (b) negative; (c) none; (d) curvilinear
@ 2012 Wadsworth, Cengage Learning
Magnitude, Scatterplots, and
Types of Relationships (cont’d.)
Relationships Between Variables
@ 2012 Wadsworth, Cengage Learning
Misinterpreting Correlations
@ 2012 Wadsworth, Cengage Learning
The Assumptions of
Causality and Directionality
• Causality: the assumption that a correlation
indicates a causal relationship between the
two variables
• Directionality: the inference made with
respect to the direction of a causal
relationship between two variables
Example: FDI and economic growth rate: in
which direction ??
@ 2012 Wadsworth, Cengage Learning
The Third-Variable Problem
• The problem of a correlation between two variables being
dependent on another (third) variable
• Partial correlation
– A correlational technique that involves measuring three
variables and then statistically removing the effect of the
third variable from the correlation of the remaining two
variables
For example: Birth control and number of electrical appliances
<==wealth/income
GNI per capital and cultural individualism: political ideology
and legal system (partial correlation) multiple regression
analysis
@ 2012 Wadsworth, Cengage Learning
Restrictive Range
• A variable that is truncated and has limited
variability
Figure 6.3 Restricted range and correlation
@ 2012 Wadsworth, Cengage Learning
Curvilinear Relationships
• When a correlation coefficient does not
adequately indicate the degree of relationship
between the variables
@ 2012 Wadsworth, Cengage Learning
Misinterpreting Correlations
(cont’d.)
Misinterpreting Correlations
@ 2012 Wadsworth, Cengage Learning
Prediction and Correlation
@ 2012 Wadsworth, Cengage Learning
Prediction and Correlation
• Example:
– Strong positive correlation between smoking and
cancer
– You may know someone who has smoked for 30
years and does not have cancer
• Person-who argument
– Arguing that a well-established statistical trend is
invalid because we know a “person who” went
against the trend
@ 2012 Wadsworth, Cengage Learning
Statistical Analysis:
Correlation Coefficients
@ 2012 Wadsworth, Cengage Learning
Statistical Analysis:
Correlation Coefficients
• Pearson product-moment correlation
coefficient (Pearson’s r)
– The most commonly used correlation coefficient
when both variables are measured on an interval
or ratio scale
• Formula for Pearson’s r:
@ 2012 Wadsworth, Cengage Learning
Statistical Analysis:
Correlation Coefficients (cont’d.)
Table 6.4 Computational Formula for Pearson’s Product-Moment Correlation Coefficient
@ 2012 Wadsworth, Cengage Learning
Correlation Coefficients
X Z1 (X-
Obs (WEIGHT) Y(HEIGHT) M)/σ Z2 Z1*Z2
1 100 60 -1.619 -1.61926 2.621577
2 120 61 -0.96154 -1.40044 1.346575
3 105 63 -1.45464 -0.9628 1.400524
4 115 63 -1.1259 -0.9628 1.084021
5 119 65 -0.99441 -0.52516 0.522229
6 134 65 -0.50131 -0.52516 0.263273
7 129 66 -0.66568 -0.30635 0.203928
8 143 67 -0.20546 -0.08753 0.017983
9 151 65 0.057528 -0.52516 -0.03021
10 163 67 0.452005 -0.08753 -0.03956
11 160 68 0.353386 0.131291 0.046396
12 176 69 0.879356 0.350109 0.307871
13 165 70 0.517751 0.568928 0.294563
14 181 72 1.043721 1.006565 1.050573
15 192 76 1.405325 1.881838 2.644595
16 208 75 1.931295 1.66302 3.211782
17 200 77 1.66831 2.100656 3.504547
18 152 68 0.090401 0.131291 0.011869
19 134 66 -0.50131 -0.30635 0.153576
20 138 65 -0.36982 -0.52516 0.194217
Mean (M) 149.25 67.4 18.81032sum
Standard
Deviation 30.42 4.57 0.940516r=(Z1*Z2)/N
@ 2012 Wadsworth, Cengage Learning
Statistical Analysis:
Correlation Coefficients (cont’d.)
• Coefficient of determination (r2)
– A measure of the proportion of the variance in
one variable that is accounted for by another
variable
– Calculated by squaring the correlation coefficient
@ 2012 Wadsworth, Cengage Learning
Statistical Analysis:
Correlation Coefficients (cont’d.)
• Spearman’s rank-order correlation coefficient
– The correlation coefficient used when one (or
more) of the variables is measured on an ordinal
(ranking) scale
• Point-biserial correlation coefficient
– The correlation coefficient used when one of the
variables is measured on a dichotomous nominal
scale and the other is measured on an interval or
ratio scale
@ 2012 Wadsworth, Cengage Learning
Statistical Analysis:
Correlation Coefficients (cont’d.)
• Phi coefficient: the correlation coefficient
used when both measured variables are
dichotomous and nominal
@ 2012 Wadsworth, Cengage Learning
Statistical Analysis:
Correlation Coefficients (cont’d.)
Correlation Coefficients
@ 2012 Wadsworth, Cengage Learning
Advanced Correlational
Techniques: Regression Analysis
• Regression analysis
– A procedure that allows us to predict an
individual’s score on one variable based on
knowing one or more other variables
• Regression line
– The best-fitting straight line drawn through the
center of a scatterplot
– Indicates the relationship between the variables
height and weight for this group of individuals
@ 2012 Wadsworth, Cengage Learning
Regression Analysis (cont’d.)
Figure 6.4 The relationship between height
and weight with the regression line indicated
@ 2012 Wadsworth, Cengage Learning
Regresion Results
. reg lngni IDV wgi_avg
Source SS df MS Number of obs = 119,175
F(2, 119172) > 99999.00
Model 118305.471 2 59152.7353 Prob > F = 0.0000
Residual 26093.9716 119,172 .218960591 R-squared = 0.8193
Adj R-squared = 0.8193
Total 144399.442 119,174 1.21166901 Root MSE = .46793
lngni Coef. Std. Err. t P>t [95% Conf. Interval]
IDV .0021468 .0000565 38.01 0.000 .0020361 .0022575
wgi_avg 1.491532 .0024 621.48 0.000 1.486828 1.496236
_cons 8.506387 .0033822 2515.04 0.000 8.499758 8.513016
@ 2012 Wadsworth, Cengage Learning
Regression Results
sum lngni IDV wgi_avg
Variable Obs Mean Std. Dev. Min Max
lngni 119,175 10.15396 1.100758 7.052728 11.37099
IDV 119,175 61.04044 27.52515 13 91
wgi_avg 119,175 1.01676 .6477148 -1.104992 1.841615
@ 2012 Wadsworth, Cengage Learning
Hypothesis Testing
@ 2012 Wadsworth, Cengage Learning
Null and Alternative Hypotheses
• Null hypothesis
– The hypothesis predicting that no difference exists
between the groups being compared
– Statistical notation:
• The hypothesis predicting that no relation exists
between factors (regression): ϐ=0
@ 2012 Wadsworth, Cengage Learning
Null and Alternative Hypotheses
(cont’d.)
• Alternative hypothesis (research hypothesis)
– The hypothesis that the researcher wants to
support, predicting that a significant difference
exists between the groups being compared
– Statistical notation:
- There is a significant relationship between the A
and B: ϐ≠0
@ 2012 Wadsworth, Cengage Learning
One- and Two-Tailed
Hypothesis Tests
• One-tailed hypothesis (directional hypothesis)
– An alternative hypothesis in which the researcher
predicts the direction of the expected difference
between the groups
– Null hypothesis:
@ 2012 Wadsworth, Cengage Learning
One- and Two-Tailed
Hypothesis Tests (cont’d.)
• Two-tailed hypothesis (nondirectional
hypothesis)
– An alternative hypothesis in which the researcher
predicts that the groups being compared differ
but does not predict the direction of the
difference
– Statistical notation:
@ 2012 Wadsworth, Cengage Learning
Type I and II Errors in
Hypothesis Testing
• Type I error: an error in hypothesis testing in
which the null hypothesis is rejected when it is
true
• Type II error: an error in hypothesis testing in
which there is a failure to reject the null
hypothesis when it is false
@ 2012 Wadsworth, Cengage Learning
Type I and II Errors in
Hypothesis Testing (cont’d.)
Table 7.1 The Four Possible Outcomes in Statistical Decision Making
@ 2012 Wadsworth, Cengage Learning
Statistical Significance and Errors
• Statistical significance
– An observed difference between two descriptive
statistics (such as means) that is unlikely to have
occurred by chance
• Which type of error is considered more
serious by researchers?
– Type I error
@ 2012 Wadsworth, Cengage Learning
Single-Sample Research and
Inferential Statistics
• Single-group design: a research study in which
there is only one group of participants
• Inferential statistics: procedures for drawing
conclusions about a population based on data
collected from a sample
@ 2012 Wadsworth, Cengage Learning
Single-Sample Research and
Inferential Statistics (cont’d.)
• Parametric test
– A statistical test that involves making assumptions
about estimates of population characteristics, or
parameters
• Nonparametric test
– A statistical test that does not involve the use of
any population parameters
• μ and σ are not needed
• The underlying distribution does not have to be normal
@ 2012 Wadsworth, Cengage Learning
Single-Sample Research and
Inferential Statistics (cont’d.)
Inferential Statistical Tests
@ 2012 Wadsworth, Cengage Learning
Probability and the
Standard Normal Distribution
Figure 7.3 Area under the standard normal curve.
@ 2012 Wadsworth, Cengage Learning
The Standard Normal Distribution
(cont’d.)
Figure 7.4 Standard normal curve with z = +1.27 indicated.
@ 2012 Wadsworth, Cengage Learning
The Standard Normal Distribution
(cont’d.)
Figure 7.5 Standard normal curve with z = - 1.0 indicated.
@ 2012 Wadsworth, Cengage Learning
The Standard Normal Distribution
(cont’d.)
Figure 7.7 Standard normal curve with z = - 1.33 and z = + 1.67 indicated.
@ 2012 Wadsworth, Cengage Learning
Advanced Correlational
Techniques: Regression
Analysis
@ 2012 Wadsworth, Cengage Learning
Summary
• Calculate basic probabilities
• Use the multiplication and addition rules
• Null and alternative hypotheses
• One- and two-tailed hypothesis tests
• Type I and Type II errors
• Single-group design: the performance of a
sample is compared with that of the general
population
@ 2012 Wadsworth, Cengage Learning