SUNY College at Buffalo
Sociology 301: Social Statistics
Study Guide for Exam #4
Dr. ZHANG Jie
**Chapter 9
Chi-Squares
Non-parametric tests
normal distribution not
interval level of data not assumed
Comparing chi-square with t and F tasts
normality
interval level of measurement
categorical level of measurement
the power of rejecting a true null hypothesis
One-way chi-square tests
concept
formula
observed frequencies
expected frequencies
df = k - 1
alpha
the obtained chi-square
the critical value of chi-square
rejecting or accepting the null hypothesis
the research hypothesis: at least one different
Two-way chi-square tests
comparing tow or more independent samples
the cross tabulation
formula
df = (r-1) (c-1)
the calculation of expected frequencies
Requirements for using two-way chi-squares
comparing two or more samples
nominal data
random sampling
expected frequency in each cell larger that 5
**Chapter 10
Correlation
Pearson's correlation coefficient r
degree of association (strength: 0 through 1)
direction of association (+ and -)
Types of correlation
positive
negative
curvilinear
zero
Scatter plot (scatter gram)
the X axis: independent variable
the Y axis: dependent variable
the scatteredness of the points
The importance of scatter plot
Requirements for using Pearson's Correlation Coefficient
1. Linearity
2. interval data
3. random sampling
4. normal distribution (if the sample size < 30)
Partial correlation
control variable
r xy.z or r ab.c
**Computer SPSS Application
Frequencies
percentages
central tendencies (mode, median, mean)
variability (range, variance, standard deviation)
t test: compare two means
F test: ANOVA, compare three or more means
MAVOVA, compare three or more means for three or more variables
Chi-Square: compare expected and observed frequencies for categorical
variables
r test: correlation test, compare two continuous variables
Bivariate correlation
Understanding the matrix in multiple correlation
Understanding the P value
END