ANT 3402 UWaterloo Occupational Prestige IBM SPSS Statistics Analysis Project first one are the sample, rest are reqirements and reading, the datafile will

ANT 3402 UWaterloo Occupational Prestige IBM SPSS Statistics Analysis Project first one are the sample, rest are reqirements and reading, the datafile will be uploaded later Statistical Analysis Project
SOC/ANT 3402
Betty Baylor
NOTE:
You may not replicate this project! Please choose your own combination of variables.
Part 1: Choosing Your Topic and Variables
My dependent variable for this project is political views. This is an important topic for
sociologists to research because we live in a two-party system. Presidents, senators,
governors, and state representatives are elected on the basis of liberal and conservative
ideologies. Understanding who aligns with which ideology and why is the foundation of
voters and voter turnout. The trends in political views that exist represent the important
issues of particular groups and a healthy democracy must incorporate the opinions of all
people.
Dependent Variable
Political Views (ordinal)
1. Extremely liberal
2. Liberal
3. Slightly liberal
4. Moderate
5. Slightly conservative
6. Conservative
7. Extremely conservative
Independent Variables
Sex (nominal)
0. Male
1. Female
Belief in Afterlife (nominal)
0. Does not believe
1. Belief in afterlife
Highest Degree Earned (ordinal)
0. Lower than High School Degree
1. High School Degree
2. Associate’s Degree
3. Bachelor’s Degree
4. Graduate Degree
Race/Ethnicity (nominal)
1. White
2. Black
3. American Indian
4. Asian American
5. Hispanic
Research Hypotheses
H1: On average, men are more conservative than women.
H2: On average, those who believe in an afterlife are more conservative.
H3: On average, those with a higher degree are more conservative.
H4: On average, minority groups are less conservative than whites.
Part 2: Descriptive Statistics
DV: Political Views
IV: Sex
IV: Belief in Afterlife
IV: Degree
IV: Race/Ethnicity
Political Views
Female
Belief in Afterlife
Highest Degree
White
Cases
1385
1457
1298
1457
1441
Cases Missing
72
1
159
0
16
Descriptive Statistics
Mode
Median
4
4
1
1
1
1
1
–
Mean
4.07
1.62
–
Standard Deviation
1.458
1.245
–
Source: GSS 2012
Political Views: Political Views is coded 1 through 7, 1 being extremely liberal, 4 being
moderate, and 7 being extremely conservative. The median and mode are both 4 (moderate)
and the mean is 4.07. The histogram shows a slight skew to the right—the more
conservative end. 37.8% of people responded that they had moderate views. 28.7%
responded as slightly to extremely liberal and 33.5% as slightly to extremely conservative.
Range
1-7
0-4
–
Sex: Sex is coded as 0 and 1. 0 is male and 1 is female. The mode is 1. 54.3% of respondents
are female and 45.7% of respondents are male.
Belief in Afterlife: Belief in afterlife is coded as 0 and 1. 0 is no belief in an afterlife and 1 is
belief in an afterlife. The mode is 1. 19% of respondents do not believe in an afterlife and
89% of respondents do.
Highest Degree: Highest degree earned is coded 0 through 4. 0 is less than a high school
degree. 1 is high school degree. 2 is associates’ degree. 3 is bachelor’s degree. 4 is graduate
degree. The median is 1 (high school degree) and the mean is 1.62. The distribution is
skewed to the left. 62.9% of respondents have a high school degree or lower. The second
most common degree was a bachelor’s degree (19%).
Race/Ethnicity: Race and ethnicity is coded 1 through 5. 1 is white. 2 is black. 3 is
American Indian. 4 is Asian American. 5 is Hispanic. The mode is 1 (white). 75.5% of
respondents are white. The second most frequent race is black (14.8%), then Hispanic
(4.4%), Asian American (3.7%), and American Indian (1.4%).
Part 3: Estimation
Confidence Intervals
Political Views
Highest Degree
Female
Belief in Afterlife
White
Source: GSS 2012
Mean
4.07
1.62
–
Proportion
0.543
0.81
0.755
95% Confidence
Interval
Lower
Upper
4.068
4.07
1.618
1.622
0.517
0.569
0.783
0.837
0.729
0.0781
Political Views: Based on a random sample of 1,358 respondents, I estimate at a 95%
confidence level that political views average between a 4.068 and 4.072 on the scale from
extremely liberal (1) to extremely conservative (7).
Highest Degree: Based on a random sample of 1,457 respondents, I estimate at a 95%
confidence level that degrees average between a 1.618 and 1.622 on the scale from 0 (no
high school degree) to 4 (graduate degree).
Sex: Based on a random sample of 1,457 respondents, I estimate at a 95% confidence level
that between 51.7% and 56.9% of respondents will be female.
Belief in Afterlife: Based on a random sample of 1,298 respondents, I estimate at a 95%
confidence level that between 78.3% and 83.7% of respondents will believe in an afterlife
Race/Ethnicity: Based on a random sample of 1,441 respondents, I estimate at a 95%
confidence level that between 72.9% and 78.1% of respondents will be white.
Part Four: Significance Testing
A. Two-Sample t-Test
Afterlife
N
Belief
1008
Non-Belief
232
Source: 2012 GSS
Two-Sample t-Test
Mean
t-Score
4.19
3.869
3.78
Sig.
0.00
H0: There is no difference in political views between those who believe in an afterlife and
those who don’t.
HA: There is a difference in political views between those who believe in an afterlife and
those who don’t.
The mean political view of respondents who believe in an afterlife is 4.19 and the mean
political view of respondents who do not is 3.78. The alpha used for this test was .05. The
test is significant; therefore, we can reject the null hypothesis. We conclude that those who
believe in an afterlife tend to be more conservative than those who do not.
B. ANOVA
Descriptives
LT HIGH
SCHOOL
HIGH
SCHOOL
JUNIOR
COLLEGE
BACHELOR
GRADUATE
Total
THINK OF SELF AS LIBERAL OR CONSERVATIVE
N
Mean
Std.
Std.
95% Confidence
Deviation Error
Interval for Mean
Lower
Upper
Bound
Bound
199
4.11
1.537
.109
3.90
4.33
Minimu Maximu
m
m
1
7
654
4.13
1.394
.054
4.02
4.23
1
7
107
4.21
1.351
.131
3.95
4.46
1
7
274
151
1385
4.13
3.54
4.07
1.508
1.513
1.458
.091
.123
.039
3.95
3.30
3.99
4.31
3.79
4.14
1
1
1
7
7
7
ANOVA
THINK OF SELF AS LIBERAL OR CONSERVATIVE
Sum of
df
Mean Square
F
Squares
Between Groups
47.245
4
11.811
5.633
Within Groups
2893.510
1380
2.097
Total
2940.755
1384
Sig.
.000
H0: There is no difference between the political views of those with different education
levels.
HA: There is a difference between the political views of those with different education levels.
There is a significant difference between the political views of those with different education
levels. We can reject the null hypothesis. At least one group is significantly different from the
others. The sample means show that education makes an individual more conservative until
an associates’ degree, when continued education seems to make an individual more liberal.
The group with the most conservative views has associates’ degrees. The group with the
most liberal views has graduate degrees.
C. Chi-Square
Female * Belief Crosstabulation
Belief
Female .00
1.00
Total
Count
% within
belief
Count
% within
belief
Count
% within
belief
.00
134
54.3%
1.00
457
43.5%
Total
591
45.5%
113
45.7%
594
56.5%
707
54.5%
247
100.0%
1051
100.0%
1298
100.0%
Chi-Square Tests
Value
df
Asymptotic
Significance
(2-sided)
.002
.003
.002
Exact Sig. (2- Exact Sig. (1sided)
sided)
Pearson Chi-Square
9.352a
1
b
Continuity Correction
8.922
1
Likelihood Ratio
9.316
1
Fisher’s Exact Test
.003
.001
Linear-by-Linear
9.344
1
.002
Association
N of Valid Cases
1298
a. 0 cells (0.0%) have expected count less than 5. The minimum expected count is 112.46.
b. Computed only for a 2×2 table
H0: There is no relationship between gender and belief in an afterlife.
HA: There is a relationship between gender and belief in an afterlife.
Gender and belief in an afterlife are significantly correlated. We can reject the null
hypothesis. According to the column percentages, women tend to believe in an afterlife
more than men. 56.5% of women believe in an afterlife while 43.5% of men believe in an
afterlife.
Part Five: Analyzing Strength of Relationships
Correlation Matrix
Political Views
Pearson’s
Political Views
Female
1
Sig.
Female
Belief in Afterlife
Highest Degree
Race/Ethnicity
Highest
Degree
-0.078
Race/Ethnicity
0.033
Belief in
Afterlife
0.109
0.219
0.000
0.003
0.414
-0.022
N
1385
1385
1240
1385
1369
Pearson’s
0.033
1
-0.085
-0.002
-0.014
Sig.
0.219
0.002
0.95
0.590
N
1385
1457
1298
1457
1441
Pearson’s
0.109
-0.085
1
0.032
0.14
Sig.
0.000
0.002
0.242
0.613
N
1240
1298.000
1298
1298
1284
Pearson’s
-0.078
-0.002
0.32
1
-0.079
Sig.
0.003
0.950
0.242
N
1385
1457
1298
1457
1441
Pearson’s
-0.022
-0.014
0.014
-0.079
1
Sig.
0.414
0.590
0.613
0.003
N
1369
1441
1284
1441
0.003
1441
Source: GSS 2012
The significant correlations in the matrix above are bolded. There is a significant positive
relationship between political views and belief in an afterlife. There is also a significant
negative relationship between degree and political views. There is another negative
relationship between race/ethnicity and degree. Lastly, there is a negative relationship
between sex and belief in an afterlife. None of these correlations are particularly strong, the
strongest being between political views and belief in afterlife (0.109).
Bivariate Regression: Political Views and Sex
Variable
b
SE
P-value
?
(Constant)
4.119
0.058
0.000
Female
-0.097
0.079
-0.033
0.219
r2
N
Source: GSS 2012
0.001
1385
Bivariate Regression: Political Views and Belief in Afterlife
Variable
b
SE
P-value
?
(Constant)
3.780
0.096
0.000
Belief
.411
0.106
0.109
0.000
r2
N
Source: GSS 2012
0.012
1240
Bivariate Regression: Political Views and Highest Degree
Variable
b
SE
P-value
?
(Constant)
4.219
0.065
0.000
Highest Degree
-0.091
0.031
-0.078
0.003
r2
N
Source: GSS 2012
0.006
1385
Bivariate Regression: Political Views and Race
Variable
b
SE
?
(Constant)
4.115
0.45
White
Reference Category
Black
-0.261
0.113
-0.063
Native American
-0.178
0.367
-0.013
Asian
-0.315
0.211
-0.041
Hispanic
0.119
0.188
0.017
r2
N
Source: GSS 2012
0.006
1369
P-value
0.000
0.021
0.628
0.135
0.526
There is not a significant correlation between political views and sex. Nor is there a
significant correlation between political views and Hispanics, Asian Americans, or American
Indians when compared to whites. The only race with a significant correlation with political
views in comparison to whites is African American. Blacks tend be less conservative than
whites. There is a significant negative correlation between degree and political views. As an
individual becomes more educated, their political views become more liberal. The strongest
correlation is belief in afterlife and political views with a beta of 0.109. Believing in an
afterlife tends to make an individual more conservative and explains for 1.2% of the
variation in political views.
Variable
(Constant)
Female
Belief
Highest Degree
White
Black
Native American
Asian
Hispanic
r2
N
Source: GSS 2012
Multivariate Regression
b
SE
4.014
0.119
-0.078
0.084
0.457
0.107
-0.097
0.034
Reference Category
-0.261
0.113
-0.178
0.367
-0.315
0.211
0.119
0.188
?
-0.027
0.122
-0.082
P-value
0.000
0.351
0.000
0.005
-0.063
-0.013
-0.041
0.017
0.021
0.628
0.135
0.526
0.027
1240
In the multivariate regression, the significance of the variables remained the same however
the strength of the significant correlations increased slightly. The bolded variables represent
significance. Combined, these variables account for 2.7% of the variation in political views.
Conclusion
After analyzing gender, belief in an afterlife, highest degree earned, and race/ethnicity with
relationship to political views, my findings only support my second null hypothesis: On
average, those who believe in an afterlife are more conservative. These findings were
significant with a t-Test as well as a bivariate regression. Although my analysis regarding
gender was insignificant, I did find a relationship between belief in an afterlife and gender
using a chi-square test. Women tend believe in an afterlife more than men. I also
hypothesized that those with a higher degree would be more conservative. In an analysis of
variance (ANOVA), I found that those with a graduate degree and those without a high
school degree are the most liberal groups. I hypothesized that minority groups would be less
conservative than whites and after running a bivariate regression using a dummy system and
white as the reference category, I found that the only minority group that differed
significantly from whites were blacks, who tended to be more liberal.
These findings prove and disprove some popular opinions about groups of people and
political ideology. For example, my tests could not conclude a relationship between gender
and political views. Generally, people believe that women tend to be more liberal. It is also
significant to note that blacks were the only minority group that differed significantly from
whites in political views. As predicted, blacks are more liberal compared to whites. It’s also
interesting to note that liberal ideologies are strong in the lower educated and the very highly
educated while conservative views tend to trend in the middle.
Statistical Analysis Project
SOC/ANT 3402
100 Total Points
Overview:
The purpose of this project is to put everything you have learned this semester about the statistical
analysis of society into one final product. You will choose one broad research topic found in the
General Social Survey (GSS) and explore it through descriptive statistics, hypothesis testing, and
regression analysis. As we will be working on this in parts together throughout the semester, you
will receive significant feedback to make this a quality paper. Therefore, I expect the final version
should be a very polished high quality piece of research.
This project is necessary to pass this class. You must turn in a SAP in order to receive any grade
higher than an “F.”
Academic Integrity
Integrity is a distinguishing mark of Baylor students, faculty, and staff. As a result, Baylor has an
honor code by which we live:
Baylor University students, staff, and faculty shall act in
academic matters with the utmost honesty and integrity.
You will do your own work.
You may NOT work with anyone else.
Statistical Analysis Project 1
Part 1: Choosing Your Topic and Variables
Look through the two 2014 GSS codebooks, which are available on Canvas. Choose an intervalratio or ordinal dependent variable (with at least four response categories, not including those in
the “missing” category) that is of interest to you.
Next, think about what would affect your chosen dependent variable (DV). Choose 4 variables that
you believe would be possible causes for your DV. These four variables will be your independent
variables (IV).
• One IV must have between 3 and 5 response categories, not including missing variables.
• Two IVs must have only 2 categories (dichotomous), not including missing variables.
• The fourth variable can be any other one you choose.
Remember, you can recode to create these variables. In other words, a variable of interest may be
ordinal in its original form, but you could recode it to make it dichotomous (such as disagree and
strongly disagree=0 while agree and strongly agree=1).
Before you estimate any procedures on SPSS, write a paragraph about why you chose your specific
DV. Why is it important to sociologists or anthropologists to better understand this variable? If we
knew more about this topic, how could it add to our society’s wellbeing?
Next, state your chosen DV and IVs and categorize their measurement level (nominal, ordinal,
interval-ratio). Write a specific research hypothesis about how each IV affects your DV. Label these
H1, H2, H3, and H4. For example if your DV was hrs1 (How many hours did you work last week?),
then a possible hypothesis relating to gender could be:
H1: On average, men will work more hours each week than women.
Part 1 Grading: The student…
chose a dependent variable that is ordinal or interval-ratio that has
at least 5 response categories.
chose one IV with 3-5 response categories.
chose two IVs with only 2 categories (dichotomous).
chose another IV of any measurement level.
wrote a paragraph explaining how the DV is an important topic.
accurately stated each variable’s measurement level.
wrote a specific hypothesis about how each IV affects the DV.
Total Points in Part 1: 15
Statistical Analysis Project 2
Part 2: Descriptive Statistics (Chapters 2-4)
After you have chosen your DV and four IVs, analyze each variable by doing the following:
Recode any dichotomous variables so that the numerical codes are 1 and 0. Describe any other
variables you recoded and how you did so.
Create a histogram (for interval-ratio or large ordinal variables) or bar chart (for nominal or smaller
ordinal variables) for each variable to summarize the overall shape of the distribution. Charts must
be appropriately labeled and titled.
Inspect the frequency distributions and graphs and choose the appropriate measures of central
tendency and (for ordinal and interval-ratio-level variables) dispersion.
For ordinal and interval-ratio variables with many scores, check the skew by using the graph and by
comparing the mean and the median.
Create a table reporting the number of cases, number of missing cases, appropriate measure of
central tendency (mean, median, or mode), standard deviation, and range for each variable.
Write a sentence or two of description for each variable, being careful to include a description of
the overall shape of the distribution, central tendency, and dispersion. For nominal and ordinallevel variables, be sure to explain any arbitrary numerical codes.
Part 2 Grading: The student…
created a histogram or bar chart for each variable.
chose the correct measure of central tendency and dispersion.
created a table of descriptive statistics.
wrote 2-3 sentences (each) describing the variables.
Total Points in Part 2: 10
Statistical Analysis Project 3
Part 3: Estimation (Chapter 7)
The next part is to estimate what all Americans look like in terms of your variables of interest.
Construct 95% confidence intervals for the mean of each of your interval-ratio or ordinal
variables.
For your nominal variables, obtain the percentage in the sample and the confidence interval using
the one-sample t-test > Analyze, > Compare Means, > One Sample T-Test. This will work for
binary variables only (0, 1). For nominals with more than two categories, like a five-category
measure of race, you will need to turn your variable into a dummy system > Transform > Create
Dummy System. Select your variable, then in the “Main Effect Dummy Variables” box, select
“Create main-effect dummies” and type in a root name. For race you could use “race.” This will
create five different binary variables, one for each race category. If white = 1 originally, then the
new variable, race_1 will code white =1 and all other races = 0. Check the data view tab to see
these changes. “Mean difference” in the output window is the actual percent that occurs in the
sample.
Create a proper table reporting each variable’s mean/proportion and the lower and upper limits to
the 95% confidence intervals.
For each variable, write a summary sentence reporting the variable, the interval itself, the
confidence level, and the sample size. Write in plain English, as if you were reporting results in a
newspaper. Most importantly, you should make it clear that you are estimating characteristics of
the population of the entire United States. For example, a summary sentence might look like this:
“Based on a random sample of 1,231 respondents, I estimate at the 95% level that U.S. drivers average
between 64.46 and 68.22 miles an hour when driving on interstate highways.”
Part 3 Grading: The student…
calculated 95% confidence intervals for each interval-ratio and
ordinal variable (with more than 3 categories).
calculated 95% confidence intervals for each nominal …
Purchase answer to see full
attachment