Paper Assignment #4 complete solutions correct answers key

Paper Assignment #4 complete solutions correct answers key 
I. Introduction.
Please revise/improve what you wrote for Assignment 2 (Motivation & Research Question)
II. Theory.
A) Once you have laid out your theory about the independent variable you think is the
most important explanatory factor of Y, present it in a path diagram that is consistent
with the theories/expectations you develop below.
B) Next, develop a theory and expectations for each of the paths between the
independent variables in your model, please label each of the paths in the path diagram
(1a, 1b etc) so we can follow your discussion.
Income
X3
Party ID
X4
(Republican)
Attitude regarding the
responsibility of
government for
poverty alleviation
Y
(Not its responsibility)
Race
X1
(White)
Education
X2
1a
1b 1c
1d
2a
3a
3b
4a
2c
2b
2
Note: if you don’t think there is a causal relationship between the independent variables used a
curved arrow with two heads as shown below and use a gray line.
X1
X1
1a ) For reasons that have to do with the historical development of the American state and the
institutional discrimination of non-white racial groups (i.e. African Americans, Asians) the educational
attainment of individuals belonging to these groups was hindered for many years. Since the access to
higher education was unequal for certain racial groups until relatively recently, an individual’s race may
be correlated with educational attainment (X2). I expect to find a positive relation between those who selfidentify
as white (coded 1) and education (in years)
1b ) For similar historical reasons a non-white respondent will be more likely to have a lower income
than a white respondent. Discrimination may still hinder non-whites’ upward mobility in some places
independently of their educational attainment.
1c) Citizens belonging to non-white racial groups such as African Americans may be less likely to
identify with the Republican Party and more with the Democratic Party due to this party’s platform and
key role in the Civil Rights movement.
1d) A respondent belonging to a non-white racial group (i.e. African Americans, Latinos) may be
more likely to agree that government should play an active role in reducing poverty. Due to disadvantages
suffered for many years that hindered the upward mobility of African Americans for example, they may be
more likely to agree that government should engage in redistributive policies.
X2
2a) The higher the educational attainment of a respondent, the more likely he/she will have a higher
income.
2b)The higher the educational attainment of a respondent, the more likely he/she may be inclined to prefer a
particular party (specialization!).
2c) The higher the educational attainment of a respondent, the more likely he/she will know that the
implementation of certain economic policies have had a huge impact on sectors of the economy resulting in the
loss of jobs and widespread poverty and may be inclined to think it is the government’s responsibility to help
those who became extremely poor.
X3
3a) The higher a respondent’s income is, the more likely he/she will identify with the Republican Party (because
of perceived pro upper class bias party pledges).
3b)The higher a respondent’s income is, the more likely he/she will disagree with government intervention
which may possibly increase one’s tax burden. Disagreement with a redistributive policy may be independent of
partisan identity.
X4
4a) The more an individual identifies with the Republican Party platform, the more likely he/she is to accept the
party’s views on tax policy and oppose a poverty alleviation policy that would increase taxes.
III. Measurement/ Operationalization.
A) Describe the data you will be using, the variables, the categories and how you recoded
them as you did in assignment 2. This time you will have at least 3 independent variables
and 1 dependent variable. Why are these valid measures of your conceptual variables?
3
Example: To test these hypotheses, we select the following variables from GSS 2008:
“race” (white, black, other), “educ” (highest year of school completed), “income06” (total
family income), “partyid” (political party affiliation) and “helppoor” (self placement on a
five point scale that ranges from 1=“I strongly agree the government should improve living
standards” to 5=“I strongly agree that people should take care of themselves”).1
B) State the main hypothesis:
Example: The main hypothesis I will be testing in this paper is (see path 3b) that holding
constant education, race and partisan identity, income will still have a positive effect on the level
of disagreement with redistributive policy (Y). My expectation is that controlling for education,
income and race, the higher a respondent’s income is, the more likely he/she will disagree with
government intervention which may possibly increase one’s tax burden.
Notice that you should now say something about the other variables in the model: why do think
there still may be a relationship even when you control for other variables?
STATA APPENDIX FOR PART III
1) Obtain the frequency distribution for the original four variables.
. tab race, miss
race of |
respondent | Freq. Percent Cum.
------------+-----------------------------------
white | 1,559 77.06 77.06
black | 281 13.89 90.95
other | 183 9.05 100.00
------------+-----------------------------------
Total | 2,023 100.00
. tab educ, miss
(output not shown)
. tab income06, miss
1. 309. I'd like to talk with you about issues some people tell us are important. Please look at CARD BC. Some
people think that the government in Washington should do everything possible to improve the standard of living
of all poor Americans; they are at Point 1 on this card. Other people think it is not the government's
responsibility, and that each person should take care of himself; they are at Point
4
. tab partyid, miss
.tab helppoor, miss
2) Recode the variables if necessary and obtain the frequency distribution of the
recoded variables.
In this assignment, you are advised not to collapse categories of any variable unless you
have a compelling reason to do so.
IMPORTANT: YOU MUST recode so that the category values start from
“0” while retaining the original number of categories. This makes it easy
to see what the regression equation looks like when the variable takes the
lowest value, i.e. 0.
Example:
Race (race → WHITE)
We recode this variable by reversing the order of the categories so that the larger value is
assigned to whites (“1”) because we believe being white will have a positive impact on the
dependent variable, we also join together “other” and “black” into a non-white category
which will be coded “0”.
. recode race (1=1) (2 3=0), gen(WHITE)
(464 differences between race and WHITE)
. tab WHITE
RECODE of |
race (race |
of |
respondent) | Freq. Percent Cum.
------------+-----------------------------------
0 | 464 22.94 22.94
1 | 1,559 77.06 100.00
------------+-----------------------------------
Total | 2,023 100.00
Education (educ → EDUC)
For this variable, we retain the original categories (because values begin with “0”) and treat
“dk” and “na” as missing.
. recode educ (.d .n =.), gen(EDUC)
(5 differences between educ and EDUC)
. tab EDUC
Income (income06 → INCOME)
5
We recode this variable so that the values start from “0” while retaining the original number
of categories, and recode “refused” and “dk” as missing.
. recode income06(1=0)(2=1)(3=2)(4=3)(5=4)(6=5)(7=6)(8=7)(9=8)(10=9)
(11=10)(12=11)(13=12)(14=13)(15=14)(16=15)(17=16)(18=17)(19=18)(20=19)(21=
20)(22=21) (23=22)(24=23)(25=24)(26 .d =.), gen(INCOME)
(2023 differences between income06 and INCOME)
. tab INCOME
[Output not shown]
Party identification (partyid → REPUB)
For this variable, we retain the original values and treat “7” (other party) and “na” as
missing.
. recode partyid (7 .n = .), gen(REPUB)
(51 differences between partyid and REPUB)
. tab REPUB
[Output not shown]
Views on the government’s role in improving the living standard (helppoor → GOVRES)
We recode so that the high value (“5”, recoded into “4”) is assigned to those who think that
people should help themselves and the low value, i.e. “0” is assigned to those that think that
it is the government’s responsibility to improve the living standard. Also, “dk”, “iap” and
“na” will be coded as missing.
. recode helppoor (1=0)(2=1)(3=2)(4=3)(5=4)(.d .i .n =.), gen(GOVRES)
(2023 differences between helppoor and GOVRES)
. tab GOVRES
RECODE of |
helppoor |
(should |
govt |
improve |
standard of |
living?) | Freq. Percent Cum.
------------+-----------------------------------
0 | 267 20.03 20.03
1 | 168 12.60 32.63
2 | 568 42.61 75.24
3 | 187 14.03 89.27
4 | 143 10.73 100.00
------------+-----------------------------------
Total | 1,333 100.00
3) Drop the cases with missing values in any of the five variables.
In order to make sure that the same cases will be used in all regressions in the following
steps, we tell Stata to drop observations with missing values in any of the five newly
recoded variables.
. drop if WHITE==.| EDUC==.| INCOME==.| REPUB==.|GOVRES==.
(862 observations deleted)
4) Correlations.
6
Obtain correlations between the four variables in the usual way.
The command “corr” uses only a subset of the data that has no missing values for any of the
variables listed (= listwise deletion).
. corr WHITE EDUC INCOME REPUB GOVRES
(obs=1161)
| WHITE EDUC INCOME REPUB GOVRES
-------------+---------------------------------------------
WHITE | 1.0000
EDUC | 0.0806 1.0000
INCOME | 0.1500 0.4232 1.0000
REPUB | 0.3059 0.0076 0.1478 1.0000
GOVRES | 0.2523 0.1230 0.2335 0.3456 1.000
End of STATA APPENDIX FOR PART III
5) Regress Y on all your independent variables (in this case X1, X2, X3 (and X4 if you
have one)).
IV. Statistical Analysis
THIS TABLE SHOULD BE IN THE BODY OF THE PAPER
Note that we are including the command, “beta” at the end of the multivariate regression,
this command tells STATA to standardize each coefficient which allows us to compare the
effects of the X’s on Y on the same scale (standard deviations).
. reg GOVRES WHITE EDUC INCOME REPUB, beta
Source | SS df MS Number of obs = 1161
-------------+------------------------------ F( 4, 1156) = 60.53
Model | 289.682462 4 72.4206155 Prob > F = 0.0000
Residual | 1383.12029 1156 1.19647084 R-squared = 0.1732
-------------+------------------------------ Adj R-squared = 0.1703
Total | 1672.80276 1160 1.44207134 Root MSE = 1.0938
------------------------------------------------------------------------------
GOVRES | Coef. Std. Err. t P>|t| Beta
-------------+----------------------------------------------------------------
WHITE | .4016828 .081009 4.96 0.000 .140263
EDUC | .0177234 .0115824 1.53 0.126 .045287
INCOME | .03228 .0063746 5.06 0.000 .151956
REPUB | .1653062 .0167309 9.88 0.000 .2798737
_cons | .3343746 .1566399 2.13 0.033 .
------------------------------------------------------------------------------
1) Write down the regression equation and “fill” in the direct effects of each of your
independent variables on your path diagram.
GOVRES = .3343746 + .4016828 * WHITE + .0177234 * EDUC + .03228 * INCOME +
.1653062 * REPUB
7
2) Interpret EACH coefficient as discussed in class/section (sign, statistical significance,
substantive significance and say something about low vs high categories on X in terms of
Y).
3) Interpret substantively in units of Y when your main independent variable is low/high and
all other variables are “0”.
4) Say something about which of the variables seems to be explaining the most variance (the
relative size of the beta coefficient).
5) What is R-squared telling you about the fit of your model? Use this discussion as a bridge to
write the conclusion.
V. Conclusions.
This part should include a discussion on alternative and competing explanations (Step 8) if it makes
sense you can include a discussion about actual inference or selection bias (Step 9) as well as an
actual conclusion: what did you find? Remember a non-finding is a finding!
Do you think you should drop your theory based on the results or test it in a different way?
Imagine that you were asked by your boss to conduct this study and find out if this theory was true,
could you tell him your findings in a nutshell in a professional manner (hint: you should recall your
causal story and what the regression coefficient for your main independent variable is telling you).
Income
X3
Party ID
X4
Race
X1
Education
X2
1b 1c .140263
2a
3a
.151956
.2798737
.045287
2b
1a
Government’s
Responsibility
Y