联系我们： 手动添加方式: 微信>添加朋友>企业微信联系人>13262280223 或者 QQ: 1483266981

EXAMINER: Mariaelisa Epifanio
DEPARTMENT: Politics
1 | P a g e
Module Code: POLI530
PAPER CODE NO.
POLI264
LGY336
JANUARY 2026 EXAMINATIONS
Module Code: POLI264
Module Title: Introduction to Data Science
Duration: 24-hour online exam
_____________________________________________________________________
INSTRUCTIONS TO CANDIDATES
This examination consists of two exercises, Exercise 1, worth 70 points, and Exercise
2, worth 30 points, with a total of 100 points available. Exercise 1 consists of six
questions: the first two questions are worth 15 points each; the remaining four
questions are worth 10 points each. Exercise 2 consists of three questions worth 10
points each. The total word limit for the whole exam is 2,000 words maximum.
Throughout both exercises, please remember to show all your calculations where
required, as partial credit may be awarded for answers that do not report the
calculations you carried out.
Module Code: POLI530 2 | P a g e
EXERCISE 1
A researcher is interested in investigating the relationship between foreign aid
dependence and corruption levels in developing countries. The central research
question asks: Does greater reliance on foreign aid increase corruption in recipient
countries
The researcher seeks to test two competing theoretical perspectives. The “resource
curse” perspective suggests that aid, like wealth from natural resource, may weaken
institutional accountability by reducing governments’ dependence on domestic
taxation, thereby increasing opportunities for rent-seeking and corruption.
Conversely, the “conditionality” perspective argues that aid donors impose
governance requirements that may reduce corruption over time.
The dependent variable measures perceived corruption levels in the public sector.
Higher values indicate greater corruption. The continuous index is constructed from
expert assessments and survey data, ranging from 0 (no corruption) to 10 (highly
corrupt).
Table 1 presents the estimates of an ordinary least squares (OLS) regression with three
successive specifications. Model 1 includes only the main explanatory variable
(Foreign aid % GNI) to establish the baseline bivariate relationship. The variable
measures total official development assistance received as a percentage of gross
national income. Model 2 adds the first set of theoretically important controls:
democracy and natural resources, which address alternative explanations for
corruption. Democracy score is a continuous index measuring the quality of
democratic institutions and practices, ranging from 0 to 10, with higher scores
indicating stronger democracy. Natural resources (% of GDP): Total natural resource
rents as a percentage of GDP. Model 3 presents the full specification including all
control variables. Here the researcher controls for Press freedom, an index measuring
the degree of freedom available to journalists and media outlets, ranging from 0 (no
freedom) to 100 (complete freedom); GDP per capita in thousand US dollars. Trade
openness measures the sum of exports and imports as a percentage of GDP;
Government size (% of GDP) measures government final consumption expenditure as
a percentage of GDP. Finally, the researcher includes two binary variables, British
colonial history coded 1 if the country was formerly a British colony, 0 otherwise, and
Island nation, coded 1 if the country is geographically an island nation, 0 otherwise.
The sample includes 134 developing countries from 1995-2022 (2,680 observations
due to randomly missing values in some years). The results of this analysis are
displayed in Table 1. The formula for the full model specification (Model 3):
= β0 + β1 + β2 + β3
+ β4 + β5 + β6
+ β7 + β8 + β9
+
Module Code: POLI530 3 | P a g e
Table 1: Foreign aid and corruption
Model 1 Model 2 Model 3
Foreign aid (% GNI) 0.142***
(0.018)
0.078**
(0.024)
0.023
(0.021)
Democracy score — -0.198***
(0.028)
-0.156***
(0.022)
Natural resources (% GDP) — 0.041***
(0.009)
0.034***
(0.008)
Press freedom — — -0.028***
(0.006)
GDP per capita — — -0.234***
(0.045)
Trade openness — — -0.012**
(0.004)
Government size (% GDP) — — 0.019**
(0.007)
British colonial history — — -0.456***
(0.089)
Island nation — — 0.234*
(0.112)
Constant 4.623***
(0.234)
5.892***
(0.389)
6.789***
(0.567)
R-squared 0.18 0.42 0.64
Observations 2,680 2,680 2,680
Notes: Standard errors in parentheses. * p < 0.05, ** p < 0.01, *** p < 0.001. The dependent variable is the Corruption Index (0-10 scale). 1. Compare the Foreign aid coefficient across the three models. What does this pattern suggest about the original relationship between aid and corruption in Model 1 Carefully justify your answer. (15 points) 2. The coefficient on Democracy score in Model 3 is -0.156***. Provide a complete interpretation of this coefficient, including its magnitude, direction, and statistical and substantive significance. (15 points) 3. Based on Model 3 results, calculate the predicted corruption level for a hypothetical country with the following characteristics: Foreign aid = 5% of GNI, Democracy score = 6, Natural resources = 10% of GDP, Press freedom = 50, GDP per capita = $8,000, Trade openness = 80%, Government size = 18% of GDP, British colonial history = 1, Island nation = 0. Show your work and interpret the result. (10 points) A country is considering democratic reforms that would increase its Democracy score from 3 to 7, while simultaneously improving Press Freedom from 30 to 70. Using the OLS results from Model 3 in Table 1, and assuming all other variables remain constant: 4. Calculate the predicted change in corruption from the democracy improvement alone. Show your calculation and provide a brief interpretation of the predicted change. (10 points) Module Code: POLI530 4 | P a g e 5. Calculate the predicted change in corruption from the press freedom improvement alone. Show your calculation and provide a brief interpretation of the predicted change. (10 points) 6. What is the total predicted change in corruption from both reforms combined Is the change substantive (10 points) EXERCISE 2 To examine the relationship between foreign aid and corruption more rigorously, the researcher adds Fixed Effects to the specification in Model 3. Table 2 presents the results of this analysis. The first column reports the simple OLS (Model 3, Table 1). The second column presents Fixed Effects estimates. Table 2: Foreign aid and corruption: OLS vs Fixed effects OLS model Fixed effects model Foreign aid (% GNI) 0.023 (0.021) 0.008 (0.018) Democracy score -0.156*** (0.022) -0.089** (0.031) Natural resources (% GDP) 0.034*** (0.008) 0.028** (0.010) Press freedom -0.028*** (0.006) -0.015* (0.007) GDP per capita -0.234*** (0.045) -0.167** (0.062) Module Code: POLI530 5 | P a g e Trade openness -0.012** (0.004) -0.009* (0.005) Government size (% GDP) 0.019** (0.007) 0.014* (0.007) British colonial history -0.456*** (0.089) (dropped) Island nation 0.234* (0.112) (dropped) Constant 6.789*** (0.567) 8.234*** (0.823) R-squared 0.64 0.81 Observations 2,680 2,680 Countries 134 134 Country fixed effects No Yes Notes: Standard errors in parentheses. * p < 0.05, ** p < 0.01, *** p < 0.001. The dependent variable is the Corruption Index (0-10 scale). 1. Why are British colonial history and Island nation dropped from the Fixed Effects model (10 points) 2. Provide some possible explanations as for why the Natural resource coefficient shrinks from 0.034 to 0.010 and is no longer statistically significant in the Fixed Effects model. (10 points) 3. The 95% confidence interval for the Democracy Score coefficient in the Fixed Effects model is [-0.150, -0.028]. Interpret this confidence interval (10 points) END OF PAPER