Statistics Research paper

Statistical Analysis in Political Science
Name of the Student
Subject/ Session
10th November, 2015
Statistical Analysis in Political Science
Introduction
Statistics is the branch of science which deals with the collection, representation, and analysis of data. The interpretation of such data helps in recognizing various endpoints or hypothesis formulated. The field of statistics is highly appealing as it provides a robust estimation of data with reliability, viability and reproducibility. Statistics is applied to medical research, clinical trials, economics and political science, apart from other areas. Whenever there is a need for interpreting data, the science of statistics can be applied and implemented for such purpose (Healey, 177-205). The present article will evaluate the usefulness of statistical tests, in interpreting data, related to political science and economics. In this article, two case studies would be evaluated from different statistical perspectives.
Study 1
Background of the Study
Election campaigns in any country involve huge financial implications. Moreover, to control the financial implications, campaigning budgets have been cut-off as an act of law by various governments. In a general election, “Workers Party” officials wants to evaluate the rationality of their promotional budget.

For such purpose, they want to target the specific group of individuals. Since unemployment is a burning issue, the party officials want to tap the unemployed segment of the society in comparison to the employed section. If the study reveals that unemployed segment has faith in politics in the creation of jobs, the promotional campaign would actively target them. However, if they do not have faith in politics as an instrument for the creation of new jobs, the party officials would not target them for ensuing elections. Rather, they will focus on he employed a segment of the society with a new vision of better jobs.
Methodology
The study was conducted to find out the faith in politics in the creation of jobs and employment. For this study, individuals were selected based on employment status in the age range of 25-27 years. To standardize for the confounding variables only male individuals were selected for the study. Further, to ensure standardization, individuals were selected with similar educational qualifications, ethnicity, race and religion. 200 individuals were interviewed and out of them 88 met the selection criteria. The individuals were interviewed as per the criteria in Table 1.
Employment status and faith in politics were assigned dummy variable ranks. These variables were considered to be dichotomous on the trait of the variables. This is because either an individual may be employed or he may be unemployed. On the other hand either the individual might have faith in politics or the individual might not have faith in politics. The dummy variables were assigned ranks 0 and 1 as per the interview results (Table 2).
Hypothesis Testing & Statistical Analysis
The hypothesis that was intended to be tested was “whether the faith levels in politics varies between individuals in the employed sector and unemployed sector.” The null hypothesis contends that there is no difference in faith levels between the employed sector and unemployed sector. Any difference that may be noted would be considered to occur due to chance factors of random sampling. The null hypothesis will be retained if the statistical test of significance yields a ‘p’ value > 0.05. This means, out of 100 observations, more than 5 observations of happening due to chance is high and null hypothesis is retained (Sproull, 49-64).
On the other hand, the alternate hypothesis contends that there is a significant difference in faith levels between the employed sector and unemployed sector. Any difference that may be noted would be considered not to have occurred due to chance factors of random sampling. The alternate hypothesis will be retained if the statistical test of significance yields a ‘p’ value < 0.05. This means, out of 100 observations, less than 5 observations might have happened due to chance. This probability might be considered too low, and the null hypothesis would be rejected (Sproull, 49-64).
The statistical test deployed for the study was “Comparison of means” done through “z test”. This test was conducted because the dummy variables that are qualitative variables were assigned a numerical value and hence they were assumed to be a continuous variable. This means that level of faith could be estimated on a scale of 0 to 1 (Sproull, 49-64).

Parameters Sex Age Are you employed (Yes/No) If yes to above, how long you are employed Do you believe political parties help in generating employment
(Yes/No) What are your education qualifications? Any other comments Table 1: Interview Questionnaire framed for the study
Results
Data for Evaluating faith in Politics and Employment Status
Faith in Politics Dummy Rank Population Dummy Rank
Yes 1 Employed 1
Yes 1 Unemployed 0
Yes 1 Employed 1
Yes 1 Employed 1
Yes 1 Employed 1
Yes 1 Employed 1
Yes 1 Employed 1
Yes 1 Employed 1
No 0 Employed 1
No 0 Employed 1
No 0 Employed 1
Yes 1 Employed 1
Yes 1 Employed 1
No 0 Employed 1
Yes 1 Unemployed 0
No 0 Unemployed 0
No 0 Unemployed 0
No 0 Unemployed 0
No 0 Unemployed 0
No 0 Unemployed 0
No 0 Unemployed 0
Yes 1 Employed 1
Yes 1 Unemployed 0
Yes 1 Employed 1
Yes 1 Employed 1
Yes 1 Employed 1
Yes 1 Employed 1
Yes 1 Employed 1
Yes 1 Employed 1
No 0 Employed 1
No 0 Employed 1
No 0 Employed 1
Yes 1 Employed 1
Yes 1 Employed 1
No 0 Employed 1
Yes 1 Unemployed 0
No 0 Unemployed 0
No 0 Unemployed 0
No 0 Unemployed 0
No 0 Unemployed 0
No 0 Unemployed 0
No 0 Unemployed 0
Table 2: represents the dummy ranks by interview questions.

Fig1: represents the % of individuals who had faith in politics compared to % of individuals who did not have faith in politics in the employed sector.

Fig 2: represents the % of individuals who had faith in politics compared to % of individuals who did not have faith in politics in the unemployed sector.

Fig 3: The figure represents the mean value of faith between employed and the unemployed group of individuals.
Discussion and Conclusion
The above study revealed that faith levels differed significantly (p<0.004) between individuals in employed and unemployed sectors. The faith levels were higher in the employed sector, so the focus of electoral campaign would be the employed sector as per the strategic decision of the “Workers Party”.
The study suffered from certain limitations as because it was biased towards the opinion of individuals who were males, and females were not included in the study. Moreover, since it is a subjective study, the personal opinion may be biased as the study was conducted by the active supporters of “Workers Party”.
Study 2
Background
The “Workers Party” in their political agenda wants to incorporate the sensitive issues of inflation and interest rate provided by the financial instruments. They want to assure the mass that although interest rates may be decreased if they come in power, the inflation will reduce too.
Methodology
The economics cell of “Workers Party”, accessed the data on the inflation rate and interest rate of various countries through the on-line website www.tradingeconomics.com. The countries were randomly selected, and there was no chance of elemental bias. Since, the interest rates and the inflation rates are highly volatile components. Therefore, intervals were assigned to such variables as depicted in Table 3.
Person product moment correlation was calculated through appropriate software, and the regression analysis was also done to estimate the line of best fit from the scatter-plot. Correlation coefficient defines the relationship of two variables. If the relation is positive, it can be interpreted that increase in the magnitude of one variable increases the magnitude of another variable too. On the other hand, if the relation is negative it can be interpreted that increase in the magnitude of one variable decreases the magnitude of another variable too. However, a statistical test of significance can only confirm whether such correlations are significant or have happened due to chance factors of random sampling (Sproull, 49-64).
The hypothesis that was intended to be tested was “whether interest rates are correlated significantly to the inflation rates.” The null hypothesis contends that there is no significant correlation between interest rates and inflation rates. Any difference that may be noted in the correlation coefficients would be considered to have occurred due to chance factors of random sampling. The null hypothesis will be retained if the statistical test of significance yields a ‘p’ value > 0.05. This means, out of 100 observations, more than 5 observations of the correlations must have happened due to chance (Sproull, 49-64).
On the other hand, the alternate hypothesis contends that there is a significant correlation between interest rates and inflation rates. Any difference that may be noted in the correlation coefficients has not happened due to chance factors of random sampling. The alternate hypothesis will be retained if the statistical test of significance yields a ‘p’ value < 0.05. This means, out of 100 observations, in less than 5 observations the correlations must have happened due to chance. This probability might be considered too low, and the null hypothesis would be rejected. Hence, it can be interpreted that inflation rates and interest rates are highly correlated to each other (Sproull, 49-64).
Results
country Inflation rate Inflation interval Interest rate Interest intervals
United States 0 0-4 0.25 0-4
Euro Area 0 0-4 0.05 0-4
China 1.3 0-4 4.35 0-4
Japan 0 0-4 0 0-4
Germany 0.3 0-4 0.05 0-4
Brazil 9.93 5 to 9 14.25 10 to 14
Russia 15.6 10 to 14 11 15 to 19
India 5 5 to 9 6.5 5 to 9
Italy 0.3 0-4 0.05 0-4
Canada 1 0-4 0 0-4
Australia 1.5 0-4 2 0-4
Korea 0.9 0-4 1 0-4
Spain 0.9 0-4 1.5 0-4
Mexico 2.48 0-4 3 0-4
Indonesia 6.25 5 to 9 7 5 to 9
Netherlands 0.7 0-4 0.05 0 to 4
Turkey 7.58 5 to 9 7.5 5 to 9
Saudi Arabia 2.4 0-4 2 0-4
Nigeria 9.3 5 to 9 13 10 to 14
Argentina 14.3 10 to 14 21.92 20 to 24
Belgium 1.28 0-4 0.05 0 to 4
Decimal places rounded to nearest values        
Table 3: Represents the data of various countries along with interest rates and inflation rates. The table also depicts the intervals taken for both the variables. This was done to provide an allowance of minor volatility in both the variables.
Pearson Product Moment Correlation – Ungrouped Data
Statistic Variable X Variable Y
Mean4.54857142857143 3.85809523809524
Biased Variance34.5627074829932 22.0314154195011
Biased Standard Deviation5.87900565427464 4.69376346011398
Covariance26.7653621428571
Correlation0.923759137206988
Determination0.853330943573398
T-Test10.5139594725214
p-value (2 sided) 2.33354935375019e-09
p-value (1 sided) 1.16677467687509e-09
95% CI of Correlation [0.818449059899329, 0.969023822671269]
Degrees of Freedom 19
Number of Observations 21
Table 4: Reflects the correlation statistics. The correlation coefficient is 0.92. The p-value is much lesser than 0.0001.

Fig 4: reflects the line of best fit from the scatter-plot and also the regression equation indicating the relation of the inflation rate and interest rate.
Discussion and Conclusion Since the correlation coefficient was positive, it can be interpreted that decrease in interest rate will certainly bring down inflation rate and vice versa. Moreover, such observations are significant because the p-value of such correlation was <0.00001. Finally, the regression analysis indicated that there were only very few outliers and the values of interest rate may be predicted from inflation rates. Therefore, “Workers Party” would incorporate in their agenda “We will ensure a decrease in interest rates by financial instruments to stimulate business potentials and at the same time will curb the inflation rates”.
The assertion of such claims may be jeopardized if inflation increases within the interval periods taken and, on the contrary, it might happen interest rates are not increased to protect the business interests of the potential stakeholders.
Bibliography
Healy, Joseph F. The Essentials of Statistics: A Tool for Social Research (2nd ed.). (Belmont,
CA: Cengage Learning, 2009): 177–205.
Sproull, Natalie L. “Hypothesis testing”. Handbook of Research Methods: A Guide for
Practitioners and Students in the Social Science (2nd ed.). (Lanham, MD: Scarecrow
Press, 2002): 49–64.