Job Search Data Analysis - Coursework Example

?Data Analysis Data Collection The data used for this study is drawn from two sets of data: Jobseekers Allowance claimants and Unemployed persons. Although the file downloaded contains datasets for the whole of England and Wales, this research will only extract data for the 32 cities of London. Data for unemployed persons was obtained from the Annual Population Survey (APS). The APS is conducted annually during which information on the number and percentage of people who are employed, economically active, unemployment rate and economically inactive is collected. The APS is intended to be representative of the whole of the population of the UK. The population covered by the survey is all people resident in private households and young people living away from the parental home in student halls of residence or similar institution during term time. The survey covers a target sample of at least 875 economically active persons for each Unitary Authority (UA)/Local Area District (LAD), and at least 450 economically active persons in Greater London Boroughs. The number of jobseekers is derived from the Work and Pensions Longitudinal Study (WPLS), which also contains data on jobseekers allowance claimants. The data was obtained from computer systems used to administer the benefits. Although the data downloaded covers the whole Great Britain, only figures relating to London were extracted for analysis. Both datasets were downloaded from government’s Neighbourhood Statistics website http://www.neighbourhood.statistics.gov.uk Selection Process The process of selecting the jobseeker’s allowance data is as follows: on the Neighborhood Statistics website, click on Topics, then Economic deprivation, then Jobseekers allowance claimants, choose the year and tick download on the right hand side, choose Microsoft Excel file. The file download is initiated. The process is repeated for all five years. For the unemployment data, click on topics and then proceed to Economic deprivation, Worklessness: economic activity, choose the time period and check download on the right hand side, choose Microsoft Excel file. The file download begins automatically. The process is repeated for all five years. Since the analysis only sampled data from cities in London, the findings may be skewed and hence not applicable to all other cities in the rest of Great Britain. However, one strength of the paper is that it uses data covering duration of five years and therefore takes account of any seasonal or periodic fluctuations. Objective This paper will investigate the relationship between the number of unemployed persons and the number of persons claiming jobseekers’ allowance in London region. To ensure that eligible persons are included in the survey, data pertaining to persons aged 16-64 years only was used in the analysis. Types of analyses to be conducted Analyses of the data provided will begin with an exploratory analysis followed by in-depth statistical analyses which will act as confirmatory tests to any hypothesized relationships made in the first stage of analysis. Under exploratory analyses, descriptive statistics will be computed, a relationship will also be hypothesized using a scatterplot and bar graphs. In the second phase of analysis, the equation relating the two variables will be modeled. Correlation analysis will also be used to investigate the strength of the relationship observed above. A residual plot will be used to confirm the presence or absence of outliers, the plot will also indicate whether there were any particular areas where the model greatly under or over-predicted the relationship between unemployed persons and persons seeking employment1. Detailed analysis and conclusions will also be made using a regression model. Data collected from the five year period has been consolidated in to two variables: unemployed (Unemployed Persons Count) and jobseekers (Jobseekers Allowance Claimants). EXPLORATORY ANALYSIS Descriptives of the data is as shown: A plot for the individual years covered during the study is shown below; From this graph, the number of unemployed persons was high in 2006, then fell in 2007 before rising, reaching its climax in 2009 and falling slightly in 2010. This period marks the global financial crisis of 2008-2009 which led into economic turmoil resulting into massive job losses. The recovery process began in late 2009 and this reduced the number of unemployed persons as seen in the figures from 2010. Variation in Data Since this data covers the duration between January 2006 and December 2010, it is expected that huge variation might be observed. This variation is particularly expected to increase as the duration between any two measures increases. However, the variation is expected to be uniform thus making it easier to spot any outliers. A scatterplot will be used to investigate possible variations in the data and also observe a possible linear relationship between the two variables. The scatterplot is showed below: Scatter plot of the number of jobseekers against the number of unemployed persons fitted with a best line Exploratory analysis from the scatterplot shown above shows that the data is uniform and there are no notable outliers throughout the five-year period the data was collected. However, the decision on the lack of outliers in the data cannot be made at this stage. Detailed conclusions will be made using a regression model and residual plots. The graph also shows that the two variables have a linear relationship, i.e. as the number of unemployed persons increases, the number of persons claiming allowance also increases. IN-DEPTH ANALYSIS In this section of the analysis, the following hypothesis will be tested: Null Hypothesis, H0: There is no relationship between the number of jobseekers and the number of unemployed persons Alternative Hypothesis, H1: A relationship exists between the number of jobseekers and the number of unemployed persons Various types of statistical analyses will be used in testing the hypothesis. Correlation analysis A correlation is a single value that defines the extent of linear relationship between any two variables. A positive correlation would imply that as one variable increases in magnitude, the second variable also increases correspondingly and vice versa. The more the value is close to +1 or -1, the greater the relation between the two variables. The correlation between the two variables is as shown: The correlation between unemployed persons and jobseekers is +0.8158. This value is close to 1 indicating that the relation between the two variables is strong. Since the value is positive, it is expected that as one variable increases, the second variable increases correspondingly. This finding corresponds to that obtained using a scatterplot earlier. Regression Analysis Regression analysis is used to come up with a model for computing the number of jobseekers based on the number of unemployed persons. The regression output is shown below: From this equation, the regression model is given as: Jobseekers = -138.48 + 0.568 unemployed The equation can be interpreted as follows: the number of jobseekers increases by 0.568 units for every unit increase in the number of unemployed persons. Since the number of jobseekers increases correspondingly with that of unemployed persons, the relationship can be said to be positively linear. An important conclusion can also be made based on the adjusted R-squared value. This value is computed as 0.6655, a rough estimate of R-adjusted can be obtained by squaring the correlation coefficient. The value can be interpreted as follows: the regression model explains 66.55% of the total variability in the data. This is a fairly good fit. A test can also be done to check whether the number of unemployed persons is significant in the model. For this test, the hypothesis is: Null Hypothesis, H0: ?1 = 0 Alternative Hypothesis, H1: ?1 ? 0 Where ?1 is the coefficient for unemployed persons in the model. The STATA output is shown below: The p-value from the analysis is less that the level of significance, hence we reject the null hypothesis and conclude that the number of unemployed persons is significant in the model2. From the various statistical tests and graphical methods shown above, it can be concluded that a relationship exists between the number of jobseekers and the number of unemployed persons. Residual Plot A residual plot (plot of residual versus predicted values) will be used to confirm the presence or absence of outliers as hypothesized earlier using a scatterplot. The plot will show deviations from the predicted model and hence inform us whether there were any particular areas where the model greatly under or over-predicted the relationship between the two variables. Plot of Residuals against Fitted values From this plot, it is observed that almost all residuals lie between +2000 and -2000, however, outliers are seen outside this range and this could be due to either wrong measures or the presence of additional influences that were not captured during data collection. Discussion of Findings Several statistical analyses have been conducted on the data as shown above. The main aim of the analysis was to investigate the nature of relationship between our two variables: number of jobseekers and the number of unemployed persons. Naturally, it is expected that the two variables would have a linear positive relationship. However, the presence of unknown covariates can shift the relation. It is observed that the regression model only explains 66.55% of the total variability in the data. This is indicative of the presence of additional factors, measurable or not, that affect the number of allowance claimants apart from the total number of unemployed persons. The presence of outliers in the residual plot seems to confirm this argument. This realization is important as it can inform researchers of the need to have additional measures. Factors such as gender, age, and the economic situation could greatly influence the number of the number of allowance claimants. For instance, for an ageing population, a large proportion of jobless people will be above the age of 60. For this category of people, the number of allowance claimants, or those looking for jobs will be quite low as compared to a young age group, i.e. between 25 and 45. Bibliography Cohen, J., Cohen, P., West, S., & Leona, S. A. Applied multiple regression/correlation analysis for the behavioral sciences (3rd ed.). NY: Psychology Press, 2002. Cox, D. R. Principles of statistical inference. Cambridge, New York: Cambridge University Press, 2006. Read More