Brians data analysis in Kingston School Closure - Research Paper Example

Brian's Data Analysis in Kingston School Closure Brian Alley works as a Policy Officer in the Department of Regional Development and he has taken leave from work to complete his Master's degree in Social Science. For his Master of Arts thesis, he has gotten approval to study how rural communities deal with the decline of public services. He has chosen the proposed school closure in a town called Kingston to be his case study of rural decline. The proposed school closure is causing concern among local residents. Brian decides to do a quantitative study which includes conducting a pre and post-test of community attitudes, developing the attitude scale, trial testing the scale, conducting a community survey and data processing using SPSS. A total of 52 residents agree to take part in the study. Questions that Form the Basis of the Research The general question that Brian wanted to address in his research project was how rural communities deal with the decline of public services. To be able to address this, Brian formulated sub-topics to form the basis for his project. These are: the status of rural decline in Australia, the impact of school closures on rural communities, alienation in rural communities, design of the study, methods of data gathering and analysis, political and economic factors influencing the proposed Kingston school closure, response of Kingston residents to the proposed closure, effects of the Kingston school closure and implications of the case of the Kingston school closure on rural Australia. In my opinion, the scope of the study was comprehensive and the formulated questions need no modifications or additions. How and why are the Chosen Methods suited to answer these Questions Brian decided on doing a quantitative study for his research project. Quantitative analysis allows the researcher to focus on a limited number of variables or factors, measure them and derive relationships between them (Seaman, 2005). In Brian's study, he defined factors which he used in his survey and on which a statistical analysis was to be performed. These are: 1) whether or not the survey respondent is a parent of a secondary age child by the next year, 2) distance of residence from the Kingston school, 3) distance of residence from the Beeganup high school, 4) age of the youngest family member, 5) number of years of residence at Kingston, 6) occupation of the family member with the largest income, 7) whether or not a residential/boarding private school would be considered if the Kingston school is closed and 8) perception about the standard of education for Years 8-10 at Kingston. Additionally, Brian designed a community attitude survey with its related alienation scale and did pre and post survey using the instrument. Data gathered regarding the factors plus the pre and post attitude survey results would have bearing on the study on the impact of school closure on rural communities, alienation in rural communities, political and economic factors influencing the proposed Kingston school closure, response of Kingston residents to the proposed closure and effects of the Kingston school closure. If I was conducting the research myself, I would follow Brian's method as he had done it as well as conduct a parallel qualitative study as Peggy was doing. Key Steps in undertaking Brian's Research Project. Brian implemented a series of steps to conduct his study. He first went to Kingston to collect the town demographics. He also managed to review local newspapers where letters of concern from residents have been published related to the proposed school closure. A month later, Brian visited Kingston again and talked with local officials and townsfolk. He also conducted interviews. First, Brian interviewed Harold, the deputy principal of the school to find out the impact of school closure on school communities, particularly that of Kingston school's academic staff and parents of students. Next, he interviewed Dr Foster, an education department director to find out the effect of school closure in rural communities in general. Finally, Brian also interviewed Alan Rixon, the Shire President to find out the impact of changes in government policy and funding which lead to school closure. After his initial activities, Brian formulated a list of factors or variables and decided that he will conduct a quantitative study with the help of the SPSS program to analyze the data that he will collect. He invited community residents to take part in the study and was able to get 52 respondents in June. He incorporated the eight factors into the community survey questionnaire. Brian developed another survey questionnaire to test community attitudes toward the school closure. Initially, there were 12 items in the attitude scale but after trialling the scale with 14 friends, Brian removed 2 items to increase its reliability. Brian did a pre-test of Kingston community attitudes at the start of his research and conducted a post-test 4 months later after the Minister of Education announced that the secondary section of the Kingston school will be closed at the end of the year. Brian used SPSS on the trial data of the test he conducted on his 14 friends to form the first SPSS matrix with 12 attitude survey items. Then he again used SPSS on the pre-test attitude data that he conducted for the 52 Kingston residents. Finally, he integrated the respondent data with the pre and post-test attitude data for the 52 respondents to form the third SPSS matrix. Strengths of the Approach taken In his study, Brian wanted to find out the effects of the proposed school closure on the local residents of Kingston as a means of determining how rural communities deal with decline in public services. Since the aim of quantitative research is to determine relationships between variables or factors, the quantitative method was appropriate for testing attitude, the dependent variable against the characteristics of Kingston residents, the independent variable. Moreover, Brian chose an experimental quantitative research design where data is gathered before and after certain events happen. The pre and post-test results would show changes in attitudes or levels of alienation of the Kingston residents before and after the announcement of the closure of the secondary section of the Kingston school. The experimental design is also the appropriate design since there were only 52 respondents for the survey, as against a descriptive design which would require hundreds or thousands of subjects (Hopkins, 2000). Quantitative research employs a linear sequence model of research as it begins with definition of a problem, formulation of hypotheses, designing of instruments to test the hypotheses and finally data collection and analysis, followed by drawing of conclusions (Shimahara, 1988). This linear sequence model is a convenient way for Brian to conduct his research. Effectiveness of the Data Collection Procedures Used by Brian With regards to the data collection procedures employed by Brian, there is no mention about what sampling method he employed to select his 52 respondents. To be able to fairly accurately generalize the results of the study back to the population, random sampling should have been done. However, since Brian's study is a form of social research it may not have been practical to do random sampling so non-probability sampling was utilized, where Kingston residents were invited to take part in the study and the 52 who agreed were used as the convenient sample (Trokim, 2006). It would also have been better if the sample size was larger. The larger the sample, the wider would be the confidence interval, and the more reliable would be the results of the study (Statsoft, 2008). Results and Data Analysis The 52 participants for the study had the following characteristics: Figure 1. Parent Table 1. Parent Figure 2. Distance of Residence to Kingston School Table 2. Distance of Residence to Kingston School Figure 3. Distance of Residence to Beeganup High School Table 3. Distance of Residence to Beeganup High School Figure 4. Age of Youngest Family Member Table 4. Age of Youngest Family Member Figure 5. Number of Years of Residence at Kingston Figure 6. Occupation of Family Member with Largest Income Figure 7. Will Consider Private Boarding School Figure 8. Perception of Standard of Education for Years 8-10 As can be gleaned from the demographics of the participants to the study, there were more respondents who resided a longer distance away from the Kingston school. According to the graphs and tables stated aboveTable 2 above showsn , 58% (Fig. 2) of respondents lived from 5 to more than 20 kilometers away from the school. On the other hand, Table 3 showsn that a big number comprising 88% (Fig. 3) of the respondents lived from 21 to more than 60 kilometers away from the high school in Beeganup. This would tend to show that moving to the Beeganup high school would indeed be burdensome to the students from Kingston. An inspection of the charts Table 4 on age of youngest family member shows that a big portion, 61% (Fig. 4) of the respondents had children either approaching high school age or is of high school age. This again would tend to show that a high school is needed right within Kingston, rather than a primary school and moving high school to Beeganup would be burdensome. The distribution (Figure 5) according to years of residence at Kingston is skewed to those who have lived in the town for more than 5 years. This would tend to show again how the closure of the Kingston school would detrimentally affect established lives of the residents, especially with the added information provided by the chart on occupation (Figure 6). Figure 5. Number of Years of Residence at Kingston Figure 6. Occupation of Family Member with Largest Income The data on preference for a private boarding school on Table 5 shows that the majority of the participants would not consider a private boarding school in case the Kingston school is closed. This may be because family finances would make this unaffordable. Table 5. Will Consider Private Boarding SchoolFigure 7. Will Consider Private Boarding School The majority of the respondents are farmers whose occupation would not be easy to displace. On the perception regarding the standard of education for Years 8-10 at Kingston, the average perception is that of medium standard as shown in Table 6.Fig Figure 78. This would tend to show that the residents find high school education at Kingston acceptable, not low standard and should be discontinued. For all the factors examined, standard deviation is low, indicating a more or less homogeneous sample (Pinkney, 2008). Table 68. Perception of Standard of Education for Years 8-10 Assuming that the pre-test and post-test data are normally distributed, conducting the paired samples T-test would allow us to compare the means by computing the difference between the two and testing if the average difference is significantly different from zero (Archambault 2000). Table 79. Paired Samples Statistics Paired Samples Statistics Mean N Std. Deviation Std. Error Mean Pair 1 pretotal 34.2692 52 8.05646 1.11723 posttot 23.8462 52 8.42588 1.16846 In the descriptive statistics table above, we note that the Post-Test mean score is lower. Table 810. Paired Samples Correlations Paired Samples Correlations N Correlation Sig. Pair 1 pretotal & posttot 52 .434 .001 Checking on the correlation between the pre-test and the post-test, we see that although the correlation is positive, it is not very strong. Table 9. Paired Samples Test.11. Paired Samples Test Paired Differences Mean Std. Deviation Std. Error Mean 95% Confidence Interval of the Difference t df Sig. (2-tailed) Lower Upper Pair 1 pretotal - posttot 10.42308 8.77685 1.21713 7.97958 12.86657 8.564 51 .000 The T value = 8.564, with 51 degrees of freedom and significance is .000. Since the significance value is less than .05, there is a significant difference in the means of the pre-test and post-test. Figure 7. Scatter Plot Table 10. Pearson Correlation Correlations pretotal posttot pretotal Pearson Correlation 1.000 .434** Sig. (2-tailed) .001 N 52.000 52 posttot Pearson Correlation .434** 1.000 Sig. (2-tailed) .001 N 52 52.000 **. Correlation is significant at the 0.01 level (2-tailed). The Pearson Product Moment Correlation is the most common measure of correlation. Between variables which reflects the degree to which the variables are related. A correlation of -1 shows a negative linear relationship while +1 shows a positive relationship. 0 correlation indicates no relationship (Lane, 2008). Although there is a positive relationship between pre-test and post-test scores, the relationship is not strong as shown in Figure 7 and Table 10.. Table 11. Chi-Square Test: Parent X Distance-Kingston Chi-Square Test - Parent X Distance-Kingston Value df Asymp. Sig. (2-sided) Pearson Chi-Square 2.674a 3 .445 Likelihood Ratio 2.726 3 .436 Linear-by-Linear Association .693 1 .405 N of Valid Cases 52 a. 2 cells (25.0%) have expected count less than 5. The minimum expected count is 3.38. Table 12. Chi-Square Test: Occupation X Boarding School Chi-Square Tests - Occupation X Boarding School Value df Asymp. Sig. (2-sided) Pearson Chi-Square 5.218a 2 .074 Likelihood Ratio 5.337 2 .069 Linear-by-Linear Association 3.846 1 .050 N of Valid Cases 52 a. 2 cells (33.3%) have expected count less than 5. The minimum expected count is 3.08. Table 13. Chi-Square Test: Parent X Pre-Test Chi-Square Tests - Parent X Pre-Test Value df Asymp. Sig. (2-sided) Pearson Chi-Square 33.359a 24 .097 Likelihood Ratio 44.346 24 .007 Linear-by-Linear Association 14.977 1 .000 N of Valid Cases 52 a. 50 cells (100.0%) have expected count less than 5. The minimum expected count is .42. Table 14. Chi-Square Test: Parent X Post-Test Chi-Square Tests - Parent X Post-Test Value df Asymp. Sig. (2-sided) Pearson Chi-Square 31.288a 17 .018 Likelihood Ratio 41.851 17 .001 Linear-by-Linear Association 15.032 1 .000 N of Valid Cases 52 a. 35 cells (97.2%) have expected count less than 5. The minimum expected count is .42. Among the chi-square tests, Pearson's chi-square is most commonly used. Test results are evaluated with reference to the chi-square distribution, the resulting distribution results when independent variables with standard normal distributions are squared and summed. The chi-square tests the null hypothesis that the sample distribution is consistent with an expected distribution. A chi-square probability of 0.05 or less is usually interpreted as justification for rejecting the null hypothesis. The alternate hypothesis is not rejected when the variables have an associated relationship. From the Tables charts11-14 above, chi-square testing breaks down since expected frequencies are too low. Acceptable frequency should not be larger that 10% for expected frequencies below 5 (Plackett, 1983). Reliability Analysis Cronbach's alpha is the most widely used estimate of internal consistency of items in a scale. It measures the extent to which item responses correlate with each other. The widely-accepted cut-off is that alpha should be .70 or higher for a set of items to be considered a scale. (David Garson, 2008) In the study, Brian wanted to test "cohesion versus fragmentation" using the following items in the attitude survey: 1) You can't trust anybody any more, 3) We stick together in this community, 5) Events like school closures bring out the best in communities, 8) There is always some kind of fighting going on in the town and 10) The wealthy members of the community don't care what happens. The reliability test for these items show Cronbach's alpha at .920 as. shown in Table 15. FigureTable 15. Reliability: Cohesion vs Fragmentation Case Processing Summary N % Cases Valid 52 100.0 Excludeda 0 .0 Total 52 100.0 a. Listwise deletion based on all variables in the procedure. Reliability Statistics Cronbach's Alpha N of Items .920 5 Scale Statistics Mean Variance Std. Deviation N of Items 17.4231 28.170 5.30758 5 With regards the test on "powerlessness versus sense of control" using items: 2) We will find a way of keeping the school open, 4) The fight against the closure is not over yet, 6) There is no future for small towns like ours, 7) There is nothing we can do once the government has made up its mind and 9) People help each other in this community, Cronbach's alpa is at .720 . as shown in Table 16. Both alphas are in the acceptable values. FigureTable 16. Reliability: Powerlessness vs Sense of Control Case Processing Summary N % Cases Valid 52 100.0 Excludeda 0 .0 Total 52 100.0 a. Listwise deletion based on all variables in the procedure. Reliability Statistics Cronbach's Alpha N of Items .720 5 Scale Statistics Mean Variance Std. Deviation N of Items 16.8462 11.427 3.38036 5 Draw Inferences from Data Analysis Assuming that the samples of 52 respondents for the study are indeed representative of the population of Kingston, the proposed Kingston school closure would negatively impact the Kingston community. Firstly, the distance of the Beeganup high school from the residents of Kingston is one significant factor. As it is, the existing Kingston school is already a great distance from the residences, much more so the Beeganup school which would entail travel of more than 20 to more than 60 kilometers. A huge number of Kingston families have high school age members and these Kingston residents would really be inconvenienced by the move to Beeganup. On the question of consideration of a private boarding school, most of the respondents did not consider this possibly because of the expenses this would entail, considering that income from farming activities may not be able to afford this. Finally, the existing Kingston residents have lived in the town for a good number of years and would definitely be displaced if they had to relocate their families closer to the Beeganup high school. The post-test for the attitude survey showed a decrease in the mean, and is interpreted as an increase of pessimism or alienation within the rural community. Although the sample used was small, the findings of the study are a good basis for studying rural decline and the effects of school closures and other instances of decline in public services on rural Australia in general. Conclusion If I were to embark on a research project involving school closure, I would like to be able to identify the demographic, economic, and educational factors that lead to school closure and the magnitude of the impact that school closure will have on the local community. I would aim to create a statistical model that can be used as a tool by policymakers and school leaders to study a proposed school closure before decisions are made. The statistical model would have already been tested using data collected from communities who have already experienced school closures. In a sense, the objective of my study would be to predict future outcomes resulting from a prospective school closure rather than documenting the effects of a foregone event. Brian used the quantitative approach to conduct his study. The study may well be enhanced by using other study approaches such as the interpretive approach and action research, popular methodologies under the qualitative genre. "Qualitative research is a systematic approach to understanding qualities or the essential nature of a phenomenon, within a particular context" (Brantlinger et al., 2005). The interpretive approach focuses on understanding the way people interpret and make sense of their experiences and the world in which they live (Holloway 1997). According to Klein & Myers (1999), the assumption for interpretive research is that knowledge is gained through social constructions such as language, consciousness, and shared meanings. Interpretive research also acknowledges the relationship between the researcher and what is being studied. In terms of methodology, the objective of this research method is not to test hypotheses, but aims to understand the social context of the phenomenon and the process by which the phenomenon influences and is influenced by the social context. As defined, this approach would be very useful in studying school closure questions as it would allow the researcher to investigate the impact of the closure phenomenon on the different sectors of the community in an intimate way. Action research on the other hand is a process of repeated inquiries that involves problem solving actions which are implemented alongside data-driven research aimed at understanding the underlying causes of the problem and enabling prediction of future events (Reason & Bradbury, 2000). Action research is an iterative process that involves three stages - input, transformation and output. It begins with planning actions in the input stage - preliminary diagnosis, data gathering, feedback of results, and action planning. The second stage is the action or transformation stage, which includes actions related to learning processes, planning and executing behavioral changes. Feedback from this second stage may alter previous planning in the first stage. The third stage is the output phase which includes actual changes in behavior as a result of corrective action from the second stage. The cycle continues with adjustments to the previous stages until the output stage produces results that are satisfactory to the researcher and the subject (Lewin, 1946). Action research, therefore, if feasible to conduct would be another ideal method to employ in studying questions regarding school closure. Not only does it allow data gathering and analysis, it also provides a practical, grounded procedure for developing and testing solutions to problems through action. References Archambault, S. (2000). Paired samples t test. Retrieved June 20, 2008, from http://www.wellesley.edu/Psychology/Psych205/pairttest.html Brantlinger, E., Jimenez, R., Klingner, J., Pugach, M., & Richardson, V. (2005). Qualitative studies in special education. Retrieved June 20, 2008, from http://cache.search.yahoo-ht2.akadns.net/search/cacheei=UTF-8&p=brantlinger+exceptional+children&fr=b1ie7&u=www.cec.sped.org/AM/Template.cfm%3FSection%3DProfessional_Standards1%26Template%3D/CM/ContentDisplay.cfm%26ContentID%3D2203&w=brantlinger+exceptional+children+child&d=O_BX4i72Q9Jl&icp=1&.intl=us David Garson, G. (2008). Scales and standard measures. Retrieved June 20, 2008, from http://www2.chass.ncsu.edu/garson/pa765/standard.htm Holloway, I. (1997). Basic concepts for qualitative research. Oxford. Blackwell Science. Hopkins, W. G. (2000). Quantitative research design. Sportscience, 4(1). Retrieved June 20, 2008, from http://www.sportsci.org/jour/0001/wghdesign.html Klein, H. K., & Myers, M. D. (1999). A set of principles for conducting and evaluating interpretive field studies in information systems. MIS Quarterly, 23(1), 67-94. Lane, D. (2008). Pearson's correlation. Hyperstat Online Statistics Textbook. Retrieved June 20, 2008, from http://davidmlane.com/hyperstat/search_hyperstat.html Lewin, K. (1946). Action research and minority problems. Journal of Social Issues, 2(4). Pinkney, M. (2008). Standard deviation. Retrieved June 20, 2008, from http://www.gcseguide.co.uk/standard_deviation.htm Plackett, R. (1983). Karl Pearson and the chi-squared test. International Statistical Review, 51 (1), 59-72. Reason, P.. W., & Bradbury, H. (2000). Handbook of action research. London: Sage Publication Ltd. Retrieved June 20, 2008, from http://www.sagepub.com/refbooksProdDesc.navprodId=Book209859 Seaman, M. A. (2005). Introduction to educational research: Types of Research. Retrieved June 20, 2008, from http://edpsych.ed.sc.edu/seaman/edrm700/Questions/researchtypes.htm Shimahara, N. (1988). Anthroethnography: A methodological consideration. In Sherman, R. and Webb, R. (Eds). Qualitative Research in Education: Focus and Methods. London: Falmer. Statsoft. (2007). Electronic statistics textbook. Tulsa, OK: StatSoft Inc. Retrieved June 20, 2008 from http://www.statsoft.com/textbook/stathome.html Trokim, W. (2006). Sampling. Research Methods. Knowledge Base. Retrieved June 20, 2008 from http://www.socialresearchmethods.net/kb/sampling.htm Read More