Self-Esteem Among Uniformed Students and the Likelihood of Graduation Among Juniors - Research Proposal Example

Proposed Mixed-Methods Studies to Assess Self-esteem Among Uniformed and the Likelihood of Graduation Among Juniors in Type High Schools NAME COURSE SCHOOL PROFESSOR DATE Table of Contents I.Introduction 1 II.Two Studies 2 III.Research Questions and Hypotheses 2 IV.Variables 3 V.Utility of the Statistical Models in Question 5 I. Introduction The discourse about mandatory uniforms and the graduation rate in public schools requires strengthening with evidence about a positive effect on student self-concept and academic achievement. Since the late 1980s, some public schools adopted a uniform policy in order to enhance behavioral discipline and the learning environment at the primary and secondary levels. Specifically, administrators and government accepted the experience of the Long Beach (California) Unified School District that adopting uniforms greatly mitigated in-campus crime and violence. However, state and Federal laws supportive of uniform policies left a loophole requiring that parents must agree. The two poles of the debate – the desire for an orderly school climate on one hand and parental insistence on children enjoying “constitutionally-guaranteed” freedom of choice therefore needs bridging with evidence that mandatory uniforms positively correlate with academic achievement and positive self-beliefs for students. In 1988, a handful of public schools adopted a mandatory dress code. The slow yet discernible trend in the public school system accelerated in the mid-1990s owing to encouraging results in a documented district-wide pilot study and subsequent encouragement from government itself, President Clinton in the White House. In 1994, the Long Beach Unified School District (LBUSD) in California required uniforms in all its schools. A month before LBUSD was to commence implementation in the fall term, California governor Pete Wilson signed SB 1269, the School Uniform Law, thus encouraging other schools in the state contemplating a school-wide uniform policy. By the following academic year, LBUSD could show that there had been drastic drops in school violence. On the other hand, public policy favors comprehensive academic success for all up to graduation from secondary school. Around the nation as a whole, this remains an elusive goal. The human factor of differential ability aside, persistence to graduation is impeded by lack of facilities, dependence on state and federal subsidies, faculty ability, administrative competence, the environment at home, the campus and the school zone, parental support, and the immaturity of students that fosters lax discipline. Study 2 rests on the assumption that Title 1 students already disadvantaged by poverty, low self-esteem, and disciplinary infractions are further handicapped by attitudinal problems. The purpose of this study is to contribute to the body of knowledge needed to address this problem. A study of current and previous students in Title I high schools will provide insight as to the negative and positive factors that contribute to student retention and success. II. Two Studies Study A: The relationship among obligatory school uniforms, enhanced academic achievement, discipline, and self-beliefs in early to mid-adolescence. Study B: How to remedy the propensity of students in Type 1 secondary schools for inattention, delinquency, mediocre academics and dropping out before graduation? III. Research Questions and Hypotheses Study A: What is the evidence that more widespread implementation of school uniforms in K-12 will impact favorably on discipline and academic achievement? RQ1: Is average self-belief profile of students more positive in schools that have mandatory dress codes or uniforms? RQ2: How well does a policy of mandatory uniforms correlate with academic achievement, more positive self-belief, and behavioral discipline? Study B: What differentiates under-achieving students in Title 1 secondary schools? H1: Being poverty-stricken is a major hindrance. H2: Meal subsidies and other income support are meant to mitigate hardship in the family and indirectly improve the chances of doing better academically H3: Lack of familial support and absence of biological father, particularly among Hispanics and African-Americans, predispose to truancy, adolescent sex, joining teenage gangs, juvenile delinquency and crime IV. Variables For study A, the two sets of variables are psychological (self-esteem) and academic achievement. The study instruments are, respectively, the Rosenberg Self-esteem Scale and the Writing Self-Efficacy Scale, the Mathematics Self-Efficacy Scale, and the Academic Self-Description Questionnaire II. Figure 1: Conceptual Model of Relationships The mix of self-administered attitude ratings (113-item questionnaire adopted from the Arkansas High School Satisfaction Survey and the Stakeholder Survey Question Bank) among students who meet the school district risk criteria and acknowledge a desire to persist with their secondary school studies and ordinal data in effect requires the researcher to grapple with the complexities of a mixed-methods study (Creswell, 2008): 1. Perceived difficulty of coursework 2. Confidence in meeting graduation requirements 3. Perceived teacher competence, empathy and support 4. Perceived aspirational level and support from parents 5. Eagerness to succeed and underlying motivation 6. Engagement and belongingness 7. Self-esteem The bipolar Arkansas Student Satisfaction Scale is about seeing the academic situation from the student’s viewpoint. The first 15 items below concerns perceptions of quality of instruction, observations about how prepared teachers are, how receptive, caring and open. These are evidently fundamental to an ameliorative attitude on the part of students, to provoking interest and maximizing the probability of graduating in timely fashion. Teachers and Quality of Instruction Not applicable Strongly agree Agree No Opinion Disagree Strongly disagree 1. My teachers give me constructive feedback. N.A. 4 3 0 2 1 2. My classes are challenging. N.A. 4 3 0 2 1 3. My teachers are prepared to teach their subjects. N.A. 4 3 0 2 1 4. My teachers are interested in helping me learn. N.A. 4 3 0 2 1 5. My teachers are easy to talk to. N.A. 4 3 0 2 1 6. My teachers encourage class discussion. N.A. 4 3 0 2 1 7. My teachers are receptive to my concepts and ideas. N.A. 4 3 0 2 1 8. I like to ask my teachers for help. N.A. 4 3 0 2 1 9. My teachers care about me. N.A. 4 3 0 2 1 10. My teachers help me when I do not understand. N.A. 4 3 0 2 1 11. I can talk to my teachers about anything. N.A. 4 3 0 2 1 12. My teachers tell me when I do good things. N.A. 4 3 0 2 1 13. My teachers know whether I have mastered the topics covered in class N.A. 4 3 0 2 1 14. My teachers regularly tell me how I am doing N.A. 4 3 0 2 1 15. My teachers care about my future N.A. 4 3 0 2 1 V. Utility of the Statistical Models in Question In Study 1, the count of dependent variables will likely be centered on a binary (uniforms or not) DV, suggesting that the range of multiple regression choices will start with logistic regression. In addition, the variable counts may be degraded by factor analysis into just five to ten independent variables (Study 2, 113 discrete variables). This suggests that non-parametric data analysis may be required. Parametric tests of hypothesis depend on large respondent counts to estimate, say the population mean or variance. Failing those, the researcher must make unrealistic assumptions about the normality of the underlying population to invoke the Central Limit Theorem. Hence, the Study 1 case requires a 2x2 contingency design – incorporating public versus private schools imposing a uniform policy or not – to even begin to project findings to the applicable county. On the other hand, nonparametric or “distribution-free” tests can often use small samples and even ordinal data, which is what Likert rating scales usually turn out to be (Doane & Seward, 2007). Figure 2: Some Common Nonparametric Tests and Their Parametric Counterparts The decision model for a multiple regression such as that conceptualized in Figure 1 can include (see overleaf): TYPE RATIONALE Linear The oldest type, amenable to computation by hand. Sensitive to both outliers and cross-correlations (both in the variable and observation domains) and subject to over-fitting. (Granville, 2015) Logistic regression Used extensively in medical research, clinical trials, fraud detection, when the response is binary (chance of succeeding or failing, e.g. for a new tested drug or a credit card transaction). Suffers same disadvantages as linear regression (not robust, model-dependent) and computing regression coefficients. Can be well approximated by linear regression after doing the logit transform. Some versions (Poisson or Cox regression) have been designed for a non-binary response, for categorical data (classification), ordered integer response (age groups), and even continuous response (regression trees, Granville, 2015). Ridge regression Puts constraints on regression coefficients and therefore more natural. More robust version of linear regression, less subject to over-fitting and easier to interpret (Granville, 2015) Logic regression Used when all variables are binary, usually in scoring algorithms. It is a specialized, more robust form of logistic regression (useful for fraud detection where each variable is a 0 or 1 rule), where all variables have been binned into binary variables (Granville, 2015). Bayesian regression Its a kind of penalized likehood estimator, and thus somewhat similar to ridge regression: more flexible and stable than traditional linear regression. It assumes that you have some prior knowledge about the regression coefficients and the error term. Bayesian regression loosens the assumption that the error must have a normal distribution though the error must still be independent across observations. In practice, however, the prior knowledge is translated into artificial (conjugate) priors, which is a weakness of this approach (Granville, 2015). Quantile regression Used in connection with extreme events, e.g. forecasting climate change and super-hurricanes (Granville, 2015). LAD regression Similar to linear regression, but using absolute values (L1 space) rather than squares of the values (L2 space, Granville, 2015). Jackknife regression This is the new type of regression, also used as general clustering and data reduction technique. It solves all the drawbacks of traditional regression. It provides an approximate, yet very accurate, robust solution to regression problems, and works well with "independent" variables that are correlated and/or non-normal. For example, Jackknife may be applicable to the several binary and ordinal variables found in Figure 1 above. Good for black-box predictive algorithms. Approximates linear regression quite well, but it is much more robust, and works when the assumptions of traditional regression (non correlated variables, normal data, homoscedasticity) are violated (Granville, 2015). Lasso regression Similar to ridge regression, but automatically performs variable reduction and permits coefficients to be zero (Granville, 2015). Ecologic regression Consists in performing one regression per stratum, if your data is segmented into several rather large core strata, groups, or bins. The caveat of the “curse” of big data applies in this context. If you perform millions of regressions, some will be totally wrong, and the best ones will be overshadowed by noisy ones with great but artificial goodness-of-fit. This concern applies to identifying extreme events and causal relationships (e.g. school uniforms for millions of public school students, global warming, rare diseases or extreme flood modeling, Granville, 2015). References Creswell, J.W. (2008). Educational research: Planning, conducting and evaluating quantitative and qualitative research (3rd ed.). Upper Saddle River, NJ: Pearson Doane, D.P. & Seward, L.E. (2007). Applied statistics in business and economics. New York: McGrawHill/Irwin. Granville, V. (2015). 10 types of regressions: Which ones to use? Retrieved from Data Science Central, http://www.datasciencecentral.com/profiles/blogs/10-types-of-regressions-which-one-to-use. Read More