Degree Type:
Master of PhilosophyDepartment:
Department of StatisticsModes of Study:
RegularEntry Requirements:
1. Holders of a good B.Ed degree from any recognized university, and should have taught for not less than two (2) years after completion of their degree programmes.
2. Holders of a good B.A or B.Sc degree must, in addition, hold a Postgraduate Diploma in Education(PGDE) or Postgraduate Certificate in Education(PGCE) from the University of Cape Coast or any recognized university and must have taught for not less than two(2) years after completion of their degree programme.
3. Applicants for MPhil in Educational Planning who have knowledge in Economics at the first degree level in addition to (1) and (2) above, will have an advantage.
4. MPhil in Administration in Higher Education: In addition to requirements under (1) and (2) above, applicants who have worked as administrators in Higher Educational Institutions for not less than two years will have an advantage.
5. Candidates would have to pass a selection interview
Programme Structure
Level 800
First Semester
STA 803: ADVANCED STATISTICAL METHODS
Exploratory Data Analysis: Data display, histograms, stem-and-leaf plots, box plots, data summary and description. Elementary Methods: Single-and two-sample problems, standard normal-theory tests and estimators, departures from assumptions, Poisson, Binomial and multinomial models, dispersion tests, goodness-of-fit, two-way contingency table. Regression Methods: Linear, multi-linear and polynomial regression, estimation of parameters, examination of residual, model checking. Analysis of variance: One- and two-way analyses of variance. Examination of residuals. Unbalanced case.
STA 807: ADVANCED TOPICS IN OPERATIONS RESEARCH
Formulating Linear Programming Models: Goal programming, Transportation problem, Case study. Mathematical Programming: Project planning and control, Dynamic programming,
Integer programming. Probabilistic Models: Application of queuing theory, Forecasting and simulation, Decision analysis (making hard decisions), Multi-criteria decision making.
STA 809: CLINICAL TRIALS
Organisation and Planning: Protocol, patient selection, response Justification of method for randomisation: Uncontrolled trials, blind trials, Placebo’s, ethical issues. The size of a clinical trial: Maintaining trials progress: Forms and data management, protocol deviations. Methods of data analysis: Binary responses, cross-over trials, survival data prognostic factors. Testing Hypothesis, Statistical Models: Inferential statistics-creating statistical hypothesis, the Z-test; designing a single variable experiment; errors in statistical decision making. Power of test used in clinical trials/maximizing tests power. Significance testing: t-test. Conducting two-way experiments and trials. Interpreting overall results of Clinical Trials. Nonparametric Procedures/Tests & Ranked Data: , Mann-Whitney U test, Kruskal-Willis, Friedman.
STA 815: TIME SERIES ANALYSIS
Stationary and non-stationary series: removal of trend and seasonality by differencing. Moments and auto-correlation. Models: simple AR and MA models (mainly AR(1), MA(1)): moments and auto-correlations; the conditions of stationarity: invertibility. Mixed (ARMA) models, and the AR representation of MA and ARMA models. Yule-Walker equations and partial auto-correlations (showing forms for simple AR, MA models). Examples showing simulated series from such processes, and sample auto-correlations and partial auto-correlations. (Other models, e.g., trend and seasonal). Model identification: Elementary ideas of identification of models based on simple acf and pacf showing difficulties with real series. Estimation of parameter: initial estimate based on sample acf and pacf only (least squares estimates by iterative method). Result for standard error of sample acf, pacf and estimators. Forecasting: use of the AR representation for forecasting. Minimum mean square error forecasts. Updating.
Second Semester
STA 802: PROBABILITY AND STOCHASTIC PROCESSES
Random point processes in time and space: Poisson process, inhomogeneous, compound and spatial generalizations. Review of transient and stability of random phenomena: the use of discrete time renewal theory, including the renewal theorem (without proof) with examples. Population models: discrete branching process, birth-and-death process, simple queuing models. Discrete time Markov chain: transition probabilities, classification of stages, equilibrium and absorption probabilities.
STA 804: SAMPLING TECHNIQUES AND SURVEY METHODS
The necessity and practical use of sample surveys: sample versus census, presentation and organisation of a survey. Methods of sampling. Simple random samples: techniques, estimation, choice of sample size. Ratio and regression estimators. Stratified random sampling: criteria for good stratification before or after sampling. Quota sampling. One-stage and two-stage cluster sampling. Systematic sampling. Comparison and choice of estimators. Estimation of treatment contrasts and their precision.
STA 806: MULTIVARIATE METHODS
Multivariate data summary and graphical displays. Multivariate normal distributions: Estimation of mean and covariance, one- and two-sample problems, analysis of variance.
Reduction of dimensionality: principal components and factor analyses. Discrimination and classification. Correlation; partial, multiple and canonical. Non-metric problems: clustering and scaling.
STA 810: SURVIVAL DATA ANALYSIS
Survival function, hazard functions, cumulative hazard function, censoring. Kaplan-Meier survival curve, parametric models. Comparison of two groups – log-ranked test.
Inclusion of covariates – Cox P.H. model, application of model checking. Competing risks – extensions of Cox’s model.
STA 816: DATA ANALYSIS
In this course the student will be cast in the role of a practicing statistician and become involved in projects within many fields of applications. Each of the selected projects has to be written in a report, some of which will be presented orally at seminars. Some presentations may be done jointly with other students. These are important aspects in the training of a practicing statistician, who must be able to present findings in a concise, but lucid manner, which can be, understood even by non-statisticians. The reports are continuously assessed, each project being graded and returned quickly. Some of the projects are designed to illustrate basic statistical techniques from various courses and methods and to introduce the use of the statistical packages MINITAB, GLIM, PLUM, GENSTAT AND SPSS. Others may be more open-ended or require more specialist techniques. The course opens with one or two non-assessed exercises designed to make the student familiar with the computing facilities.