Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Dissertation
Why a dissertation? Just for fulfilling the requirement for a doctoral degree program. For learning how to conduct medical research scientifically. Basic requirement for starting career as a researcher.
Types of research
Quantitative research: Aim is to measure certain variables. Estimation of population parameters. Statistical testing of hypothesis. Qualitative research: Mainly concerned with obtaining information about certain population characteristics such as socio- economic status, concepts about health, health seeking behaviour, utilization of health services.
Planning
Administrative feasibility: Check for available research
facilities in your institution or the collaborating agency. Procure administrative clearance from the ethical committee. Procure permission from the agency where the study has to be conducted. Procure Informed consent of study participants/patients. Operational feasibility in field based research studies/community based studies assess the participation and co-operation of the study population and the resources required. Financial feasibility.
a. Author index. b. Subject index. Collecting references from National and International journals.
List of some Important Internet websites:1. http://www.isoc.net 2. http://www.medlineplus.gov 3. http://www.clinicaltrials.gov 4. http://www.hon.ch 5. http://www.nlm.nih.gov 6. http://www.biosis.org 7. http://www.sis.nim.nih.gov 8. http://www.healthnet.org.za 9. http://www.pubmedcentral.nih.gov 10. http://www.reliefweb.int 11. http://www.nic.in 12. http://www.pubmed.com
Research question. Rationale and apriori hypothesis on the study. Review of literature. Aims and objectives. Preparation and validation of questionnaire, by pilot study. Material and methods. Summarization of data and statistical analysis. Appendices. Summary. Bibliography.
Sampling strategies
Why sample size? For addressing the research question scientifically and for avoiding certain biases and fallacies. Chance bias. Statistical fallacies.
Basic concepts
The Process or Technique of selecting a sample of appropriate characteristics and adequate size
Sample:
Subset of the population Total elements of the survey population, redefined according to certain specifications
Sampling Frame:
Parameters:
Sample Statistics:
Sampling error:
Efficiency:
Representativeness:
Similarity to reference
population.
Sampling Procedures
Probability Samples
Multi Phase Sampling: Sequential Sampling (Quality control): Panel Sampling:
Random Sampling
All members have equal chance of selection List all members of population Random selection: Dice, Coins, Lottery, pages of book Random tables, calculator, computer
Area Sampling
Stratification of area through maps Random selection: Dice, Coins, Lottery, pages of book Random tables, calculator, computer
Systematic Sampling
Cluster Sampling
Enumeration of sampling Units Cumulative counting. Random selection: First Cluster unit selection Each cluster 7 units selection by random. Total 30 clusters. A total of 210 units Most commonly used method
Multiphase Sampling
Collect basic data from large sample Collect details from sub sample.
Sequential Sampling
First select small sample If questions are not answered, increase the sample size. Quality control methods in industries utilize this method
Panel Sampling
Randomly select sample. Collect data at frequent intervals. Trend has to be studied. Sentinel surveillance data; Nutritional monitoring data; social changes.
Examples:
G E-
error and
F-
error
Pre-requisites
G
Design of the study:
* estimating average or proportion * Difference between two averages or proportions
Sampling Error
Disparity between the true parameter and the sample statistic:
Sampling error =
1/n
Sampling Error
Errors Sampling Non-Sampling Coverage Type of controls Processing errors Observational errors
Prevalence
If the prevalence of the condition is more; then sample size will be smaller:
Error
Less tolerable error = more the sample size:
1 1 1
Consult a statistician Refer to tables already available a. For prevalence studies b. For comparison groups Calculate a. Prevalence study( mean, proportions) no = Z2 pq / e2
Z2 = abscissa of the normal curve that cuts off an area E at the tails
1 Statistical Software.
Thank You