Statistical Research Methods
Rahul's Noteblog Notes on Biostatistics Statistical Research Methods
Research Methods
• There are twp types of variables: independent and dependent.
• Research attempts to find a relationship between the two.
Experimental Studies
• There are two types of studies: experimental and nonexperimental (observational).
Experimental:
• The researcher conducts experiment to find a relationship between dependent and independent variables.
Non-experimental (Observational):
• Nature is allowed to take its course and observations are made.
• However, experiments can have two negative points: they may be unethical, or impractical (eg., results may take a long time to appear).
Clinical Trials
• Clinical trials are used to evaluate the effects of a treatment.
• They isolate one factor and vary all others.
• They use control groups and randomization.
Control groups:
• Patients in clinical are divided into two groups: experimental group and control group.
• Experimental group are given the new treatment, control group aren't given treatment.
• There are two types of control groups: no-treatment control groups (they are given no treatment at all), and placebo control group (they are given an inert medication).
Single-Blind Studies:
• Test subjects don't know whether they're the experimental or control group.
Double-Blind Studies:
• Neither test subjects nor researchers know whether the given test subjects are in the experimental or control group.
• To avoid confounding variables (age, previous health status, gender, race, etc.), randomization is used.
Control Group Matching:
• Pairing patients from control group to their closely related counterparts (same age, previous health status, gender, race, etc.) in the experimental group.
Stratified Randomization:
• Poplation is first divided on the basis of age, previous health status, gender, race, etc., and then equal numbers of patients from each subgroup are randomly added to control and experimental groups.
Crossover design:
• The experimental group becomes the control group (and vice-versa) after a certain washout period.
Non-Experimental Studies (Observational studies)
• These are two types: descriptive studies and analytical studies.
Descriptive:
• They just offer a simple description of the frequency of the disease.
Analytical:
• They aim to test hypothesis or provide explanations about a disease or other phenomena. They are four types: cohort, case-control, case-series, and prevalence surveys.
Cohort:
• Answer questions like, how does the disease develop? A group of people (some have been exposed, and others will eventually be exposed) are observed. This study allows us to study the natural history and incidence of the disease. Also known as follow-up or longitudinal studu because it involves prolonged and repeated observations. They are expensive and sometimes impractical.
Case-control:
• This study compares disease-free people with people who actually have the disease. This study starts with the outcome of the disease, and then looks back into the past Thus, this study is also called retrospective study. They can be done cheaply, with few subjects, and quickly.
Case-series:
• This is basically a report or presentation of a disease in a number of patients. These studies are used to present new information about patients with rare diseases.
Prevalence Surveys:
• This is a survey of the whole population, and examines the relationship between the disease and other characteristics of the polulation.
Additional Readings:
General Biostatistics
1. Inferential Statistics
2. Descriptive Statistics
3. Correlational Techniques
4. Hypothesis Testing
5. Statistical Research Methods
Random Pages:
Please Do Not Reproduce This Page
This page is written by Rahul Gladwin. Please do not duplicate the contents of this page in whole or part, in any form, without prior written permission.