Week 2 Flashcards
What are the Classifications of study design?
Descriptive study designs-
Describe the relationship between exposures (independent variable/IV) and outcome (dependant variable/DV)
eg case study, case report and series,cross sectional etc
Analytic study designs-
Analyse the relationships between the IV and DV. The process includes correlation, association & causation
Observational Analytic designs-
-Cross sectional
-Ecological/correlational
-Cohort
-Case control
Intervention/Experimental Analytic Design
-Randomised design (Randomised Controlled Trialls/RTC)
-Non-Randomised design or Quasi-experimental design
How do you write a research question for Intervention studies
PICO – Used for intervention studies
P=Population (what group of people do you need to look at to answer the question)
I=Intervention (what are you wanting to learn the effect of? What is your choice of exposure?)
C=Control (or comparison) group
O=Outcome – It is not the result of the study, it is the something that is used to see if the intervention had an effect or not
Example: Does Magnesium supplement reduce the frequency of leg cramps among older adults?
Magnesium supplement- Intervention (I)
older adults – Population (P)
frequency of leg cramps – Outcome (O)
How do you write a research question for an observational study?
PECO- Used for Observational studies
P=Population (what group of people do you need to look at to answer the question)
E=Exposure (what variable determines the outcome?)
C=Control (or comparison) group
O=Outcome – It is not the result of the study, it is a measurement that is used to see if the intervention had an effect or not
What are the 2 common measures used in morbidity testing?
Prevalence, Cumulative incidence, and Incidence Rate
What is the difference between sample and sampling?
Sample= Subset of a population
Sampling=Process of how we choose a subset of a population
What is a probability sample?
The probability of selection is known for the participant. When this is known and equal throughout the sample it reduces bias.
Probability samples meet the requirements of the tests and statistical inference. Inferencial statistics are created about a population using data, questionaires, etc. to create the sample
What is Sampling error?
error (also called chance) – while great effort is made to generate participants who meet inferencial statistics sometimes choices for samples are incorrect or inappropriate
What are the 2 categories of samples?
Probability samples- Random sample
Non-Probability- Non random sample
What considerations should be made when choosing a sample?
-Nature of research question
-Availability of good sampling frame
-time
-Cost
-Level of accuracy required in the sample’s estimate
-Method by which data is stored and collected
What are the Probability sample types and what are their characteristics?
Simple random sample- Every subjest has an equal opportunity/ probability to participate in the sample.
-most representative of the population
-requires extensive sampling frame/list
-not cost effective, high effort
Systematic random sample- Selection of specific items in a series according to some predetermined sequence
-representative of population
-requares sampling frame
-if there is a repeating pattern in characteristics it may over or under represent characteristics
Stratafied Random sample- subdividing population into smaller homogenous groups (strata) to get more accurate representation.
Population is broken down into non overlapping groups (ie male, female, urban rural, health variables) then simple random sample is extracted from each
-more representative esp. for sub groups
-estimate is more precise
-if number is small in stratum there is loss of precision
Cluster sample- Sampling of naturally occurring clusters ie schools, households, suburbs
Single stage sampling- Target population is divided into clusters then either all units in clusters are chosen to sample or random clusters are chosen as units of sample
Two stage cluster sampling - random school chosen, random class chosen and all students within classes are sampled
Multistage ccluster sample: select States>town>School>class
-Low cost
-when details of population aren’t known this can aid in finding samples
-typically requires bigger sample size to get unbiased results
What are the non probability samples
Convenience samples- selecting most convenient available persons
Quote sample-Stratum are decided apon as well as those that fit the criteria – it is not random selection
Purposive sample-People are selected according to those who best fit the study. Commonly used in qualitative studies
Snowball sample- a person who suits the testing criteria is interviewed then they are further asked for recommendations of subjects who may fit the sample
What does Validity and Reliability mean?
Validity- accuracy or correctness
Reliability- are results repeatable, is data consistent or precise
What are the 2 types of measurement errors?
Random error -
-it is a type of error governed by chance
-Occurs in every study die to normal and natural variation
-Causes unpredictable consequences
-Small random error means good reliability
Systematic error (bias)-
-an error caused by bias in a systematic way ie. An subject with unknown undesirable characteristic,effecting test results.
-no systematic error means good validity
What are the 3 types of bias?
Selection Bias-Destortion of results caused by inappropriate sampling technique
Information Bias – Distortion of results caused by systematically inaccurate measuring of exposure or outcome. Eg participants giving inaccurate information during interview
Confounding- Destortion of results when exposure on the outcome is mixed with the effects of other variables, eg testing for lung health in a smoker, lifestyle habits may influence outcome
