Wk 6 Part 1: Study Designs Flashcards
What is a variable?
Variable is a characteristic that we can record about the subjects or objects in a study
- Different types of variables require different types of analysis
- The way you ask questions will affect the type of answer you get
What are 2 types of variables?
- Numeric variables
- Categorical variables
What are 2 types of numeric variables? What are features?
What are 2 types of categorical variables? What are features?
What is the population in sampling?
Population is the larger group to which research results are generalised
- Usually the population is too large for the researcher to attempt to include all of its members in a study.
What is the sampling?
is selecting a small subgroup of the population (sample) that serves as a representative group for drawing conclusions about the population.
What are 2 features of sampling?
- Generalising results from the sample to the target population is done on the assumption that the sample reflects the characte ristics, heterogeneity and variations of the target population.
- Must specify the inclusion & exclusion criteria so that it is clear who would/not be classified as a member of the target population.
What are 3 advantages of sampling?
- Economical
- Time efficient
- Can be more accurate because there is greater control over the measurements and procedures
What is sampling bias?
Sampling bias occurs when the sample over/under-represent attributes of the target population.
- Can be conscious or unconscious
What are the 3 factors that sample size depends on?
- Size of the effect you are trying to detect
- Small effect needs big sample. Vice versa.
- Amount of residual variability
- Small variability needs small sample. Vice versa.
- Research design, including the sampling methods.
What is power analysis?
Power analysis is used to quantify relationships between these factors and the necessary sample size.
What are 2 types of sampling methods?
Sampling methods are as either probability or non-probability.
- In probability samples, the probability of each person being selected is equal.
- Non-probability samples does not use randomisation in sampling, but they can still use randomisation in allocation to treatments.
What are 5 types of probability sampling?
- Random sampling
- Systematic sampling
- Stratified random sampling
- Disproportionate sampling
- Cluster sampling
What are 6 features of random sampling as probability sampling?
- Purest form of probability sampling - each person have an equal chance of selection.
- Subjects are randomly selected from a population
- Reduces risk of systematic bias
- Computer generated lists of random numbers for
- Subject assignment to groups
- Not always practical to implement - need a master list of participants
What is a feature of systematic sampling as probability sampling?
Every nth subject of a population list is selected.
What are 3 features of stratified random sampling as probability sampling?
- The population is split into groups of similar individuals (strata) from which a simple random sample is drawn.
- Strata are chosen to correspond to important subgroups needing proper representation of the population.
- Includes under-represented subgroups
- Involves identifying relevant characteristics, and partitioning members of a population into homogeneous, non-overlapping subsets based on these characteristics.
What are 2 features of disproportionate sampling as probability sampling?
- If the strata in a population are of substantially unequal size, then stratified random sampling may give inadequate sample sizes for comparison.
- The probabilities of a subject being selected are unequal to ensure an equal representation of unequally sized strata.
What are 2 features of cluster sampling as probability sampling?
What are 4 types of non-probability sampling?
- Convenience sampling
- Quota sampling
- Purposive sampling
- Snowball sampling
What are 2 features of convenience sampling as non-probability sampling?
- Subjects are chosen based on availability.
- e.g. Friends, family etc.
- Huge potential for bias, may not represent the population
What is a feature of quota sampling as non-probability sampling?
The researcher guides the sampling process until the participant quota is met.
What are 2 features of purposive sampling as non-probability sampling?
- The participants are “handpicked” based on certain criteria.
- Useful for qualitative research.
What are 2 features of snowball sampling as non-probability sampling?
- Original participants identifies or refers other people with similar characteristics.
- Used when the desired characteristics of the required sample are rare.
What are 2 features of experimental designs?
Researcher has some control over the independent variable, experimental conditions and the comparison of groups of subjects.
- Involves some sort of treatment or intervention.
- Aim to provide evidence that the independent variable is the cause of changes in the dependent variable by controlling other possible influences.
What are 5 features of the Randomised Controlled Trial (RCT) for experimental designs?
- ≥2 groups are allocated using randomisation.
- 1 group receives the experimental variable (treatment)
- 1 group is the control group.
- Measurements are taken before & after treatment.
- Changes from pre- to post-test that appear in the experimental group but not the control group can be attributed to the treatment.
What is the double blinding for study blinding for experimental designs?
Double blinding (gold standard): Neither patient nor assessor knows which treatment the patient is receiving.
- Sometimes need a placebo
- Not always possible to achieve double blinding
What are 3 factorial deigns for experimental designs?
What are 2 outcome variables for experimental designs?
-
Primary outcomes must be specified and chosen carefully.
- If the primary outcome is statistically significant, then we may conclude that the treatment “worked” .
- Secondary outcomes (if relevant) should also be specified.
What are 5 features of Quasi-Experimental Designs?
- One group posttest-only design
- One group pretest-posttest design
- Non-equivalent control group design
- Non-equivalent control group pretest-posttest design
- Single subject designs
What are 6 features of single subject design of Quasi-Experimental Designs?
- Within-subject, single system, or “n of 1” study
- Studies a single patient or institution over time
- Differs from simple case report study as it is a structured experiment
- Outcomes are measured before & after the intervention
- Difficult to conclude that the treatment alone resulted in any of the differences as other factors may change over time
- External validity and ability to generalise is weak
- Most useful for making decisions about that particular patient
- Less useful for making broader inferences about the effects of an intervention
What are 4 features of observational studies?
Subjects are observed in their natural state.
- Group memberships are self-selected
- Subjects may be observed and measured but there is no treatment
- The most basic approach to trying to understanding variability in a population - similar groups with only 1 key different characteristic
- Have to be very careful about interpreting results as we cannot control outside factors.
What are 2 typesof observational studies?
- Surveys and polls ask questions to obtain data from a subset of the population to characterise attributes
- Other types of studies attempt to make inference about cause-and-effect relationships through observation
What is a retrospective observational study?
Retrospective studies look at past data, such as medical records, for potentially influential factors and outcomes.
What is a rprospective observational study?
Prospective studies follow participants into the future, observing exposure and then tracking outcomes.
What are Longitudinal Cohort Studies?
A population of subjects is identified by a common link.
What are 3 ways of thinking about the information from Longitudinal Cohort Studies?
- The researcher can follow them over time to see what happens to them. This is useful for establishing the natural history of a condition.
- The cohort can be divided into subgroups whose experience is to be compared. They are followed over time and the incidence of outcomes of interest is compared between groups. This is helpful for considering possible causative factors or for establishing predictive factors.
- The cohort may be followed over a period of time or until the event of interest occurs. The characteristic of those who have the event of interest with those who do not are then compared. This helps identify those most likely to develop the outcome
What is attrition?
Attrition is the loss of subjects to a study and is likely in studies with multiple follow-up or lengthy time periods between follow-ups.
What are 5 attritions of Longitudinal Studies?
- Death
- Patients moving away from follow-ups occur
- Patients transferring to other sources of health care
- Patients being institutionalized
- Patients withdrawing from the study
High attrition rates can badly deplete _____ and _____.
sample size; power
What are 3 features of high attrition rates?
- Attrition needs to be taken into consideration when first calculating the sample size.
- Bias is introduced where attrition is related to the factor/s of interest in the study
- Attrition rates/reasons can also be a valuable source of information. Attrition rates can inform health care policy makers, f or example, about natural “drop-out” rates.
What are 3 features of Case-Control Studies?
- Retrospective
- Similar individuals (one with a condition, one without) are matched (e.g. by age/gender), then the presence or absence of an event of interest (possible causal agent) in the past is examined.
- Typically examine the effects related to cause of a condition rather than its treatment
What are 6 features of Cross-Sectional Studies?
- A sample of subjects is studied at a single point in time
- Often used to determine the prevalence of a disease
- Provide a “snapshot” of certain issue at a point in time
- It tells you little about the development of the disease or condition or the order of events
- Ensure you obtain a sample that is representative of the population
- Ensure the results are not biased by a poor response rate
What are case reports?
Case reports describe the clinical history of a single patient in the form of a story
What are 2 features of case reports?
- Presents clinically important information and can be the basis for future studies.
- Not a strong design for determining cause-effect
What are 6 features that survey research describes?
- Attitudes
- Behaviours
- Opinions
- Beliefs
- Preferences
- Attributes
What are questionnaires?
Structured surveys, typically self-administered.
What are 3 advantages of questionnaires?
- Efficiency and convenience
- Can reduce bias
- Respondents may be more candid in responses
What is a disadvantages of questionnaires?
Potential for misunderstanding questions
Where are 4 locations where questions come from in the design of surveys?
- Literature
- Published studies
- Clinical/research experience
- Discussions with others
What is pilot testing & debriefing?
Pilot testing & debriefing with respondents are important to assess bias, validity, whether questions are understood, answered and worded appropriately - to reduce misinterpretations & bias.
What are 3 features of open questions?
- Allows people to answer in their own words
- Coding answers can be difficult
- More appropriate for interviews
What are 4 features of closed questions?
- Respondents select an answer from several options
- Easier coding and better uniformity of responses
- Does not allow respondents to express own opinion, leading to risk of bias
- Questions should be exhaustive (all possible options should be covered) and mutually exclusive (only 1 answer can be selected at a time)
What are 3 types of Response Formats?
- Dichotomous response: Yes/No. Agree/Disagree.
-
Likert scales: 3-6 categories of response
- Decision required on whether to include a neutral response (e.g. I don’t know)
- Visual analogue scale: Measure responses along a continuum between two extreme opposites
What is a dichotomous response as a response format?
Yes/No. Agree/Disagree.
What is a Likert scales as a response format?
3-6 categories of response
- Decision required on whether to include a neutral response (e.g. I don’t know)
What is a visual analogue scal as a response format?
Measure responses along a continuum between two extreme opposites
What are 3 features of question wording?
- Question wording can have a large impact on results
- Some surveys are deliberately biased in wording to pressure respondents to answer one way or another.
- Leading questions should be avoided
What are 6 features of unintentional bias?
- Avoid long words, ambiguity, jargon and lengthy sentences
- Avoid double barreled question (e.g. “How satisfied are you with your pay and job conditions?” )
- Avoid biasing or leading questions
- People lie to other and to themselves, particularly with sensitive questions.
- Ensuring confidentiality can help, but very sensitive topics are still problematic.
- Pilot studies is a good idea before large surveys!
