Quantitive Methods Flashcards
Randomised controlled trials
The aim: test a hypothesis regarding a treatment or intervention. Half the participant will receive the conventional treatment (experimental group) and some will receive the placebo (the control group).
Randomised controlled trials
The outcomes for the two groups are compared and the findings will determine whether the hypothesis about the new treatment has been supported. Three essential characteristics: o Randomisation o Manipulation o Control
Cohort studies
An idea about a possible cause and effect relationship.
- The pp that are recruited are exposed to the presumed cause and followed up over the years to see if they experience the presumed effect.
- Also follows other people who are in similar cases but not exposed to the problem
Case-control studies
- The differences in the group give the researchers an indication of the likely causes of the condition or problem.
- Randomisation and control are missing
- Can reveal correlations between variables
- Compare cases, whereby the researcher compares people with a condition or problem.
- Done retrospectively
- The researcher recruits participants with the problem and then recruits other people to be in the control group- these should be similar to the case as possible.
Surveys
starts with a research question.
- Allows researchers to gather information about attitudes, beliefs and behaviours and prevalence
- Key advantage- anonymity
- Look back to book about the low response rate
Experimental Methods
- Underlying theory of positivism
- Active manipulation of variables under the control of the researcher and attempt to study how participants will react to the manipulated conditions through monitoring one or more outcome measures.
- Attempts to establish laws and principles by developing hypothesis that can be accepted or rejected through structured tests under controlled conditions.
- Data collected are counted and analysed by using statistics.
Randomised controlled trials:
Hypothesis:
attempt to confirm that the information from an experiment is the best explanation we have at present.
- Making a statement that predicts a direct relationship between a cause and its effect.
researcher’s job to rule out possible causes leaving only the most plausible explanation remaining- trying to isolate the cause and ruling out any other. - Independent variable- the cause and the dependent variable- the effect.
Hypothesis:
- It is usual to propose a further hypothesis at the same time, one which stated the opposite position to the original.
- First hypothesis is labelled H1 and the null hypothesis is labelled H0.
Population and sampling
- Finding pp who are eligible for the experiment
- The participant need to exhibit the same characteristics
- Called the study population
- Specifying eligibility criteria for inclusion in the study
- Randomly sampling from the study population
Eligibility criteria
- Potential inclusion of all those people who might have the characteristics needed for the study.
- Often easier to state who is excluded first.
- Might want to limit the age range
- Should have a clear explanation of why the particular criteria has been selected.
- informed consent
Random sampling
- Method of picking a sample from the study population where everyone has an equal chance of being chosen
Sample size:
Random assignment:
- Sample size calculator
- Assignment of the participants to receive either the new treatment or the placebo.
- The purpose: create two groups that are characteristically equal at the baseline.
Concealment :
- Blind studies decrease biasness
- Controversy about placebo effect- the idea that people getting the placebo demonstrate relief of symptoms.
- RCT need to report the process of concealment
Intention to treat analysis (ITT)
- All data are analysed according to the trial protocol
- Avoids over-optimistic estimates
Selection bias-
Allocation bias-
systematic difference between those selected into the sample and those not selected. Therefore, not a representative.
a systematic distortion of the data.
Detection bias-
Observational bias-
differences in the assessment of the outcome between the experimental and control group.
contaminated observations by the observer’s belief, prejudice or background assumptions
Quantative methods of data collection:
Questionnaires:
improves ability to make better decisions
- A logical argument as to why you are using questionnaires as opposed to another form of data collection, if pros outweigh the cons.
Research question and design must be clear. Not be used if you want to find out how often people do things or self-reports people under/overestimate
Using existing questionnaires
- Validity in a QA involves collecting the data that usually answers the question
- Reliability in a QA provides an indication that the qa will give the same results when filled out by like-minded people in similar circumstances. Expect consistent answers
Questionnaire design
Closed questions:
short as possible
short or negative answer. Quick to answer and easier to code for analysis. Often contain researcher bias, the questions are researcher defined so can be misled
Open questions
Ask the pp to write longer answers, allow greater freedom to respond, depends on ability and time. More time consuming to code can easily misinterpret responses but more info.
Quantative data analysis: measures of clinical effectiveness
one establishes statistical significance and the other measures clinical effectiveness
Statistics:
Statistics is about dealing with uncertainty and variation, and the aim of statistical analysis is to establish that an intervention or medication will work on most people- numerical value
Relative risk
Relative risk is the ratio of the risk of the event in one regimen compared with the risk of an event in another regimen
Median
To find the median, order the numbers and see which one is in the middle of the list.
Mode
To find the mode, order the numbers lowest to highest and see which number appears the most often.
Mean
To find the mean, add all the numbers together then divide by the number of numbers.
Range
To find the range, subtract the lowest number from the biggest number.