Fichas
topic 7
Fichas del conjunto (20)
- 1Observational research designs- can be defined as non-interventional and non experimental
- variables are not manipulated by researchers
- observing natural relationships between factors and outcomes
- types include cross sectional studies, longitudinal studies, case-control studies, cohort studies, and survey studies - 2cross sectional studies- assess a population at a single point in time
- reflect the situation of a disease or outcome at a particular moment, in a particular population - 3longitudinal studiescontinuous or repeated measures to follow particular individuals over prolonged periods of time
- 4case control studies- identify participants (diseased or not diseased)
- quantification of the number of individuals among the cases and the controls who are exposed allow for statistical association between exposure and outcomes to be established - 5cohort studiesidentify study participants based on exposure status and follow them through time to identify which participants develop the outcomes of interest
- 6survey studies- common descriptive research method in health sciences
- respondents are asked a series of questions in a standard manner so they can be easily quantified and analysed statistically
- enables researcher to describe characteristics of the sample and make generalisations to the larger population of interest
- particularly useful for collecting info about a research phenomena that is not directly observable or measureable, and collecting data from a wide geographical range
- 3 main types of surveys are epidemiological, surveys on attitudes to health service or intervention, and questionnaires assessing knowledge on a particular issue or topic - 7interventional research designs- also called experimental, researcher intervenes or manipulates a variable/s
- types include RCTs, pre-post study design, and non-randomised control trials - 8randomised control trials- subjects are randomly assigned to either experimental, or comparison group
- experimental group receives intervention being tested
- comparison or control group receives an alternate conventional treatment
- ideally participants and those collecting the data won't know which is which, to eliminate confounding variables, and stops bias being introduced - 96 features of an RCT1. the sample will be appropriate to the hypothesis so results are generalisable
2. effective randomisation of subjects
3. both groups treated identically in all respects apart from intervention, and patients and investigators will be blinded
4. investigator assessing outcome will be blinded
5. patients are analysed within the group they were assigned
6. analysis focuses on testing the research question that initially led to the trial - 10pre-post studies- measure the occurance of an outcome before and after a particular intervention
- single arm, one group measure, or multiple arms with comparison between groups
- the no-intervention arm acts as the control group
- strength of temporality to be able to suggest that the outcome is impacted by the intervention
- however do not have control over other elements that are also changing
- suffers many threats to internal validity, though can provide preliminary evidence for intervention effectiveness - 11non-randomised control trials- compare a group where an intervention was performed with a group that had no intervention
- convenient
- can suggest possible relationships between the intervention and the outcome
- good choice where RCT is ideal, but practical considerations make it infeasible
- subject to many types of bias and error, so not considered strong - 12variability in a single quantitative variable- the variability pattern is the distribution, pattern typically involves a central tendency where observations gather around a central value
- interested in describing 3 main aspects of the distribution:
1. the location of the centre of variability
2. the spread of the variability - how far values extend from the centre
3. the shape of the variability in whether or not values are spread symetrically on either side
- once described, look for outliers - 13Histograms- useful for plotting the distribution of a continuous variable when you have a lot of observations - good picture of location, spread, and shape but are somewhat subjective, and it is difficult to compare more than 2 groups
- split the range into intervals called bins, and count the number of observations falling in each bin
- the issue with histograms is that changing the no. bins gives a different picture, less bins gives a better picture of distribution
- if the tail is to the right, then it is positively skewed. to the left = negatively skewed - 14density plots- show a continuous estimate of density
- useful for comparing distributions between groups as it is east to overlay them
- downside is you need software to create them, where you can do histograms by hand - 15percentiles and quartilesthe median is the middle in value, in an order from smallest to larges. As there are 50% above and below, this is called the 50th percentile, or 0.5 quantile
- the 25th percentile is the 0.25 quantile,or first quartile
- the 75th percentile is the 0.75 quantile, and 3rd quartile
- the 0th percentile is the 0.00 quantile, or minimum
- the 100th percentile is the 1.00 quantile, or maximum - 165 number summarythe 1st and 3rd quartiles, plus the median, minimum, and maximum give a 5 number summary
- 17interquartile rangeis the distance between the 1st and 3rd quartiles
- 18box plots- uses quantiles to give useful visualisation of the distribution of a continuous variable, showing location, spread, and shape, as well as flagging unusual observations
- the 'box' covers the middle 50% of data (1st quartile, median, and 3rd quartile)
- if the whiskers are roughly balanced, the distribution is symmetrical
- outliers are values in the data that don't match the overall distribution observed. They can be genuine, but can also arise from errors. Outliers are 1.5 times the length of the box (called a yard stick) if a value exceeds this then it is an outlier
- compare groups by making side by side box plots
- when the whisker above the median line is the largest, it is skewed right/+ve, when below, it is skewed to the left/-ve - 19bar charts- categorical variables (nominal or ordinal) are most often graphed with a bar chart
- shows frequencies in each category or percentage of participants in each category by using different bar heights
- similar to histogram, but has separate bars compared to continuous
- useful for comparing frequencies across categories - 20pie chartsare less reliable and should be avoided
Información
- Descripción
- HRSS3101 Fichas sobre topic 7, creado por Soph B el 23/06/2020.
- Creado por
- Soph B
- Visitas
- 1
- Valoración