Fichas
Statistics overview
Fichas del conjunto (14)
- 1Descriptive statistics: Inferential statistics:Descriptive statistics: summarise information about a population.
Inferential statistics: to make inferences derived from the research data or other analyses. - 2Reasons for consulting a statistician:Consulting a statistician:
Consult as early as possible, especially when unsure about data.
Helpful for calculating the sample size required to detect whether an estimated difference reaches the predetermined clinical significance threshold.
Helpful for ways to control confounding.
Helpful to consult about right and wrong ways to approach data. - 3Standard deviation:Standard deviation: This measures the variation in the data and does not depend on the sample size.
- 4Standard error of the mean:Standard error of the mean: This measures the uncertainty of the sample mean and does depend on the sample size (SE=SD/√n).
- 5Confidence interval:Confidence interval:
The range within which you can be confident, to a specified level, that the true value of the population being approximated lies. For example, if this test were repeated 100 times, 95 times out of 100 the true population mean would lie within this interval. Similarly, we can be 95% confident that the true population mean lies within this interval.
The width of the confidence interval is smaller with increased sample size.
95% CI= mean ± 1.96 x SEM) - 6Null hypothesis:Null hypothesis:
The hypothesis which predicts that there is no difference in outcome between the intervention and control group.
Usually opposite to the researcher’s hypothesis. - 7P value:P value:
Probability value used to assess the probability that the result has occurred due to chance.
The lower the P value the less likely the outcome occurred due to chance. - 8Statistical significance:Statistical significance:
If the P value is less than 0.05 then the observed outcome is unlikely to be due chance and therefore is statistically significant. we can reject the null hypothesis.
If the confidence interval does not contain the null value (0 for AR, 1 for RR or OR), then the result is statistically significant at the 0.05 level and we can reject the null hypothesis. - 9Sources of error:Sources of error:
Selection bias.
Measurement or information error.
Confounding. - 10Selection bias: Sources:Selection bias: Bias introduced when selecting participants for a study.
Sources:
Selecting participants.
Volunteers (self-selection)
Healthy worker effect
Loss to follow up
Detection/ascertainment bias (over or under diagnosis)
Low response rates. - 11Measurement error: Measurement of information error: Random error: Systematic error: Sources of measurement error:Measurement error:
Measurement of information error:
Misclassification of case and control.
Random error:
Measure of precision-variation in measurements
Systematic error:
Measuring instrument not calibrated properly leading to a systematic error
Sources of measurement error:
Recall bias
Interviewer or observer bias
Inaccurate/uncalibrated measuring instruments - 12Confounder requirements:Confounding:
Must be a risk factor for the outcome.
Associated with the the exposure of interest.
Not lie on the causal pathway - 13Controlling for confounding:Controlling for confounding:
Randomization: Only possible in RCT; even distribution of confounders in control and intervention groups.
Restriction: Restrict participants to those only with or without confounders.
Matching: Cases and controls are matched on confounders. Must account for this during statistical analysis; stratification or multivariable modelling. - 14Study validity:Study validity:
Internal validity: The extent to which the study results reflect the true situation within the study sample in the absence of any alternative explanations; chance, bias and confounding.
External validity: The extent to which the results of the study are generalizable.
Información
- Descripción
- Public health Fichas sobre Statistics overview, creado por Ashutosh Kumar el 17/03/2017.
- Creado por
- Ashutosh Kumar
- Visitas
- 5
- Etiquetas
- Valoración