What is accuracy? systematic errors
Improved by
Def: How close a measurement is to the true value.
To analyse:
- suggest systematic errors (consistently false results in one direction that decrease accuracy)
- suggest changes to method
Improved by: Correcting for systematic errors or using calibrated equipment.
Precision (def, how to analyse)
Def: How close repeated measurements of the same experimental group are to each other
(experimental group= not control group)
- analyse data from within experimental groups, not between
- look for random errors that decrease precision
- suggest changes in method that could improve precision (e.g. use equipment with finer scales- more significant figures)
- to determine if precision of data is being achieved (repeatability)- repeat trials and if consistent- precise
better: effects of random errors reduced by multiple trials and average
validity
how well an experiment truly measures what it intends to measure.
Improved by: Designing the experiment well, controlling variables, control group, and using appropriate methods, large sample size
quantitative vs qualitative data:
quantitative: numerical
tells you how much, how many, how often
allows for more statistical analysis and less subjectivity
qualitative: descriptive, tells you what something is like, can observe or describe
repeatability vs reproducibility
repeatability
the precision obtained when the same or similar results are produced by the same student, using same method and equipment, under the same conditions in a short time frame
reproducibility
degree of agreement between results of experiment produced by different students, working underdifferent conditions, with different equipment at different times but SAME method
(usually under same conditions but in a different lab)
reproducible if yields same or similar results in equivalent conditions
Define IV, DV, controlled variable
IV: variable that is changed by the experimenter
DV: variable that is being measured / changes as a result of changing the IV
controlled variable: variables kept constant (To ensure the change in the dependent variable is caused only by the independent variable- improves validity)
What is an outlier?
Suggest some causes of outliers
How can they be managed in data analysis?
- result that is a long way from other results and seen as unusual
causes:
- personal errors (e.g. contamination of beaker with another substance)
- random errors
- data should still be accounted for and analysed
- don’t include in average if the cause is identified
- include in average if cause is not identified to avoid experimenter bias
what causes random errors and reduces precision vs systematic errors and reduces accuracy?
How to reduce impact of random and systematic errors?
random errors: unpredictable variations in measurements that result in a spread of readings
causes:
measurment instruments
reduce the effect of RE:
- mutiple trials and average results
systematic errors: cause readings to differ from the true value by a consistent amount each time, shifted in one direction from the true value, methododology
Causes:
uncalibrated equipment or used incorrectly
flawed method
- cannot be improved by repeating experiment
Justify the value of a control group in scientific investigations.
- serves as a baseline to compare whether the IV is the only factor causing changes to the DV
- allows researches to check the validity of their experiment
types of methodologies: case study
- real world or hypothetical situation
- reflects what is expected to happen in the real world
- allows for analysis of causes and consequences of outbreaks which can lead to plausible recommendations for future outbreaks
types of methodologies: field work
- observations external to classroom environment
- sampling, surveys, observation
- determine correlational rather than causal relationships
(requires contact with environment / pathogen etc)
types of methodologies: controlled experiment
- determines relationship between IV and DV
- ALL other variables are controlled
types of methodologies: correlational study
- to determine the associations between variables
- determine which factors are more important
- to make predictions
- do not control or change variables
types of methodologies: modelling
- creating a physical model (e.g. representation of an object), conceptual model, mathematical model
- predict, simulate, explain behaviour
types of methodologies: product, process or system development
- designing a scientific instrument, method or system to meet human need
- solve practical problems
- may have technological applications
Developing a mobile health monitoring system that tracks patients’ vital signs in real time.
types of methodologies: simulation
- using a model to study the behaviour of a real or theoretical system
- useful when direct experiment is expensive, difficult or unethical
types of methodologies: literature review
- systematic analysis of previously published scientific studies
- determine gaps in research
- learn what is already known about a topic
