Physics - Working scientifically

Accuraccy

Closeness of an observed value to its true value

Absolute error

| True value - Observed value |

Relative error

| true value - observed value| / True value

Percentage error

| (True value - Observed value) / True value | x 100

Systematic error

When repeated, observed values are displaced in the same direction from the true values. They may read consistently higher or lower than the true values

Causes of systematic error

improperly calibrated measuring instruments (e.g. forgetting to zero a scale”

Random error

When repeated, observed values are scattered randomly above and below the true value

Cause of random error

caused by random fluctuations in the ambient conditions or uncontrolled variables

How to reduce absolute error

Use measuring instruments with appropriate resolution, read analogue instruments directly front on to reduce parallax

How to reduce relative error

Design experiments to increase the size of measurements being made

How to reduce systematic error

ensure that measuring instruments are properly calibrated and zeroed

How to reduce random error

Hold control variables constant so that they dont add fluctuations to the measurements. Conduct repeated trials of measurement and calculate an average

Uncertainty

The possible margin of error within an experiment

when does a measurement agree with an accepted value

If the accepted value falls within the measurements uncertainty range

Precision

When the uncertainty of an experiment is small

Absolute uncertainty

normal uncertainty e.g. 5 +- 0.2

percentage uncertainty

(Absolute uncertainty / measured value) x 100

How do you calculate uncertainty in a measuring device

Halve the smallest readable graduation on the scale of a device

Uncertainty in LOBF

Find the uncertainty of each measurement included within the line of best fit, then draw two lines of worst fit with one line starting at the lowest uncertainty and finishing at the highest uncertainty and vice versa.

Once you have done that, the uncertainty is calculated using:

+- (max LOWF gradient - min LOWF gradient)/2

How can uncertainty be reduced

same ways as reducing errors

combining uncertainty

+-x/ the approximate lengths then find the minimum and maximum lengths by +-x/ the minimum and maximum values

Reliability

The consistency in results

Internal reliuability

When repeated trials within an experiment are consistent (precision)

External reliability

When the results from one experiment are consistent with those from other experiments conducted in the same way

How can reliability be asessed

Quantifying spreads of trials around an average. large spreads are unreliable and vice versa. OR if the relative uncertainty of the trials is less than an arbitrary value e.g. 5%

How can you improve reliability

reducing the effect of random errors

Validity

An experiment which examines what is intended aka the relationship between the independent and depended variable

How can we assess validy

An experiment is valid only if the independent variable can be varied

when is an experiment valid

1. reliable results

2. accurate results