How Science works

An accurate measurement is one which is close to the true value.

Anomalous data are those measurements that fall outside the normal, or 

expected, range of measured values. Variation is a characteristic of all living

organisms, and it is often difficult in biological investigations to distinguish

between data that reflect this variation and those that are genuinely

anomalous. A large number of readings allows anomalous data to be

identified with greater certainty.

When using a measuring instrument, calibration involves fixing known points

and constructing a scale between these fixed points.

A change in one variable that results from or is caused by a change in

another variable

Chance is essentially the same as luck. If a coin is tossed in the air, whether 

it comes down heads or tails is purely due to chance. The results of any

investigation could have a genuine scientific explanation but they could be

due to chance. Scientists carry out statistical tests to assess the

probability of the results of an investigation being due to chance.

A confounding variable is one that may, in addition to the independent

variable, affect the outcome of the investigation. Confounding variables must

be kept constant or the investigation will not be a fair test. In some

investigations, ecological investigations in particular, it is not always possible

to keep confounding variables constant. In such cases, these variables

should be monitored. In this way it may be possible to decide whether or not

the factor concerned affects the outcome of the experiment. Confounding

variables are sometimes referred to as control variables.

A control experiment is one that is set up to eliminate certain possibilities. In

a well designed investigation, the independent variable is changed and all

confounding variables are kept constant.   The possibility exists, however,

that something else other than the independent variable might have produced

the results that were obtained. A control experiment is one that is designed to

eliminate this possibility.

A control group is one that is treated in exactly the same way as the

experimental group except for the factor that is being investigated. This

allows scientists to make a comparison. It ensures that the data

that are

collected are valid because any differences between the results for the

experimental group and those for the control group will be due to a single

independent variable.

A confounding variable is one that may, in addition to the independent

variable, affect the outcome of the investigation. Confounding variables must

be kept constant or the investigation will not be a fair test. In some

investigations, ecological investigations in particular, it is not always possible

to keep confounding variables constant. In such cases, these variables

should be monitored. In this way it may be possible to decide whether or not

the factor concerned affects the outcome of the experiment. Confounding

variables are sometimes referred to as control variables.

A correlation shows that there is a relationship between two variables,

however, it might not be a causal one.

The dependent variable is the variable the value of which is measured forvariable         each change in the independent variable.

A trial, usually used in the context of medicine, when assessing the effects of a new drug or treatment on humans. Neither the patients nor the scientists concerned know which treatment a particular individual is receiving until after completion of the trial. This helps to avoid bias and increase the validity of the trial.

Errors cause readings to be different from the true value.

The data or observations that are used to support a given hypothesis or

belief.

A fair test is one in which only the independent variable has been allowed to affect the dependent variable. A fair test can usually be achieved by keeping all other variables constant or controlled.

Sometimes known as an experimental hypothesis, this is a possible explanation of a problem that can be tested experimentally.

The independent variable is the variable for which values are changed by the 

 investigator.

A statistical test requires a clear hypothesis to test. It is often difficult to predict what would happen as the result of an investigation. It is much easier to phrase a hypothesis in terms of there being no difference or no association. A hypothesis worded in this way is called a null hypothesis. As the result of carrying out a statistical test, a  decision can be made about whether to accept or reject this null hypothesis.

A placebo is a dummy pill or injection given to members of a control group

in medical trials. Where a placebo is in the form of a pill, it should be identical

to the pill used with the experimental group. The only difference should be

that that the placebo does not contain the drug being trialled. The use of

placebos helps to ensure that the data collected from a trial are valid.

Precision is related to the smallest scale division on the measuring instrument that is being used. A set of precise measurements will have very little spread about the mean value.

Precision is related to the smallest scale division on the measuring instrument that is being used. A set of precise measurements will have very little spread about the mean value.

Probability is the likelihood of an event occurring. It differs from chance in that it can be expressed mathematically. In statistical tests, probabilities are usually expressed as a decimal fraction of one. Thus a probability of 0.05 means that an event is likely to occur 5 times in every 100.

Once an experimental method has been shown to produce valid and reliable results, it becomes a protocol used by other scientists.

A random distribution is one that arises as a result of chance. When investigating, for example, variation in living organisms, the data collected will only be valid if they have been collected at random. This avoids observer bias and allows statistical tests to be used in an analysis of the results of the investigation.

Random errors occur in an unpredictable way. They may be caused by human error, faulty technique in taking measurements or by faulty equipment.

Raw data are instrument readings and other data collected at the time of the investigation. These data may subsequently be processed and used to calculate percentages and standard deviations.

The results of an investigation may be considered reliable if they can be repeated. If other scientists get the same results, then the results of the initial investigation are more likely to be reliable. The reliability of data within a single investigation can be improved by carrying out repeat measurements.

These errors cause readings to be spread about some value other than the true value. In other words, all the readings are shifted in one direction from the true value. Systematic errors may occur when using a wrongly calibrated instrument.

This is the accurate value which would be found if the quantity could be measured without any errors.

Data are only valid if the measurements that have been made are affected by a single independent variable only. They are not valid if the investigation is flawed and control variables have been allowed to change or there is observer bias. Conclusions are only valid if they are supported by valid and reliable data measured to an appropriate level of accuracy.

Zero errors are caused by instruments that have an incorrect zero. A zero error may occur when the needle on a colorimeter fails to return to zero or when a top-pan balance shows a reading when nothing is placed on the pan.