ICAS Science key terms

Information gathered directly through the senses or with instruments, such as noticing that a liquid turned cloudy or that a seedling is 12 cm tall. It records what actually happens rather than what it might mean.

A number paired with a unit that describes a quantity, such as 250 mL, 3.4 seconds or 18 degrees Celsius. It requires choosing a suitable instrument and reading its scale carefully.

An agreed standard amount used to express a measurement, such as metre, gram, second or litre. Units allow different people to compare and communicate results meaningfully.

The marked divisions on a measuring instrument and the value that each division represents. Reading the scale correctly allows a measurement to be reported with a sensible number of digits.

A device used to extend the senses and obtain measurements, such as a ruler, thermometer, stopwatch or microscope. Choosing an instrument with a suitable range and scale improves the data collected.

How close a measured value is to the true value. Accurate results come from correctly calibrated instruments, careful technique and reading scales from the correct viewing position.

How close repeated measurements are to one another, or how finely an instrument's scale is divided. A set of readings can be precise yet still inaccurate if the instrument is wrongly calibrated.

An approximate value worked out without exact measurement, often by comparing with a known quantity. Estimates are useful for planning and for checking that a measured answer is reasonable.

Observations described in words rather than numbers, such as colour, smell, texture or whether a change occurred. These can be sorted into categories and compared across trials and conditions.

Observations recorded as numbers with units, such as mass in grams, time in seconds or temperature in degrees Celsius. Numeric data can be averaged, graphed and compared between trials.

An explanation drawn from observations together with prior knowledge, such as concluding that a cloudy liquid now contains a new solid. An inference is an interpretation and could be wrong.

The observations and measurements collected in an investigation, usually kept in tables or graphs so patterns can be examined. Data become evidence when they are used to support or challenge an explanation.

An organised grid of rows and columns used to record data, normally with a heading and unit for each column. A well set out table makes values easy to compare, check and graph.

A visual display of data, such as a bar, line or pie chart, in which position, height or area represents values. Graphs make trends and comparisons easier to see than a column of numbers.

One of the two numbered lines that frame a graph, usually the horizontal axis for the independent variable and the vertical axis for the dependent variable. Each axis needs a label and a unit.

The general direction or pattern shown by a set of data, such as steadily increasing, decreasing or staying level. Describing the overall trend is more useful than describing single points.

A data point that lies clearly apart from the pattern of the other readings. It should be checked for measurement or recording error before being excluded, and any exclusion explained.

A single value found by adding all the readings and dividing by how many there are. Averaging repeated measurements reduces the effect of random variation and gives a more dependable result.

A testable statement predicting a relationship between variables, such as 'plants given more light grow taller'. It must be possible to design an investigation that could show it to be wrong.

A statement about what is expected to happen in a particular situation, based on a pattern, a hypothesis or prior knowledge. A prediction can be checked against the results that are collected.

A short statement saying what the results show, whether the hypothesis was supported and which evidence backs this. It should stay within what the collected data can reasonably support.

The relevant data and observations used to support or challenge a claim. Strong evidence is sufficient in amount, directly related to the question and consistent across repeated trials.

Using a pattern found within the data to predict values outside the measured range, such as continuing a trend line beyond the last point. Predictions far beyond the data are uncertain.

Estimating a value that falls between two measured points by assuming the pattern continues smoothly from one to the other. It is usually more dependable than predicting beyond the measured range.

Any factor in an investigation that can change or take different values, such as temperature, time or amount of light. Identifying the variables is necessary to plan a fair test.

The variable the investigator deliberately changes or selects in order to observe its effect. Its values are usually placed along the horizontal axis when the results are graphed.

The variable that is measured in response to the changes made, such as growth in height or time taken to react. It is the outcome recorded as data during the investigation.

A factor kept the same across all trials so that it cannot influence the outcome, such as using the same container or the same volume of water. Keeping controls makes comparisons meaningful.

An investigation in which only the independent variable is changed while all other relevant factors are held constant. This allows any measured difference to be attributed to the change that was made.

A comparison group that receives no treatment, or a standard treatment, so that the effect of the tested factor can be judged. Differences between the groups show what the treatment did.

Carrying out the same measurement or experiment more than once, ideally under the same conditions. Repeats reveal random variation, help identify outliers and make results more trustworthy.

The extent to which repeating an investigation gives consistent results. Reliability is improved by repeating trials, using larger samples and applying careful, consistent technique each time.

The extent to which an investigation really tests what it claims to test, using suitable controls and a measurement that reflects the variable of interest. A reliable study can still be invalid.

The number of individuals or items included in an investigation. Larger samples usually give results that better represent the whole population and reduce the influence of chance variation.

A relationship in which two variables tend to change together, either in the same direction or in opposite directions. Correlation on its own does not prove that one variable causes the other.

A relationship in which a change in one factor directly produces a change in another. Establishing it requires a controlled investigation rather than merely noticing that two variables move together.

A simplified representation of a real object or process, such as a diagram, a physical replica or a computer simulation. Models highlight key features but always leave some detail out.

Something accepted as true without being tested in the current investigation, such as that a thermometer is working correctly. Unstated assumptions can weaken a conclusion if they turn out to be wrong.

Grouping objects or organisms according to shared features, often by working through a series of yes-or-no questions called a key. Classifying reveals similarities and differences and makes large sets of items manageable.