AQA A-L evel Biology How Science Works Glossary

Types of Variable

Types of Data

Dependent Variable
dependent variable

The variable you measure, to see

how it is affected by the
independent variable.
Independent
Variable

Data

Confounding Variables
Any variables that could also affect the
dependent variable. Confounding
variables should be
controlled in a fair test.

(measurements,
singular datum)

Quantitative or
Numeric Data

Qualitative or
Categoric Data

(numbers)

(words)

Control variables

Confounding variables that

The variable you choose
are kept constant
independent variable
to change, to see how it
(controlled) during the
affects the dependent variable.
experiment. If you can't control a variable
You may also measure it when
(such as weather in a field investigation),
you change it.
you should at least monitor it

Continuous Data

Discrete Data

Ordered Data

Nominal Data

can have any value

e.g. 7.34, -294.6, 2x105

only whole numbers

e.g. no. of atoms

can be ranked

can't be ranked
e.g. male, female

Errors
Bias

Random Errors

Experimental Design
Controlled Experiment (FairTest)
When all relevant variables are controlled, so that observed changes in the
dependent variable must be due to changes in the independent variable.

Control Experiment (Control)

An additional experiment designed to eliminate alternative explanations for the
main experiment and so show that observed changes in the dependent variable
must be due only to changes in the independent variable.

Inaccuracies due to mistakes, poor

technique, or random variation.
Random errors are very common, but
can be improved by taking many
replicates. Data with a small random
error is said to be precise.

Systematic Errors

A group or sample treated in the same way as the experimental group, except
for the factor being investigated e.g. a placebo group in a drugs trial. By
comparing the results for two groups it can be shown that observed changes in
the dependent variable must be due only to changes in the independent variable.

Inaccurate measurements in one

d i re c t i o n o n ly, d u e t o p o o r
calibration or poor technique.
Systematic errors can not be
improved by taking more replicates.
Data with a small systematic error is
said to be reliable.

Placebo

Zero Error

Control Group

A dummy pill, injection or treatment that has no physiological effect (e.g. a sugar
pill or saline injection). Used in a clinical trial to allow for the placebo effect the observation that symptoms can improve when patients believe they are
being helped.

RCT
The best experimental design for a drug trial. RCT stands for Randomised
Controlled Trial, or in more detail, a Randomised, Placebo-Controlled, Doubleblind Trial. This design ensures that the trial is valid free from bias.
! Randomised means the study and control groups are allocated randomly
! Placebo-controlled means the study group (taking the drug to be tested) is
compared to a placebo group (who are given a placebo).
! Double-blind means that neither the subjects nor the investigators know
who is in the study or placebo groups.This avoids bias.

Protocol
A method or technique that has been shown to produce valid and reliable
results.

Hypothesis
A suggested explanation of observations or results that can be tested. Also
known as a scientific hypothesis. A good hypothesis can be used to make
predictions.

Quality of Data

e.g. small, medium, large

When the observer chooses some

results and ignores others, to support a
particular view. Also called cherry
picking.

Anomaly or Outlier
A measurement that falls far outside
the expected range and is therefore
probably due to experimental error.
Anomalies should be rejected, since
they skew the mean, but it is very
difficult to distinguish between
anomalies and normal biological
variation.

Calibration
Ensuring that a measuring instrument

A particular kind of systematic error, gives correct readings by fixing known

where the instrument does not return to points then construction a scale
zero.
between them.

Simple Analysis
Replicates

Standard Deviation (SD)

A measure of the dispersal of the replicates

about the mean. In a normal distribution 68%
of the replicates will be within 1 standard
Raw Data
deviation of the mean, and 95% will be within 2
The original measurements or standard deviations of the mean.
re c o r d i n g s b e f o re a ny
manipulation or processing.
Standard Error of the mean (SEM)*
A measure of the uncertainty, or error, of a
calculated mean.The smaller the standard error,
Mean or Average
T h e m i d - p o i n t o f t h e the more reliable the mean.
replicates.
= sum of replicates / N
95% Confidence Interval (CI)*
Another measure of the error of the mean.We
Range
can be 95% confident that the true mean lies in
The highest and lowest the range (mean CI). The top and bottom of
replicates, or the interval this range are called the confidence
limits.
between them.
Repeats of a measurement.

Accurate Data
TrueValue

Measurement that are close to the

The real value of a measurement, if it true value.
could be measured with no errors at
Valid Data
all.
The best quality data, i.e. data that is
precise and reliable and obtained
Precise Data
1. Measurements that give similar from an unbiased, controlled
values when repeated. The experiment that addresses the stated
replicates therefore have a small aim. Valid data is assumed to be
range.
accurate.
2. Data measured on sensitive
equipment with a suitably fine scale, Evidence
e.g. 20 mm is more precise than 2 Any data or observations that are used
cm.
to support a particular hypothesis.

Reliable Data

Anecdote

Findings that can be repeated. This

includes by the original investigator; by
other scientists; by other techniques;
or those that agree with secondary
sources.

An observation or story from real life.

Anecdotes are not evidence and
cannot be used to support a
hypothesis, but they can be useful to
suggest a new testable hypothesis.

HGS Biology

Statistical Analysis
Correlation (or Association)

Null Hypothesis*

When one variable changes with

another variable, so there is a relation
between them. However it may not be
a causal relation. The strength of a
correlation can be measured using the
Pearson or Spearman test.

The statement that is tested by a

statistical test. The null hypothesis is
fixed for each test, but always says that
there is no difference or no
association. The null hypothesis has
nothing to do with a scientific
hypothesis.

Causal Relation

When changes in one variable cause

the changes in another variable. Can P-value*
only be shown by a controlled The result of a stats test, expressed as a
experiment.
probability. It is effectively the
probability that the null hypothesis is
StatisticalTest*
true. If P<0.05 then we reject the null
Something that tests whether
hypothesis, otherwise we accept it.
observed differences or associations
are significant, or just due to chance.
* = A2 only

NCM 01/09

AQA A-L evel Biology How Science Works Glossary

Types of Data
Data

dependent variable

Types of Variable
Dependent Variable

(measurements,
singular datum)

The variable you measure, to see how it is

affected by the independent variable.

Quantitative or
Numeric Data

Qualitative or
Categoric Data

(numbers)

(words)

Continuous Data

Discrete Data

Ordered Data

Nominal Data

can have any value

e.g. 7.34, -294.6, 2x105

only whole numbers

e.g. no. of atoms

can be ranked

can't be ranked
e.g. male, female

e.g. small, medium, large

Confounding Variables
The variables that could also affect the
dependent variable. Confounding variables
should be controlled in a fair test. If you
can't control some variables (such as
weather in a field investigation), you should
at least monitor them.

independent variable

Independent Variable

The variable you choose to change, to

see how it affects the dependent
variable.You may also measure it
when you change it.

Control variables

Experimental Design

Confounding variables that are kept constant (controlled) during the experiment.

Controlled Experiment (Fair Test)

When all relevant variables are controlled, so that observed changes in the
dependent variable must be due to changes in the independent variable.

Control Experiment (Control)

An additional experiment designed to eliminate alternative explanations for the
main experiment and so show that observed changes in the dependent variable
must be due only to changes in the independent variable.

Control Group
A group or sample treated in the same way as the experimental group, except for
the factor being investigated e.g. a placebo group in a drugs trial. By comparing
the results for two groups it can be shown that observed changes in the
dependent variable must be due only to changes in the independent variable.

Placebo
A dummy pill, injection or treatment that has no physiological effect (e.g. a sugar
pill or saline injection). Used in a clinical trial to allow for the placebo effect the observation that symptoms can improve when patients believe they are being
helped.

RCT

Errors
Random Errors

Bias

Inaccuracies due to mistakes, poor

technique, or random variation. Random
errors are very common, but can be
improved by taking many replicates.
Data with a small random error is said
to be precise.

When the observer chooses some

results and ignores others, to support a
particular view. Also called cherry
picking.

Anomaly or Outlier

A measurement that falls far outside the

expected range and are therefore
Inaccurate measurements in one
probably due to experimental error and
direction only, due to poor calibration
should be ignored. It is very difficult to
or poor technique. Systematic errors
decide if measurements are anomalous
can not be improved by taking more
or part of normal biological variation.
replicates. Data with a small systematic
error is said to be reliable.
Calibration
Ensuring that a measuring instrument
Zero Error
gives correct readings by fixing known
A particular kind of systematic error,
where the instrument does not return points then construction a scale
between them.
to zero.

Systematic Errors

The best experimental design for a drug trial. RCT stands for Randomised
Controlled Trial, or in more detail, a Randomised, Placebo-Controlled, Doubleblind Trial. This design ensures that the trial is valid free from bias.
! Randomised means the study and control groups are allocated randomly
! Placebo-controlled means the study group (taking the drug to be tested) is
compared to a placebo group (who are given a placebo).
! Double-blind means that neither the subjects nor the investigators know who
is in the study or placebo groups. This avoids bias.

True Value

Valid Data

The real value of a measurement, if it

could be measured with no errors at
all.

Protocol

Accurate Data
A measurement that is close to the
true value.

The best quality data, i.e. data that is

precise and reliable and obtained
from an unbiased, controlled
experiment that addresses the stated
aim.Valid data is assumed to be
accurate.

Precise Data

Evidence

1. Measurements that give similar

values when repeated. The replicates
therefore have a small range.
2. Data measured on sensitive
equipment with a suitably fine scale,
e.g. 20 mm is more precise than 2
cm.

Any data or observations that are used

to support a particular hypothesis.

A method or technique that has been shown to produce valid and reliable results.

Hypothesis
A suggested explanation of observations or results that can be tested. Also known
as a scientific hypothesis. A good hypothesis can be used to make predictions.

Quality of Data

Reliable Data
Findings that can be repeated. This
includes by the original investigator; by
other scientists; by other techniques; or
those that agree with secondary
sources.

Simple Analysis
Replicates
Repeats of a measurement.

Raw Data
The original measurements or
recordings before any
manipulation or processing.

Mean or Average

Standard Deviation (SD)

A measure of the dispersal of the replicates
about the mean. In a normal distribution 68% of
the replicates will be within 1 standard deviation
of the mean, and 95% will be within 2 standard
deviations of the mean.

The highest and lowest

replicates, or the interval
between them.
HGS Biology

Statistical Test

Another measure of the error of the mean. We

can be 95% confident that the true mean lies in
the range (mean CI). The top and bottom of
this range are called the confidence limits.

P-value

The result of a stats test. It is effectively

Something that tests whether observed the probability that the null hypothesis
differences or associations are
is true. If P<0.05 then we reject the null
significant, or just due to chance.
hypothesis, otherwise we accept it.

Correlation (or Association)

A measure of the uncertainty, or error, of a

95% Confidence Interval (CI)

Range

An observation or story from real life.

Anecdotes are not evidence and cannot
be used to support a hypothesis, but
they can be useful to suggest a new
testable hypothesis.

Statistical Analysis

Standard Error of the mean (SEM)

calculated mean. The smaller the standard error,

The mid-point of the replicates.
the more reliable the mean.
= sum of replicates / N

Anecdote

Null Hypothesis

When one variable changes with

another variable, so there is a relation
between them. However it may not be a
causal relation.

The statement that is tested by a

statistical test. The null hypothesis is
fixed for each test, but always says that
there is no difference or no association. Causal Relation
The null hypothesis has nothing to do
When changes in one variable cause the
with a scientific hypothesis.
changes in another variable.

NCM 01/09

How Science Works

Types of Data

Data
(measurements, singular datum)

Quantitative or Numeric Data

Qualitative or Categoric Data

(numbers)

(words)

Continuous Data

Discrete Data

Ordered Data

Categoric Data

can have any value

e.g. 7.34, -294.6, 2x105

only whole numbers

e.g. no. of atoms

can be ranked
e.g. small, medium, large

can't be ranked labels

e.g. blue, red, iron, steel

Errors

Dependent Variable
The variable you measure, to see
how it is affected by the
independent variable.

dependent variable

Types of Variable

Random Errors

independent variable

Control Variables
The variables you keep constant
because they could also affect
the dependent variable. If you
can't control some variables
(such as weather in a field
investigation), you should at
least monitor them.

Independent Variable
The variable you choose to
change, to see how it affects the
dependent variable. You may
also measure it when you
change it.

Controlled Experiment (Fair Test)

When all relevant variables are controlled, so that observed
changes in the dependent variable must be due to changes in the
independent variable.

Inaccuracies due to mistakes, poor

technique, or random variation. Random
errors are very common, but can be
improved by taking many replicates.

Systematic Errors
Inaccurate measurements in one
direction only, due to poor calibration or
poor technique. Systematic errors can
not be improved by taking more
replicates.

Zero Error
A particular kind of systematic error,
where the instrument does not return to
zero.

Control Experiment or Group

Used when a fully controlled experiment isn't possible (e.g. in a
field investigation or medical study). An experiment or group set
up to show that observed changes in the dependent variable must
be due to changes in the independent variable e.g. a placebo
group in a drugs trial.

Measurements

Bias
When the observer chooses some results
and ignores others, to support a
particular view.

Quality of Results

Replicates

Accurate Data

Reliable Data

Repeats of a measurement.

A measurement that is close to

the true value.

Findings that can be repeated

by other people, or by other
techniques, or agree with
secondary sources.

Range
The highest and lowest replicates,
or the interval between them.

Mean or Average
The mid-point of the replicates.
= sum of replicates / N

True Value
The real value of a measurement,
if it could be measured with no
errors at all.
HGS Science

Precise Data
1. Measurements that give similar
values when repeated. The
replicates therefore have a
small range.
2. Data measured on sensitive
equipment with a suitably fine
scale, e.g. 20 mm is more
precise than 2 cm.

Valid Data
Data obtained from a
controlled experiment that
addresses the stated aim.

Evidence
Data that has been checked
and is considered sufficiently
valid that we can confidently
use it to make conclusions.
NCM 01/07

upper confidence limit

95% confidence interval
mean
95% confidence interval
lower confidence limit

4th quartile (maximum)

3rd quartile
2nd quartile (median)
1st quartile
0th quartile (minimum)
A

num ber of tim es eac h

v alue oc c urs

mean
95% CI

95% CI

leaf area (mm)

normal
distribution
curve

values

Open

normal

skewed

flattened

Shaded

bimodal

text