1

MEASURES AND ERRORS IN EXACT (PHYSICAL) SCIENCES

PHYSICAL SCIENCES ARE EXPERIMENTAL

Physical Sciences and the Scientific Method

A confusing aspect of science is that not all fields of science arrive at conclusions in the
same way. The physical sciences, like physics and chemistry, use experimental forms
of the "scientific method." The physical sciences do experiments to gather
numerical data from which relationships are derived, and conclusions are made.
The more descriptive sciences, like zoology and anthropology, may use a form of the
method that involves gathering of information by visual observation or interviewing.
What is common among all sciences, however, is the making of hypothesis to explain
observations, the gathering of data, and based on this data, the drawing of conclusions
that confirm or deny the original hypothesis. The difference is in what is considered
data, and how data is gathered and processed.

Data for a physical scientist are numbers. The numbers are often plotted on graphs.
Graphs can be used to derive equations that can be used for making predictions. Data,
for an anthropologist, could be a recorded interview. Interviews can be compared to
other related information. The distinction between the exact sciences and other
sciences lies in the fact that exact (or physical) sciences mainly use numbers to
measure and calculate results, and the others mainly use descriptions and
inferences to arrive at results.

What is the Experimental Scientific Method?

The steps listed below are followed systematically to investigate observations that can
be tested with the experimental method. Not all questions can be dealt with by the
experimental scientific method. Tests done to check hypothesis are called experiments.
To design a suitable experiment we must think about the things that affect the system
you want to investigate (variables). This requires thought, information gathering, and a
study of the available facts relating to the problem. As experiments are carried out, the
data that measure the effect of variables are recorded. Using these data results can be
calculated. Results are presented in the form of tables or graphs. These results will
show trends related to how the variables affect the system. Based on these trends,
conclusions are drawn about the hypothesis being tested.
The existence of "cause and effect relationships" in nature is what makes
experimental science possible. Hypothesis can only be verified using the scientific
method described here if there is a cause and effect relationship between the variables
you have chosen and the system you are studying.

What Is Experimental Science?

Experimental science is actually the search for cause and effect relationships in nature.
A hypothesis is our best guess at what this cause and effect relationship is. Conclusions
allow us to predict the result of future cause and effect relationships. These relationships
are expressed in terms of mathematical equations (laws) and a coherent group of laws
and general propositions used as principles of explanation for a class of phenomena:
constitute a theory.
 2

Technology is the area that applies the findings of the sciences to produce machines,
or do things for us.

MEASUREMENTS AND UNITS

Basic concepts

The experimental method of physical sciences requires measuring. So what is “to

measure”? Let’s state some few concepts about this fundamental concept.

A magnitude is anything that can be measured and to measure is to compare a

magnitude with an arbitrary standard called the Unit for that specific magnitude.
Strictly speaking the concept of “measuring” precedes the concept of magnitude.

A Measurement is the act and result of a measuring experiment; it tells me how many
times the unit is contained into the magnitude we are measuring. A mass of 5,4
kilograms means that the unit “kilogram” fits 1,5 times in the magnitude we are
comparing to it.

The measure is the final result being informed (after averaging for example).

Any measure or measurement is given as a number (some magnitudes need more than
just a number) and a symbol for the unit being used, for example: 5 m - 36 s - 14 km

The International System or “Système International” (SI)

As stated before units are absolutely arbitrary. To day all scientists and most countries
have agreed in using the so called SI (International) system. A set of four units for four
magnitudes respectively, has been adopted as the base from which all other units are
derived. These four fundamental units are:

• The METRE (m) for LENGTH

• The SECOND (s) for TIME and
• The KILOGRAM (kg) for MASS.
• The AMPERE (A) for ELECTRIC CURRENT

There is nothing special in them. Their power lies on being standards universally
accepted. We will not give the strict meaning of each of these units, but just say that
every country has its own set of perfectly checked standards and that every measuring
instrument is directly or indirectly calibrated (checked) against them.

SUB-UNITS AND PREFIXES

The magnitudes that have to be measured both in everyday life and in a science lab
differ dramatically in their orders of magnitude ranging from thousands of billions
to thousandths of billionths and more. To avoid the use of a huge number of zeros,
 3

science uses very frequently the scientific notation (exponential notation) as

learned in previous maths courses. In addition, universally accepted submultiples
of the units are regularly used. The following chart lists the most commonly
employed in the metric system.

Metric system prefixes

Prefix Symbol Meaning

femto- f x 1/1,000,000,000,000,000 (10-15)
pico- p x 1/1,000,000,000,000 (10-12)
nano- n x 1/1,000,000,000 (10-9)
micro- µ x 1/1,000,000 (10-6)
milli- m x 1/1,000 (10-3)
centi- c x 1/100 (10-2)
deci- d x l/10(10-1)
kilo- k x 1,000 (103)
mega- M x 1,000,000 (106)
giga- G x 1,000,000,000 (109)
tera- T x 1,000,000,000,000 (1012)

Thus, if a potential difference is just 2 millionths of a volt (0,000002 V) it will be

represented as 2 x 10-6 V or as 2 microvolt μV

0,000002 V = 2 x 10-6 V = 2 μV (2 micro volts)

Another example: if the frequency of a radiation is 5.000.000.000 hertz it will be written

as 5 x 109 Hz or as 5 GHz

5.000.000.000 Hz = 5 x 109 Hz = 5 GHz (5 giga hertz)

EXPERIMENTAL ERRORS

When we measure several times, say, the length of a table, the different measurements
will generally be different. Nevertheless, there is just one possible value for the
length: which of them is the proper one? How can we get the “true” value of the
length? What we do is to average the data (there is a theoretical reason for doing
s but we will not explain it).In fact, doing so, we are not getting to the “true” but
to the “most probable” value.

The true value – the value that would be obtained in the absence of errors – can only be
known in a practical sense through standards, which are true values by agreement,
rather than real true values. Measured values can approach true values more and
more through careful experimentation designed to eliminate more sources of error,
and confirmed by duplication of results in other experimental setups.

If there are N measurements X1, X2, …, XN, the accepted value (most probable) will
be:
 4

The uncertainty or error (ε) of each individual measurement is found subtracting the
average (most probable value) to it.

As can be appreciated, errors can be either positive or negative.

Random Errors

Some randomness is always present in nature. No two measurements are exactly the
same so errors are impossible to predict and hence to eliminate. If we measure the
length of a 10 m courtyard several times with a meter tape, its is very probable that we
will not put the tape at the very same point each time we move it, so the results cannot
be predicted. In this case we are dealing with a random error (accidental error) and
consequently we will not be able to get rid of it. This kind of error is always present in
an experiment and cannot be eliminated. They are “at random” so they can be, as
previously stated, either negative or positive. Random errors cancel each other
statistically so the more times we measure, the closer to the “true” (“most probable”)
value we get. So they can be minimised despite their inevitability: precision increases
with the number of measurements we make.
Scientists use various statistical tests to determine if the difference between runs is due
to randomness in nature, or to the way they are doing the experiments.

Systematic Errors

Uncertainty is inevitable in every measurement as we have seen. Some uncertainties or

errors appear because of the skill of the operator, or the use of an inadequate measuring
devise. Inspecting carefully the procedures these errors can eventually be evaluated and
eliminated. These errors are called systematic errors (they are repeated once and again
until they are taken into account and eliminated). A systematic error will be always of
the same sign: positive (gives results higher than the true value) or negative (readings
are smaller than true value. They will have the same value. For example if there is a mm
missing in a ruler, all the lengths that are made with it will be 1 mm bigger than they
should (positive error). On the contrary, if the pointer of an unloaded spring balance
stands below the zero mark of the scale, all weights measured will be less than the real
value (negative error). They cause the average measure to be shifted from the correct
value
Systematic errors are very frequently harder to find. Your data and results may look
consistent and reproducible. Here you may be doing something you are not aware of,
that is causing all your measurements to be off the same amount. As stated before, if
you were not aware that a piece of your ruler had been cut off and now starts at 1"
instead of 0", all your measurements would be one inch too long. This is a systematic
error because all your data is affected the same amount, and in
the same direction. This “zero point error” for example, is
eliminated re-calibrating the measuring instrument.
 5

Parallax errors are another easy-to-find systematic error. They appear when a pointer
moving on a scale is looked from a slanted position. The line of vision eye-pointer falls
on a different mark from the one is looked by looking at the scale at right angles.
Whenever the index mark (pointer) and scale lie in different planes, the "reading" you
observe can change as you move your head to different positions. This effect is called
"parallax." To eliminate its effect on measurements, be certain that your line of sight is
exactly perpendicular to the scale.

One way to check for systematic errors is to run experiments of a different design that
should give the same answers. Scientists often do different kinds of experiments to
cross check their results. Another way to locate errors is to have an independent
investigator repeat your experiments. Others should get the same results you did.

There are four main sources of experimental uncertainties (experimental errors):

• Limited accuracy of the measuring apparatus (decimal places in a digital
balance).
• Limitations and simplifications of the experimental procedure (assume that
there is no air friction if objects are not moving fast)
• The observer (lack of experience, inattentiveness, etc.)
• Uncontrolled changes in the environment. (Small changes of the temperature
and the humidity in the lab).

ACCURACY AND PRECISION

These two terms appear as synonyms even in dictionaries but they have very different
meanings.

Accuracy

When we state that a length is 100 m, we just mean that it lies somewhere between 99 m
or 101 m. 99 m < 100 m < 101m.

In case we use a better instrument (for example a metre tape that has dm marked on it)
we can compare with fractions of the unit for example we can state a length as
100,5 m. In this case we have a better accuracy (we get closer to the “real” value).
Now the length lies in a shorter interval 100,4 m < 100,5 < 100,6 m

Accuracy shows how close the value of a measurement is believed to be to the true
value. It depends on the estimated magnitudes of both systematic and random
errors. If the metre tape is graduated every millimetre instead of every centimetre
our accuracy will increase

Precision

Precision, on the other hand, refers to how big the scatter of repeated measurements
about the mean value – depends on the magnitude of random errors only. It is different
from sensitivity, which is the smallest change in the variable being measured that can
be sensed by an instrument.
 6

If we measure the same length with the same instrument many times probably we will
not get to the same result because of random errors But if the different results are all
of them close to each other ( for example in the previous case 110,4 m, 100,5 m,
100,4m, 100,6 m, 100,6 m, 100,5 m) the measurements have been precise. Evidently
we can be quite sure that things seem to be properly done, that is results are predictable.
The following diagrams may clarify these concepts

Random and systematic errors

(Average shifted from true value)

Just random errors

Briefly

• A result is said to be accurate if it is relatively free from systematic error

• A result is said to be precise if the random error is small

ABSOLUTE AND RELATIVE ERRORS

Errors must be minimised both from a technological and a scientific point of view. An
error of 1 mg when weighing a powerful psychoactive drug in a medicine can lead to
unexpected and dramatic results. On the other hand, nobody would complain if a ten kg
package of the same drug (bought for the preparation of a medicine) is lacking 1 mg. In
both cases the error is of just 1 mg but the relative importance is absolutely different.
 7

An error in 1 mm when measuring a ball bearing 2 mm wide is unacceptable. The

length of a football field (100 m) will be the same for practical purposes although it is 1
mm shorter than it should be! In both cases for both examples the error is the same. We
call this uncertainty the absolute error (ε).

The relative error (η) is the ratio (generally expressed as a percentage) between the
absolute error the actual measure.

So, for the ball bearing’s diameter uncertainty

ε = 1 mm and η = (1 mm / 2mm) 100 = 50 % (more than relevant, unacceptable)

For the football field uncertainty (100 m or 1x 108 mm)

ε = 1 mm and η = (1 mm / 1 x 108) 100 = 10-6 % (incredibly accurate)

QUOTING RESULTS AND ERRORS

The following are some simple rules and examples to show how to quote a measure
(result) and its error.

• Generally errors are stated to one significant figure (although if this significant
figure is around one or two then two significant figures may be used).
• Quote result to same significance as error
• When using scientific notation, quote value and error with the same exponent

• If value is 44 kg and error is 5 kg ⇒ 44 ± 5 kg

• If value is 128 m and error is 32 m ⇒ 30 ± 130 m
• If value is 4.8x10-3 V and error is 7x10-4 V ⇒ (4.8 ± 0.7)x10-3 V
• If value is 1092 s and error is 56 s ⇒ 1090 ± 60 s
• If value is 1092 cal and error is 14 cal ⇒ 1092 ± 14 cal
• If value 12,345 g and error is 0,35 g ⇒ 12,3 ± 0.4 g

Never ever over quote results to a level inconsistent with the error:
36.678935372± 0.5 is absolutely nonsense

ESTIMATING READING ERRORS

The following are examples to show how to estimate the errors according to the
measuring situation involved

1- Oscilloscope – Error related to width of trace

Read: 3.8 divisions. (If 1V/division = 3.8V)

Trace width is ~0.1 division = 0.1V

Result: (3.8± 0.1) V

 8

2- Digital meter – Error taken as ± 5 in next significant figure

Result: (3.360± 0.005) V

3- Analogue meter – Error related to width of pointer

The value is 3.25V (work it out). The pointer has width 0.1V

Result: (3.3± 0.1) V

4-ALinear scale (e.g. a ruler)

You have to estimate precision with which measurement

can be made but it might be a subjective choice (how
accurate you think you can be)

Result: (16.77± 0.02) cm

The reading error may be dependent on what is being

measured. In this case the use of greater precision equipment
may not help reduce the error. What is the use of being
scrupulous in the example to the right? Which is the length,
the upper or the lower edge’s length?

PROPAGATION OF ERRORS

In general we will calculate a result using a formula which has as an input one or more
measured values. For example: volume of a cylinder V = π r2h How do the errors in the
measured values feed through into the final result? You will be given approximate
equations or “formulae” to be used without any further discussion. In them

• A, B, C and Z are the absolute values

• εA, εB, εC and ε Z are the absolute errors in A, B, C and Z
• ηA is the relative or fractional error in A (∆ A/A)
• εA and A will have the same units obviously

• Assume errors in numerical or physical constants (e.g. π, e, c etc) are much

smaller than those in measured values – hence can be ignored.

For additions / subtractions:

Z=A± B±C εZ = εA + εB + εC2

For multiplications

Z = A.B.C or Z = A.B/C ηZ = ηA + ηB + ηC
 9

If powers are involved

Z = An.Bm.Cl ηZ = n.ηA + m.ηB + l.ηC

Some Practical Examples on Error Manipulation

Example 1: Area of a circle with r = (5 ± 0, 5) m. As A = π r 2 A = 78,5398m2. The

error is estimated as follows:

ηr = 0,5 / 5 = 0.1 ηA = 2ηr = 0,2

Hence εA = ηA A = 0,2 . 78,5398m2 = 16 m2 Answer: A = 79 ± 16 m2

Example 2: Perimeter of a rectangle where L (length) = (4 ± 0.2) m and W (width) =

(5 ± 0.2) m

P = L+L+W+W so P = 18m εP = 0.2+0.2+0.2+0.2 = 0.8

Final result: P = (18.0 ± 0.8) m

AGREEMENT BETWEEN TWO RESULTS:

Error estimates are necessary to be able to say whether the two independent
measurements of the same thing agree within the stated errors or disagree. For
example, two students measured the melting point temperature of ice and obtain results
of -0.3 °C and +0.8 °C. Without an estimate of error, we cannot say whether these
measurements agree.
Suppose the results had been stated with errors: Tmelt = -0.3 ± 1.0 °C and
Tmelt = +0.8 ± 1.0 °C. Since the first result admits values between -1.3 °C and 0.7 °C and
the second one between -0.2 °C and 1.8 °C, there is an overlap (between -0.2 °C and
0.7 °C) and the results are in agreement within experimental errors

In case that the errors were stated as ± 0.2 °C there will be no overlapping: the first
result would be lie between -0.5 °C and -0.1 °C and the second between +0.6. °C and
+1.0 °C. Evidently there is no overlapping zone and the values do not agree.
If one cannot find an overlap between the error bands, then the results do not agree with
each other.

USING SIGNIFICANT FIGURES

Significant figure rules are a set of rules for approximating the propagation of
uncertainty in scientific or statistical calculations. These rules can be used to find the
appropriate number of significant figures to use to represent the result of a calculation.
If a calculation is done without analysis of the uncertainty involved, a result that is
written with too many significant figures can be taken to imply a higher precision than
is known, and a result that is written with too few significant figures results in an
 10

avoidable loss of precision. Understanding these rules requires a good understanding of

the concept of significant and insignificant figures.

Significance arithmetic rules rely on the assumption that the number of significant
figures in the operands gives accurate information about the uncertainty of the operands
and hence the uncertainty of the result.

An important caveat is that significant figures apply only to measured values. Values
known to be exact should be ignored for determining the number of significant figures
that belong in the result. Examples of such values include:

• integer counts (e.g., the number of oranges in a bag)

• legally defined conversions, such as international currency exchange
• constants that are defined arbitrarily (e.g., a millimetre is 0.001 meters)
• scalar operations, such as "tripling" or "halving"
• mathematical constants, such as π and e

Physical constants such as Avogadro's number (number of atoms in 12,0000 g of carbon

atoms) have a limited number of significant digits, because these constants are only
known to us by measurement.

The rules for estimating significant figures were given in the first tutorial
(00 maths for physics.doc)

QUESTIONS AND PROBLEMS

1- Why sociology and physics are both sciences?

2- Why physics is an exact science but sociology is not?
3- Causality is in the very foundations of science. Explain
4- Physicists and engineers study mechanics, heat, electricity, etc. What do you
think makes the difference between both careers?
5- State the difference between measurement and measure.
6- Change as required:
a- 15,6 m to km
b- 125 ng to kg
c- 125 ng to g
d- 34,6 THz to Hz
e- 20 m / s to km / h

7- Find the most probable value (mean value) and the errors (deviations), for the
current through a circuit if the different experimental data obtained by Jean are:

Current 25 25 25 24 24 25 25 25 25 252
(mA) 0 4 4 9 8 0 2 0 1

State which errors are positive and which are negative.

8- Quote properly the following results:

 11

a- The amount of copper in a sample as determined by a method with a 0,5

% accuracy is 35 mg.
b- The content of a bottle of wine is 975 cm3 with a 2 % allowance (percent
relative error).

9- Steel balls for a ball bearing should have the same diameter (15 mm) within a 2
% allowance. Which are the maximum and minimum values allowed for the ball
bearings?

10- The figure shows the scale of a voltmeter and the position of the pointer as seen
from above (top view).

What is the reading when Tommy looks:

a- At right angles (just in front of the pointer)
b- At 30° right of the pointer
c- At 45° from the left side
Estimate the pointer’s width as covering 1/5 of the distance between marks.
Trace the lines of vision and read your result in the paper.

11- Gather in foursomes to carry out the following experiment. Measure the length
and width of one of your desks to the next centimetre and to the next millimetre.
With your results calculate its surface’s area and write down your results
properly. Put them in a chart as shown below. Measures and results must show
their errors. Use significant figures for calculations. Discuss your results briefly.

L (cm) W (cm) A (cm2) Agreements L (mm) W (mm) A (mm2) Agreements

Pupil 1
Pupil 2
Pupil3
Pupil 4

12- Two students are tested to assess their skills using a two pan balance. The
sensitivity of the balance is ± 0,01 g (10 mg).
They are both given in turns the same object and asked to find its mass ten
times. You are given the results of both pupils and the mass of the object as
previously measured by the teacher with an electronic balance (a tenfold more
sensitive).

Jean’s Masses 1,54 1,5 1,52 1,5 1,52 1,5 1,5 1,53 1.5 1,54
(g) 3 5 2 5 4 2 3
Paul’s Masses 1,49 1,5 1,43 1,5 1,48 1,4 1,5 1,50 1,5 1,46
(g) 5 0 3 9 1 5 0
 12

Teacher’s value = 1,495 ± 0,001 g

a- Calculate and quote properly the measures reported by both students

b- Compare their results in terms of accuracy and precision
c- Discuss the “personal equation” the human factors that produce these
differences.