Science

Common mistakes in working scientifically

11 mistakes learners make with working scientifically, each one named and explained.

All common mistakes

A result only answers the question when one thing was changed and everything else was held still.

A bar chart used for continuous data

What it looks like

Temperature against time is drawn as separate bars, so the values in between look as though they do not exist.

Why it happens

Bar charts are the first chart anyone learns and they work for almost everything in primary school.

A worked example

Question. Temperature measured every 2 minutes

A common answer. Draw a bar for each time

The answer. Draw a line graph with time along the bottom

Why. Time and temperature both run smoothly, so the points in between are real.

Put the two side by side

Bars suit categories such as eye colour.
Lines suit measurements such as temperature.

Ask whether a value halfway between two readings could exist. If it could, use a line.

A conclusion the results do not reach

What it looks like

You tested two fertilisers on cress and conclude that this fertiliser is best for all plants.

Why it happens

A conclusion feels weak unless it sounds general, and the experiment did work.

A worked example

Question. Two fertilisers, tested on cress, over one week

A common answer. Fertiliser A is best for growing plants

The answer. Fertiliser A grew taller cress than fertiliser B over one week

Why. The conclusion may only mention what was actually tested.

Put the two side by side

The wide claim covers plants you never grew.
The narrow claim is one nobody can argue with.

Write the conclusion using the words from your own method.

A link read as a cause

What it looks like

Ice cream sales and sunburn rise together, so you conclude that ice cream causes sunburn.

Why it happens

Two things moving together is exactly what a cause looks like on a graph, and the graph is real.

A worked example

Question. Ice cream sales and sunburn, both rising in June

A common answer. Ice cream causes sunburn

The answer. Both rise with hot, sunny weather

Why. A link is a reason to look for the cause, not the cause itself.

Put the two side by side

A correlation tells you two things move together.
A cause needs a mechanism you can explain.

When two things rise together, ask what third thing might be moving them both.

Repeat readings treated as a cure for a wrong method

What it looks like

You say the result is accurate because you repeated it five times, although the balance was never zeroed and every reading is 2 g too high.

Why it happens

Repeating is drilled as the thing that makes results trustworthy, so it gets used as a general repair rather than for what it actually fixes.

A worked example

Question. A balance reading 2 g with nothing on it

A common answer. Repeat the measurement five times and take the mean

The answer. Zero the balance, then repeat five times and take the mean

Why. Repeats reduce random scatter. They cannot remove an error that repeats itself.

Put the two side by side

Five readings of 52 g are precise.
They are still 2 g from the true 50 g.

Repeats deal with scatter. Only checking the equipment deals with a shift.

The measuring cylinder read at the top of the curve

What it looks like

Every volume comes out a millilitre or two high, because the reading is taken where the water climbs the glass rather than at the bottom of the curve.

Why it happens

The water at the edge is the part at eye level and the part you notice first.

A worked example

Question. Water in a measuring cylinder, curve between 24 and 25 ml

A common answer. Read 25 ml, at the top of the curve

The answer. Read 24 ml, at the bottom of the curve, with your eye level with it

Why. The curve is the water gripping the glass, so the flat middle is the real level.

Put the two side by side

Reading the edge adds a little every time.
Reading the bottom of the curve gives the same answer to anyone who repeats it.

Get your eye level with the liquid and read the flat part.

The odd result left inside the mean

What it looks like

Readings of 24, 25, 24 and 62 are averaged to 33.75, and the mean now describes none of them.

Why it happens

Leaving a number out feels like changing the result to suit yourself, so every number is kept.

A worked example

Question. 24, 25, 24 and 62

A common answer. Mean = 33.75

The answer. Leave 62 out as an anomaly, mean of the rest = 24.3

Why. An anomaly is a reading the method went wrong on, not a reading you dislike.

Put the two side by side

33.75 is close to nothing you measured.
24.3 is close to three of your four readings.

Identify anomalies before averaging, say which ones you left out, and say why.

The points joined up instead of a line of best fit

What it looks like

Your graph is a zigzag from point to point, so the trend disappears and a reading taken between two points means nothing.

Why it happens

Joining the dots is what a dot to dot taught, and it does use every point, which feels honest.

A worked example

Question. Eight points that rise with some scatter

A common answer. Join each point to the next with a ruler

The answer. Draw one straight line with the points balanced either side of it

Why. The line shows the pattern. The scatter is the measuring, not the science.

Put the two side by side

A zigzag says the quantity jumps about.
A line of best fit says it rises steadily and the readings wobble.

One line through the trend, with roughly as many points above it as below.

The small marks on the scale counted wrongly

What it looks like

You read a thermometer marked every 2 degrees as though every mark were 1 degree, so the reading is out by half.

Why it happens

Most scales a learner meets are marked in ones, so the eye fills in ones without checking.

A worked example

Question. A scale with 5 gaps between 20 and 30

A common answer. Each small mark is 1 degree

The answer. Each small mark is 2 degrees, because 10 degrees is split into 5

Why. Count the gaps between two labelled marks, then divide.

Put the two side by side

Assuming the marks are ones is a guess.
Counting the gaps and dividing is a check.

Before the first reading, work out what one small division is worth.

The thing you change mixed up with the thing you measure

What it looks like

You call the temperature the independent variable in an enquiry where temperature is what the thermometer reads at the end.

Why it happens

Both variables are numbers that change during the experiment, so the two roles blur.

A worked example

Question. Does more insulation keep water warm for longer?

A common answer. Independent variable: the temperature of the water

The answer. Independent variable: the number of layers of insulation

Why. You choose the layers. The temperature is what the layers do to the water.

Put the two side by side

The one you set before you start is the independent variable.
The one you read off at the end is the dependent variable.

Ask which one you could write into the plan before touching any equipment.

The variables put on the wrong axes

What it looks like

What you changed is up the side and what you measured is along the bottom, so the graph reads backwards against every other graph in the class.

Why it happens

Nothing in the numbers says which way round they go, so the first axis drawn takes whichever column was written first.

A worked example

Question. Layers of insulation against temperature

A common answer. Temperature along the bottom, layers up the side

The answer. Layers along the bottom, temperature up the side

Why. What you chose goes along the bottom, what you measured goes up the side.

Put the two side by side

Swapped axes make the gradient mean something else.
The usual order lets anyone read your graph at a glance.

Independent along the bottom, dependent up the side, every time.

Two things changed at once

What it looks like

Your plan changes the mass and the height of the ramp together, so whatever the result is, it cannot say which change caused it.

Why it happens

Changing two things feels like covering more ground in one go, and the plan still looks and reads like a proper experiment.

A worked example

Question. Testing whether a heavier trolley rolls further

A common answer. Use a heavier trolley and make the ramp steeper

The answer. Use a heavier trolley and leave the ramp exactly as it was

Why. One change gives one explanation, two changes give two.

Put the two side by side

Two changes leave two possible causes.
One change leaves one.

Name the one thing you will change, then write the list of everything you are holding still.

WAJD spots these patterns in your child's answers and names the one behind their wrong answers, instead of just marking them wrong.

More science topics