Suspend a small ball from a string, pull it aside and let go. It swings back and forth. Now make the string longer. Will the swinging slow down? What if you replace the ball with a heavier one instead? Intuition suggests an answer, but different people may predict different things. How can we decide which explanation works better?
Physics investigates patterns in nature: how objects move and interact, and how matter, light and other physical systems behave. It connects observations with models that produce predictions we can test. This helps us explain familiar events and make reasoned predictions about things we have not yet observed. OpenStax: the scope of physics and the role of models.
Turn “it swings more slowly” into a measurable question
A ball hanging from a string is a simple pendulum. To compare two pendulums, we must agree on what to compare. A longer path, the speed at the lowest point and the time of one complete oscillation are different quantities. For our question, it is useful to measure the period: the time it takes the ball to return to the same position moving in the same direction. Travelling from one extreme to the other is only half a cycle.
The initial question “will it slow down?” becomes more precise: how will the period change if we quadruple the distance from the pivot to the centre of the ball? We keep the ball’s mass and its small initial displacement the same. We have isolated one change whose effect we want to test.
A record stating “this ball completed ten full oscillations in this many seconds” is an observation if we actually measured it. Saying “the longer string caused a change in period” already includes an explanation. To support it, we need to know what else changed in the apparatus and how reliably we measured the time.
A model retains the relevant features and names its simplifications
A real pendulum has a string of a particular thickness, a ball of a particular shape, a support and surrounding air. Describing every detail is cumbersome for a first prediction. We start with a model: imagine the mass concentrated in a small ball, the string light and unable to stretch, and air resistance and friction small enough to neglect.
When the ball is displaced, gravity has a component along its arc that pulls it towards the lowest point. The ball speeds up and passes that point, beginning the next half of the oscillation. For small displacements, this model gives a simple relationship between length and period. Pendulums like these are also used in university demonstrations. Idaho State University: the simple pendulum model.
A simplification is useful while it preserves the effects relevant to the question. Here, the model helps us compare periods. If we wanted to know how many minutes it takes the pendulum to stop, the energy losses we neglected would become central to the problem.
Keep a prediction separate from an experimental result
Before measuring, write down the prediction: in the small-displacement model, quadrupling the length makes the period approximately twice as long. Quadrupling the ball’s mass while keeping the other conditions the same leaves the period unchanged. These statements are specific enough to compare with a result. OpenStax: the period of a pendulum.
Compare the three cases below. The numbers are calculated from the model; we have not measured these pendulums here. Looking at a correctly drawn diagram does not yet test whether nature behaves that way.
Three predictions for a pendulum
Predict first: which changes the time of a complete oscillation — string length or bob mass? Compare the consequences of one model.
Reference pendulum
- Pivot to the centre of the bob
- 0.25 m
- Mass relative to the reference
- 1×
- Predicted period
- 1.00 s
Reference prediction
Four times the length
- Pivot to the centre of the bob
- 1.00 m
- Mass relative to the reference
- 1×
- Predicted period
- 2.01 s
Twice the period
Four times the mass
- Pivot to the centre of the bob
- 0.25 m
- Mass relative to the reference
- 4×
- Predicted period
- 1.00 s
The same period
Branch · With formulasOptional · you can skip this
How a formula produces a testable prediction
For an idealized pendulum at small displacements:
T ≈ 2π√((L)/(g)).
Here, T is the period in seconds, L is the distance from the pivot to the centre of the ball in metres, and g is gravitational acceleration in metres per second squared. For this example, we use the rounded value g = 9.81 m/s². With L = 0.25 m we get about 1.00 s; with L = 1.00 m, about 2.01 s.
Multiplying L by four multiplies its square root by two. Mass does not appear in the formula. These conclusions apply to the stated model; at a large initial displacement, the period also depends on the swing’s amplitude. A formula needs its assumptions to be applied correctly.
Now let the apparatus provide an answer
For your own test, compare a pendulum 25 cm long with the same pendulum lengthened to 100 cm. Measure to the centre of the ball rather than just to the knot. Each time, release it without pushing, from a small and approximately equal angle — around 5°, for example. The support should remain still.
Time ten complete oscillations and divide that time by ten. Start and stop timing as the ball passes the same marker in the same direction. Repeat the series and record every result, rather than just the one closest to your expectation. The apparatus, lengths, angle, timing procedure and records let someone else understand exactly what you tested.
You can now compare the periods and their ratio. If the longer pendulum has roughly twice the period, that supports the prediction under the conditions tested. One such result does not establish that the model works for every length, a large swing or a completely different apparatus.
Branch · DeeperOptional · you can skip this
Why the numbers will vary slightly
A person starts the stopwatch with a delay, does not always see the marker crossing precisely, and may have a support that moves a little. For the same start-and-stop timing error, timing ten cycles reduces its contribution to the time of a single cycle. Other sources of uncertainty remain.
If you repeatedly measure the length to the bottom of the ball instead of its centre, repetition will not correct that offset. In metrology, uncertainty includes components evaluated statistically from repeated measurements and components evaluated in other ways. More digits on an instrument’s display do not automatically guarantee a more accurate result. NIST: components of measurement uncertainty.
If prediction and measurement disagree, first investigate the cause: did we count cycles correctly, measure the required length and really use a small angle? The measurement procedure may need improving. The model may need an effect that we omitted for simplicity. Data should help us revise an explanation, rather than be adjusted to produce the answer we wanted.
The same approach extends far beyond the room
In real research, questions, calculations, observations and tests can lead back to one another many times. Different people carry out different parts of the work. A new apparatus reveals a previously unnoticed effect, while a new model suggests what else to measure. “Question → hypothesis → experiment → conclusion” captures a useful relationship, but it is not an unchanging recipe for all science. UC Berkeley: how science works.
In astronomy, we often cannot change a star and repeat its life under different conditions. Instead, we test the consequences of models against available observations: do predicted positions, light or other measurable features agree with what instruments record? The requirement to connect explanations with evidence remains, even though the test differs from a tabletop experiment.
This also explains the practical value of physics. Understanding motion helps us calculate mechanisms and flights; knowledge of electricity and light helps us build instruments and transmit signals. A tested model lets us choose a design with a reasoned expectation of its behaviour. We still need to check whether the future device will operate under conditions for which the model is suitable.
Return to our pendulum. If we lengthen it and also release it from a larger angle, is that a clean test of length’s effect? No: two conditions have changed. To identify the cause of the result, we must separate their effects or use a model that accounts for both.
Physics helps us turn impressions of the world into explanations with testable consequences. You can begin with a ball on a string and careful records of what was actually measured. Observing the familiar night sky provides material for the same questions on a larger scale. Next in the course, we will examine how to describe motion and speed precisely.