# Quick question about uncertainty in using a stopwatch

## Homework Statement

Watching an object fall, along a ruler, using a stopwatch to take the time from when the object flies passed the upper part of the ruler, and again when it hits the floor (bottom of the ruler)

## The Attempt at a Solution

My friend is doing introductory mechanics and says that since using a stopwatch requires someone to react to something visually, there'll have to be an uncertainty of about 0.2 seconds, both at the beginning and at the end, for a total of 0.4 seconds. From my experience this is correct, hence why they shouldn't have done it this way, as the fall takes about 1 sec, so the fractional error is enormous. But I'd be happy to have it confirmed before I agree with him. Is this correct?

(People in his group argue for 0.05 seconds in total, which given how they performed the experiment, sounds unreasonable to me)

Orodruin
Staff Emeritus
Homework Helper
Gold Member
2021 Award
There are several things going on here. The visual reaction time is only one thing, but is somewhat countered by the person actually expecting the fall. Unless you have a very long ruler, the best thing would be to capture the process on video and then go frame by frame to determine the times more precisely.

One thing about reaction times is that both of the reaction times will be delayed. If you had exactly the same reaction times to both measurement points, then this effect would cancel out in the difference. Of course, that is not going to be the case, but there is an overall systematic error (mean reaction time - which is the same for both times) and a random error (spread of reaction times - which is different from both points but I would expect it to be smaller than the systematic error) at play here.

There are several things going on here. The visual reaction time is only one thing, but is somewhat countered by the person actually expecting the fall. Unless you have a very long ruler, the best thing would be to capture the process on video and then go frame by frame to determine the times more precisely.

One thing about reaction times is that both of the reaction times will be delayed. If you had exactly the same reaction times to both measurement points, then this effect would cancel out in the difference. Of course, that is not going to be the case, but there is an overall systematic error (mean reaction time - which is the same for both times) and a random error (spread of reaction times - which is different from both points but I would expect it to be smaller than the systematic error) at play here.

If I had done the experiment, and had the same tools, I would have used a phone camera and loggerpro. But they choose this way, so better methods wont' change much at this point.

But you're saying that because the person is expecting it, the reaction time is similar at both ends and is thus reduced? Doesn't expecting it, and viewing it from a distance make reacting too quickly just as likely?

Orodruin
Staff Emeritus
Homework Helper
Gold Member
2021 Award
Expecting it is likely to reduce the bias from reaction time alone. The random part (spread in reaction times) which is actually the only one that enters into the measurement will probably remain.

Expecting it is likely to reduce the bias from reaction time alone. The random part (spread in reaction times) which is actually the only one that enters into the measurement will probably remain.

Excuse my ignorance, but what is the spread of reaction time on average for a human?

haruspex
Homework Helper
Gold Member
Excuse my ignorance, but what is the spread of reaction time on average for a human?
@Orodruin is not professing to know that, and certainly I do not, but it would be safe to assume it is much less than the individual reaction times.
If the experiment was repeated several times with the same parameters then you could find the spread from the data.

Orodruin
Staff Emeritus