Understanding Scientific Measurements: Accuracy, Precision, and Uncertainty, Study notes of Astronomy

Download Understanding Scientific Measurements: Accuracy, Precision, and Uncertainty and more Study notes Astronomy in PDF only on Docsity! Scientific Measurement & Data Understanding measurements through making measurements Author: Sean S. Lindsay Version 1.1 created August 2018 Version 3.1 by S. Lindsay in September 2019 Learning Goals In this lab, students will engage with taking direct and indirect measurements, examining collected data to understand accuracy and precision, and performing remote distance calculations similar to stellar parallax distance determinations. In this lab, students will • Perform simple scientific measurements • Carefully record data • Examine the distribution of data points over several different variables • Explore the concepts of accuracy, precision, and uncertainty • Perform remote distance calculations • Play a game Materials • Game board, bean bags, direct distance measurement tools, and devices to measure angles, including a protractor. • Prepared Google response form: 1. Background 1.1 Introduction to Scientific Measurement How do we go about determining things like the number of galaxies in the universe, how much “stuff” the universe contains, the value of the gravitational constant of the universe, and the distance to the stars or how bright they are? Where do we get the evidence to evaluate some of our biggest astronomical questions, such as how did the universe, the solar system, and Earth get their start? To answer all of these questions, scientist must collect observations and empirical evidence to test the predictions of their hypotheses and theories against. The collection of these observation and evidence almost certainly involves measuring some quantity of something. Whether that be the amount of light being received by a distant supernova, the total amount of mass (both luminous and dark) in a galaxy cluster, or the distance to the stars of the Milky Way, observations must be made, and data must be collected. How well we can measure the related quantities is intimately tied to how well we can evaluate our deepest scientific understandings of the universe. Answering some of the above questions is a daunting task, and often times we only have a ballpark idea of what the answer might be. In more intriguing cases, we find ourselves in completely new scientific territory where we have little theory to rely upon. Here, science must tackle challenging questions in face of the unknown. There may be no known methods from which we can start trying to solve the problem. We are left to rely on our own creativity and ingenuity to develop new problem-solving techniques. If we are clever, we often can devise multiple ways to address a question. Some of the methods will provide more reliable answers than others. With hypothesis, theory, and ingenuity as our resources, we then have to determine, using the principles of science, which methods are best. This frequently comes down to how those methods make their measurements, what they are measuring, and which one gives the explanations we have the most confidence in. This means we need to understand the empirical evidence, frequently just called data in science, itself. How reliably does it give an answer close to the correct one, and how much does the data vary from measurement to measurement? In this lab, we will explore the some of the techniques scientists and astronomers use to begin addressing how to approach scientific problems and gather the required evidence. We will explore the difficulty in making measurements, the uncertainty and errors associated with measurement, and how we can compare measurements of the same thing made with different tools or techniques. 1.2 Scientific Measurement Measurement often is thought of as a simple task. For most people, daily life measurements are simply measuring how long something is with a measuring stick, how much volume you need with a measuring cup, the mass of something with a scale or a mass balance, or how long something takes with a stopwatch. If you are off by a few centimeters, milliliters, grams, or seconds, then so be it, there was no real harm done. In science, however, we strive to arrive at the real value as close as possible, or rather, to have measurements with high levels of accuracy. To have an accurate measurement, or set of measurements, is to be close to the real value with the measurement, or the average of the set of measurements. To accomplish accurate measurements, much more care and sophisticated instruments are needed than you experience in everyday life. In science, we also never rely upon a single measurement. Science demands that we make measurements repeatedly to check and re-check our evidence. In these repeated trials, scientist aim for measurement methods that consistently give values that are nearly the same, or rather, have high levels of precision. How close they are together, or how precise your measurements are, determines how certain we can be that our evidence provides a good test for the predictions of our hypotheses and theories. In other words, the precision of our measurements quantifies the uncertainty in our measurements, or rather, what is the probability that our measurement is accurate within a certain range of values. A graphical representation of these concepts is provided in Figure 1. Figure 1. A graphical representation of accuracy and precision. For repeated measurements, the accuracy is how far the average of the measurements is from a reference value, represented by the vertical line through the measurement distribution. The precision is the spread of all measurements. A quantification of the width of the distribution of measurements would be the uncertainty. 1.3 The Difficulty of a Scientific Measurement – A Direct Measurement Example Scientific measurement is a deceptively difficult task with a large amount of nuance. Consider the simple task of measuring exactly how long a room is. For simplicity, let us imagine that the room’s floor print is a square, so that the length and width are equal. How would you go about measuring the length of the room? Likely, you would get a measuring tape, a meter/yard stick, or some other standard of length measurement and just see how long the room is. Since we are working in the world of science, let us imagine that you chose to measure the length in meters. Unfortunately, you only have a meter stick in your room, but nevertheless, it seems like a good one with centimeters (cm) and millimeters (mm) clearly marked. So, you set to the task with confidence that you can determine the length of the room. You measure 5 full meters and have a bit leftover where the meter stick is too long to measure the remaining distance to the wall. Being clever, you mark the 5-meter mark on the floor and then measure from the wall out to that point to and find that length to be 35.6 centimeters. Proudly, you declare the length of the room to be 5.356 meters (m) in length! How certain are you that 5.356 m is the true length of the room? You recall that the uncertainty for any measurement is half the smallest division, so +/- 0.5 mm, or 0.0005 m. That seems wildly accurate and precise for a single measurement, and you suspect that your own measurement method prevented such high precision. Thinking on your method of measurement, first you begin to wonder how good of a job you did when you moved the meter stick from one position to the next. Did you truly put it exactly where it left off? Did that amount vary each time you did it? You certainly weren’t perfect, so how much were you off by? A few millimeters? Over the multiple measurements that could easily add up to being off by a centimeter or more. You also are unsure if it was always extra length added, or did you underestimate sometimes? Worried, you repeat the task again and this time you get 5 full meters, but the extra measurement this time comes out to be 34.3 cm given a total length of 5.343 m. A different value by 5.356 m – 5.343 m = 0.013 m! Clearly, you can’t be certain about the length of the room from these measurements alone. A fundamental rule of measurement dawns on you, … All measurement has some level of uncertainty in it. A scientist’s goal is to reduce the uncertainty as much as possible. Entire theories may depend on it! A few days later, you are still curious about how long your room really is. You ask a friend to come over and measure the room. You have shared your concern with not being able to perfectly move the meter stick and how you measured that last remaining bit and aren’t sure if that was the best method. Your friend decides that the best way to avoid this is to lay out a string from one end of the room. She then measures the length of the string, and she is very careful to mark each meter with marker before moving the meter stick. Using this method your friend determines the room to be 5.352 m long. You double check and measure the length of the string as well and get 5.348 meters long. This time only a difference of 5.352 m – 5.348 m = 0.004 m, or 4 mm! You are beginning to feel more confident in the length of the room, but your friend has a few concerns. She noticed that there is a bit of extra material at the start and end of the meter stick, so the meter stick is likely slightly longer than a full meter. Being an acute observer, she noticed that the meter stick is a bit worn on the ends, and that she measured along the edge while you measured in the middle of the meter stick. Even worse, she points out that if the millimeters were subdivided even more into 1/10,000th of a meter, then you could have a more accurate answer still. Being a skeptic, she also brought up doubts about each demarcation being perfect. Perhaps one of the millimeter markers is really 1.1 mm and another is 0.96 mm. In frustration with the problem, you decide to take the midpoint of your two estimates and declare the room to be 5.350 m give or take a few millimeters. You have embraced the uncertainty and given your answer with some give or take a bit. The give or take a bit is a quantification of the uncertainty, or error, of a measurement. This is just a ballpark guess at the uncertainty. It would be great if you could truly quantify it. 1.4 Repeated Trials and Averages to Improve Measurements and Quantify Uncertainty Your friend is not satisfied and wants to push the measurement further and to higher levels of accuracy. She is certain she can determine the true length of the room. She calls up her science major friends to come help. One of them has a tool with marks down to the 1/10,000th (0.0001) meter. As a team, you all set to the task. You measure the following lengths in meters: 1st Try 2nd Try 3rd Try 4th Try 5th Try 6th Try 7th Try 8th Try 5.3488 5.3507 5.3512 5.3391 5.3510 5.3517 5.3602 5.3506 One of the people just learned about the law of large numbers in his mathematics class that states that average of repeated trials (done with the same level of precision using the same method) will tend toward the true value as the number of trials increases. That is, the more trials (i.e., number of measurements made) the closer the averaged value will be to the “real” value. Averaging the eight trials together, and using a bit of statistics, you all find the average 5.3504 m with a standard deviation (standard error) of 0.0057 m. You decide this is good enough for your purposes and report the length of your room to be 5.3504 ± 0.0057 m, where ± means plus and minus and is the standard way of listing scientific uncertainty. The above example is meant to point out that measuring something can be quite difficult and comes with many considerations. You could improve on the measurement through a better method (your friend with the string), and/or a more precise measuring instrument (a more finely divided measuring instrument). You can also make use of the law of large numbers and average many trials together to get a better answer. Regardless of all these efforts though, you could always add another decimal place to your answer, do more trials, and list smaller and smaller uncertainty. You could accurately measure the room to be 5.3505028 ± 1.0x10-7 m, but someone else could come along with a better measuring instrument and push that to 5. 5.35050284 ± 1.0x10-8 m, and so on and so forth. In theory could keep the one more decimal point game up until you get to the Planck length, which is theoretically the smallest physical distance measurable, but you still would have some problems. That number has its own uncertainty in it. It is calculated from three other fundamental physical constants of the universe: the speed of light in a vacuum, the Planck constant, and the gravitational constant. Two of these constants (the Planck constant and the gravitational constant) are empirically measured, themselves, and therefore have their own uncertainty in them! A major goal of astrophysics is to know such quantities as Planck’s constant and the gravitational constant to higher and higher accuracy, but such determinations depend on making other, almost always, indirect, measurements, which have their own limitations and uncertainty. 2. Collecting the Data I hope you have gained an appreciation for how difficult and important measurement is to science. It requires incredible care and often relies on clever indirect measurement techniques. Today in lab, you will be engaging with some of those concepts, but in a very simplified, and hopefully fun, way that demonstrates core concepts of measurement methodology, accuracy, precision, and the how uncertainty is quantified. Measurements will be made, and the data for the whole class will be collected and examined to reinforce the concepts learned in this lab. For this lab, you will work in student pairs. Lab Activity 1 – Playing the Game and Taking Measurements For our exploration this labs core concepts, you will collect x- and y-position data of bean bags you will toss onto a board. You will also measure each bag’s angular location on the board with respect to a reference line. To increase interest, and hopefully have fun while doing so, your tossed bean bag locations on the board will be made into a game. The goal of the game is to get the highest score possible while also carefully measuring the (x, y)- positions. The general layout of the board and the rules of the game are shown in Figure 3. Figure 3. The rules of the game. Two competitions will be taking place while you play. You will directly compete against your lab partner, and as a team, you and your lab partner will be competing with all other lab pairs. The goal is to get the highest total score. While one person throws the bean bags, their lab partner will record the x- and y-positions. The right-hand side of the board is positive x and radius. The top of the board is positive y. Lab Activity 2 – Examining the Distributions of Measurements With all the data collected, it is time to calculate the radial distance away from the center of the board. Use Equation 3 (𝑟 = @𝑥B + 𝑦B) to determine r. Once you have calculated all the radial distances, access the Google Form “Measurement Lab Data Entry Form” and enter your measurements. The form will prompt you to input the x-position, the y-position, the radial distance, and the score earned. Once finished with an entry, choose the prompt to do another entry. You will do this repeatedly until all the data on all 10 tosses is entered. Once all the data is entered, your instructor will graphically display the results on the projector. The distributions for the x, y, angle, radius, and score will be displayed as histograms, which shows how frequently each measurement was made within a certain range. Using this graphical information, you will need to answer questions relating to what the distributions look like, the accuracy, and the precision of each distribution. You will also calculate the averages of your bag trials and compare that to the overall class distribution. Measurement & Data Lab Student Worksheet Name: Lab Instructor: Lab Meeting Time: Measurement Data Records Input your measured data in the table below. Put the name of each player under Player 1 and Player 2. Below the all the data entry, calculate your average x, y, radius, and score for all ten of your team’s measurements. If you need to save calculating the averages for the take-home portion of this lab, then that is okay. You can skip them for now if you do not have time. Measurement Data Table / Game Scorecard Toss # x y Angle Radius Score Player 1: 1 2 3 4 5 Total Score Player 2: 1 2 3 4 5 Total Score Player 1 + Player 2 Scores = Team Score Team Averages x average y average Average angle Average radius Average score Indirect Measurement Results Record the measurements you made to determine the official game play distance and height for the bag tossing platform. Use these measurements to calculate the distance and height. 1. Describe the method the class used to determine the distance to and height of the light pole or fire detector (if lab took place inside). a. [Word description of method] Describe the complete method the class used including all angles and distances measured. b. [Graphical description of method] Draw one or more diagrams demonstrating the method you used. Label all distances and angles measured or determined. 2. What is the distance to and height of the object? Height in meters: ________________ Distance in meters: ________________ 3. Describe at least 3 sources of error that would have propagated into the final answer. That is where could better measurements have been made? 7. For the x-position, y-position, and radius, we should expect the average distance to be zero (a bull’s eye). a. Are the class’s distributions of these measurements near the expected, “true” value? b. If an average is not close to the expected value of 0 then the data is telling us something about an error inherent to the measurement. This is what we call a systematic error, which an error from the method used to obtain the data. Such an error can negatively affect our ability to appropriately test hypotheses in science. Do you observe any systematic errors in the data? If yes, what do you think may have caused them? 8. The width of the distributions characterizes the precision of the measurement. A narrow distribution would mean that every toss was close to the average with little variance. This would be high precision. A wide distribution means that the position measurements varied widely from toss to toss. This would be low precision. a. Do the x-position and y-position distributions have the same precision? If not, which is more precise? b. c. What could have caused any differences in the x-position and y-distribution precisions? Note: the width of a distribution, or how much each toss varies from other tosses is capture by the quantity called the standard deviation, which is one of the displayed values. The standard deviation gives the uncertainty. 9. The radial distance distribution combines information from the x-positions and the y- positions. Therefore, characteristics of the x-position distribution and y-position distribution should be present in the radial distance distribution. a. Describe the shape of the radial distribution with respect to the shape of the x- and y-position distributions. b. Are there any systematic offsets present in the radial distances? If so, provide your hypothesis on what caused the offset. c. Describe how you could use bean bags and the target board to test your hypothesis. What experiment would you run to either confirm there is no systematic offset, or to confirm your explanation of what caused the offset. 10. Any measurement has some inherent error in it. You certainly had some when measuring the x- and y-positions of the bean bags. List and describe at least TWO sources of error you encountered when measuring a bag’s x- and y-positions. That is, what things that you can think of made it difficult to measure the bags exact (x, y)- location? Lab Follow-up Questions [Complete the following questions at home if you do not have time to during class. 11. In your own words, write definitions for accuracy, precision, and uncertainty. 12. Use the triangulation method to determine the distance d using the image to the right. The baseline distance is 5 meters, and the measured angle is 70°. 13. Describe the difference between a direct and an indirect measurement.