11.1 UNCERTANTIES AND ERRORS IN MEASUREMENT AND RESULTS FOCUS CONCEPT All measurement has a limit of precision and accuracy, and this must be taken into account when evaluating experimental results. INTRO Error analysis is the study and evaluation of uncertainty in measurement. In science, the word “error” does not mean “mistake” or “blunder”. “Error” in a scientific measurement means the inevitable uncertainty that attends all measurements. It is the job of the scientist to keep these errors to a minimum. Assessing the magnitude of these errors is necessary in order to determine the reliability of a final result. “Experimental error" (aka: experimental uncertainty) refers to variability in results due to limitations in the experimental design and measuring DEVICE; it's the reason scientists perform multiple trials of any given experiment, and report their results as a statistical average with a plus or minus deviation included. In this course, although we will often average measurements, we'll rarely do more advanced statistical analysis. In the “Conclusion and Evaluation section of your formal lab reports, do list any observed reasons that you feel may have contributed to errors in your experiment, including: specific problems with the equipment, difficulties in reading the equipment; or, limitations in the design of the experiment. Don't just guess about why your experiment might have gone wrong - only mention specific sources of error that you have a legitimate reason to believe affected your results, and explain how those sources of error affected your results, e.g. "The average final velocity of the pendulum was 15.6% smaller than it was at the top, which we presume is at least partly caused by loss of energy due to air friction as the pendulum swung down." *SIGNIFICANT DIGITS IN MEASUREMENTS (http://youtu.be/F9R8johYCww) COUNTED QUANTITIES MEASURED QUANTITIES - _____________exact - ______ uncertainty - _____________ number of significant digits e.g 22 people, 3 eggs, 12 pencils - _______________exact - last digit in a measurement is always ____________ (an estimate) e.g. Consider a mass reading on three different scales. Bathroom scale 56 1 kg (___ sd) Physician’s office 55.8 0.1 kg (___ sd) Science lab 55.778 0.001 kg (___ sd) *In any measurement, the digits that are reliably known are called __________________ digits. These include the digits known for certain and the single last digit that is _____________________. (↑ decimal places = ↑ precision) Rules for Determining the Number of Significant Digits Associated with a Measurement RULE MEASUREMENT (EXAMPLE) 1. All NON-ZERO numbers ARE significant. 1,292 mm 2. “CAPTIVE” ZEROS (i.e. zeroes placed between other digits) ARE significant. 400,006 cm 3. LEADING ZEROS are NEVER significant. e.g. 0.0000089 kg = 8.9 x 10-6 kg 4. TRAILING ZEROS: a) ARE significant if they indicate a measurement (i.e. the zeros are to the right of a decimal). b) Zeroes at the end of a number are significant only if they are indicated to be so using scientific notation. (e.g. 1000 seconds has an unspecified # of sig digs) - 5.800 km - 900.0 mL, - 703.0 N - 1 x 103 s - 1.0 x 103 s - 0.45 mol dm-3 - 4.5 x 10-1 mol dm-3 - 1.00 x 103 s - 4.50 x 10-1 mol dm-3 -1.000 x 103 s - 4.500 x 10-1 mol dm-3 SIG DIGS PRACTICE ACTIVTY 1: *Determine the number of significant digits in the following numbers: NUMBER SIG. DIGS NUMBER 0.02 0.0005 0.020 0.1020 501 10 001 501.0 3.50 x 104 3 5.000 x 10 2 x 109 5 000.0 8.4572 x1023 6 052.00 8040 10 800 000 3.01 x 1021 SIG. DIGS NUMBER 80405 0.0300 699.5 2.000 x102 0.90100 90 100 4.7 x 10-8 0.000 410 SIG. DIGS. PRACTICE ACTIVITY 2: Round the following to the indicated number of significant digits. *When performing calculations that involve a series of mathematical steps, don’t round off numbers until you are ready to state the final answer…use all the decimal places reported by your calculator re: intermediate calculations. a) 2150 (2 s.d.) -___________ b) 0.0256 (2 s.d.) - _________ c) 0.0346 (2 s.d.) - __________ d) 0.0050129 (3 s.d.) - _________ e) 1.994 (3 s.d.) - _________ f) 2149.99 (3 s.d.) - ________ *SIGNIFICANT FIGURES IN CALCULATIONS ADDITION AND SUBSTRACTION MULIPLICATION AND DIVISION RULE: Round the answer to the LEAST number of places in the _________________ portion of any number in the problem. RULE: The LEAST number of __________________ figures in any number of the problem determines the number of significant figures in the answer. e.g. Add the following measurements: 123.0 cm + 12.40 cm + 5.380 cm e.g. A cyclist travels 4.00 x 103 m on a racetrack in 292.4 s. Calculate the average speed of the cyclist. Solution: Solution: Average speed (v) is distance (d) divided by time (t): 123.0 cm 12.40 cm 5.380 cm cm v= *Thus, ______________ cm should be rounded to _____________ cm since 123.0 cm has only one decimal place. 4.00 x 103 m d = = t 292.4 s *Thus, the average speed should be reported to _____ significant digits – ____________ m/s. PRACTICE: Perform the following operations expressing the answer in the correct number of significant digits. a) 1.35 2.267 = b) 1 035 42 = c) 12.01 + 35.2 + 6 = d) 55.46 – 28.9 = e) 0.021 3.2 100.1 = f) 1.278 103 1.4267 102 = g) 0.2129 + 0.002 + 0.03 = h) 101.4 + 25 + 201 = i) 1.0 + 2.04 + 5.03 = j) 2.5 x 1.1111 = k) 8.314 x 2.5x10-2 = l) 35.45 x 2.25 = m) 150 4 = n) 505 – 450.25 = o) 1.252 0.115 0.012 = p) 0.15 + 1.15 + 2.051 = q) 41.11 + 20.5 + 18.333 = *EXPERIMENTAL ERRORS Random vs. Systematic Error (http://youtu.be/x5Euj2d39kI) RANDOM ERRORS SYSTEMATIC ERRORS cause a series of measurements to fluctuate above ________ below a ______________ _______ – sided, predictable, unavoidable, quantifiable error. Repeated measurements _________ reveal this type of uncertainty. cause a series of measurements to be EITHER too high _______too low (i.e. either all higher than an accepted value OR all lower than an accepted value) ONE – sided, quantifiable error. Repeated measurements WILL ________ reveal this type of uncertainty regardless of the number of trials performed. Usually difficult to detect unless the expected/actual value is known. Can sometimes be remedied/ameliorated/lessened by careful experimental design. *E.g. A sample of metal is weighed 5 times on the same scale. TRIAL 1 2 3 4 5 MASS / g 2.86 2.87 2.85 2.84 2.88 mean = __________ mmax = __________ mmin = __________ absolute uncertainty = ________ mass = _________________ absolute uncertainty = range = random error, quantified *The five different masses are due to random errors. What causes random errors? *Cannot always be identified – possible sources: - ________ currents - ____________ fluctuations in balance - human _______________ variance (for analogue devices) – e.g. reading a burette - human ______________ time - judging a colour change *E.g. Let’s say the actual mass of the metal in the adjacent example is 2.92 g. The 5 measurements, therefore, are ALL _____________ than the actual value – the measurements are _________________ due to _______________ error. *What causes systematic errors? Usually/Examples: (i) faulty/broken or ____________________ equipment: - ________________ running two seconds too fast - _____________________reading 2° higher than actual - measuring length using stretched _______________ - ______ meter not calibrated (ii) improper use of measuring instrument / poor experimental technique - forget to “_______” electronic balance - reading a ______________ incorrectly / parallax error - reading scale incorrectly - unaccounted __________ loss in a calorimetry experiment - adding too much titrant in titration experiment - using an inappropriate indicator in a titration experiment Precision vs. Accuracy PRECISION how __________ a series of measurements are to each other *Which measurement is more precise?....has less random error? a) 2.85 0.01 g or 2.85 0.02 g b) 2.850 0.001 g or 2.85 0.01 g - (↑ decimal places = ↑ precision = ↑$ ) usually quoted as a value = range = _____________ error ↓ random error = ↑ ______________ ↑ # of ___________ = ↓ random error = ↓ standard deviation ACCURACY how close a result or measurement is to the ____________, accepted value Accuracy / systematic error quantified using: PRACTICE: Consider the following ideal gas constant measurements: (Accepted = 8.314 J mol-1 K-1) 0.03 J mol K Group 2 8.513 0.006 J mol K Group 1 8.34 -1 -1 -1 -1 Which of the measurements is: a) more accurate? b) more precise? c) Calculate the relative(percent) error associated with each measurement. SUMMARY DEFINITION AFFECTED BY EVALUATE BY LOOKING AT HOW IMPROVE PRECISION - how close a series of measurements are to each other - ______________errors (caused by air currents, voltage fluctuations etc.) - use more precise measuring instrument (↑ $) - perform multiple trials ACCURACY - how close a result or measurement is to the actual, published value - ______________ errors (caused by faulty equipment / improper use of equipment) - uncertainty range (i.e. absolute error / magnitude of value) - magnitude of ____ error - calibrate / use equipment properly - compare measurement readings with more reliable instrument *UNCERTAINTY / ERRORS IN MEASUREMENT (SUMMARY) “Uncertainty in measurements (http://youtu.be/2JRwdeaX_w0) “Absolute vs % Uncertainty (http://youtu.be/kfsoxLuC570) ABSOLUTE UNCERTAINTY - the actual uncertainty in the value / measurement e.g. 13.10 0.01 g A.U = _______ PERCENTAGE (RELATIVE) UNCERTAINTY absol . uncert . % uncert . x 100 measurement e.g. 13.10 PERCENTAGE ERROR 0.01 g % uncert. = Measurement 13.10 g ________ *Identify two ways by which the % uncertainty can be reduced. 1. ↑ ____________ size; 2. ↑ _____________ of measuring device Analogue vs. Digital Uncertainty ANALOGUE DIGITAL - If not available, use ½ of _________ count (i.e. smallest subdivision / gradation on measuring instrument) - If not available, use the value associated with the smallest ____________ value. *EVALUATING PRECISION USING % UNCERTAINTY *Consider the following: SCALE MEASUREMENT Electronic Balance 0.25 Triple Beam Balance 100.00 0.005 g 0.05 g % UNCERTAINTY (Relative uncert.) Which scale is more precise? Explain. ____% (0.25 g ____% ) ______ % (100.00 g ____ %) *REPORTING UNCERTANTIES (in “Conclusion” of formal lab) RULE 1 report uncertainties to ______ SIG. FIG. (almost always) RULE 2 experimental result and uncertainty must have SAME _______________ SIG. FIG. EXAMPLES INCORRECT CORRECT d = 5.67 a = 1261.29 q = 2.34 0.0415 g cm-3 199.43 cm/s2 .000225675 J d = 18.8 g cm-3 9.63 % (large uncert.) (very small uncert.) Which measurement is more precise? Explain. *PROPAGATION OF UNCERTAINTIES IN CALCULATED RESULTS (http://youtu.be/EvQOaPNhdxI) Propagating uncertainty taking uncertainty in measurements and carrying them through to the calculated result Shows the impact of the uncertainties on the final result. ADDITION AND SUBTRACTION e.g. Initial burette reading (Vi) = Final burette reading (Vf) = 3.10 36.45 MULTIPLICATION AND DIVISION e.g. density of tap water – sample data 3 0.10 cm 0.10 cm3 d= Volume added (ΔV = Vf -Vi ) = __________ __________ *Volmax = ________________________ 9.90 0.005 g 10.1 0.1 cm3 R.Umass =________ R.Uvolume = _________ Σ R.U =__________ *Calculation of density with unrouded R.U sum: *Volmin = ________________________ * (Volmax - Volmin) ÷ 2 = ______________ *Next, Convert to absolute uncertainty (round to ONE sig dig) and make sure density value has same rightmost sig dig. RULE: When ADDING or SUBTRACTING measurements, ADD THE _________________ UNCERTAINTIES attached to each measurement. RULE: When MULTIPLYING or DIVIDING measurements, ADD THE ___________________ (RELATIVE) UNCERTAINTIES attached to each measurement. *DISCUSSING ERRORS AND UNCERTANTIES IN FORMAL LAB RESULTS (http://youtu.be/a3L1H0aflyM) *Your discussion of experimental errors, in a formal lab report, should focus on the influence of random and systematic errors on your results. Note that almost all measurements are subject to both systematic and random uncertainties. Refrain from referring to "human error". Examples of so-called human error include misreading a ruler, adding the wrong reagent to a reaction mixture, mistiming the reaction, miscalculations, or any kind of mistake / blunder. Scientists would never report the results of an experiment affected by human error - instead, they repeat the experiment more carefully. PRACTICE: Evaluate the influence of random and systematic error on the following results. LAB Determine molar mass of methane. (Accepted value = 16.04 g mol-1) SAMPLE RESULTS 0.5 g mol 13.2 0.5 g mol-1 15 5 g mol-1 10 4 g mol-1 15.1 RANDOM ERROR SYSTEMATIC ERROR -1 *When the final uncertainty arising from random errors is calculated, this can then be compared with the % error. If the % error is larger than the total uncertainty, then random error alone cannot explain the discrepancy and systematic errors must be involved. FIXING RANDOM ERROR (how get better precision) 1. Identify largest contributor of ______ uncertainty. e.g. d= 153 g 6% 25 cm3 0.3% *Aim would be to decrease the % uncertainty with respect to ____________ since it is a much larger (6% vs. 0.3%) source of random error. FIXING SYSTEMATIC ERROR (how get better accuracy) 1. a) “Seek and destroy” – find and then eliminate source of systematic error. (E.g. Are all the measuring devices working properly? Are the displayed values accurate? Have I used the equipment properly? Have I committed any avoidable procedural errors? - forgot to “zero” the balance when obtaining a mass, for example.) b) Re-________________– Have all devices been calibrated properly? 2. Use ____________ sample size. e.g. 0 .1 g (100) = 10% 1g 0 .1 g (100) = 5% (ii) % uncert = 2 g 2. Evaluate the validity of any ___________________ that were made. (e.g. room temperature/pressure, density/specific heat capacity of water) (i) % uncert = 3. Perform more _____________ – will statistically decrease random error. 4. Change procedure use more ____________ measuring instrument(s), for example *ASSIGNMENT Complete problem set handout. 3. Re-evaluate ____________ Look for places where something you measured/calculated might have been different from the true value. *N.B. Repeating experiment will not eliminate systematic error.