Chemical Thinking Lab Assignment 2 Guide v1.

0 | 1

How do we assess data quality?

Overview of LA2
Lab Assignment 2 concerns assessing the quality of experimental data in terms of precision and accuracy. To this
end, we take up the graphical determination of density and assess the precision and accuracy of several results.

In research, industrial and clinical laboratories, spreadsheet programs such as EXCEL are prevalent. CHEM 151
makes extensive use of EXCEL for data workup, especially for graphical presentations, analysis and inferring
physical/chemical values. Thus, another goal for Lab Assignment 2 is to develop basic EXCEL graphing skills. This
will include the proper labeling and titling of the graphs generated. You will likely find such proficiency applicable
to data processing and presentation in other courses and future careers – so this is a valuable skill set to develop.

All measurements require proper handling of significant figures and units. So, in Lab Assignment 2 you will further
experience managing significant figures and controlling units. Skills learned here will be important throughout the
semester, in other classes, and in many professions.

Density Revisited
As you know from Lab Assignment 1, density (d) is a measure of
the amount of mass (m) per unit volume (V) of a given substance
Graph LA2-1
(d = m/V). As indicated previously, this suggests that density can Density Determination:
be determined by measuring the mass and volume of the Mass vs. Volume
substance of interest and dividing.

Now, if you have a set of volume and corresponding mass m = 1.9291 + 0.4089
measurements for a given substance, then density can be
determined graphically by plotting in a spreadsheet program
mass (m) versus volume (V) and using the spreadsheet’s linear
fit tool to calculate and display a linear trend-line as illustrated
on the right (Graph LA2-1). The slope of this trend-line is m/V or
density. This is actually a superior way to determine density, R2 = 0.9853
since a single measurement error, and certain types of random
errors are suppressed. Put another way, the graphical approach
has the effect of weighting and “averaging out” the individual
density values, giving a result that you can have more
confidence as being the “true” or actual density.

Experimental Errors
Experimental errors fall into two broad categories: systematic and random. Systematic errors generally have
assignable (specifiable) causes and can be quantified. In most cases systematic errors cause the measurements to
be either too large or too small. So, they have a direction. Such errors include instrumental errors, method or
design errors (build-in procedural faults) and most human errors (mistakes, personal bias, etc.). Random errors on
the other hand have causes unknown or multiple causes such that the error source is unspecifiable; the
measurements fluctuate in a random (non-directional) pattern. Errors can also propagate. Some are additive,
others tend to cancel out.

In general, random errors primarily affect the precision, while systematic errors influence the accuracy.

CHEM 151 Lab Assignment 2 v1.1 MY © 2020 8-4-20 Chem Think Technical Press CBC UofA
Chemical Thinking Lab Assignment 2 Guide v1.0 | 2

Precision
Precision is the reproducibility of a measurement – how close together repeated measurements of one type on
the same thing are to each other. Remember, random errors cause measurements to fluctuate in a chance (non-
directional) pattern, so random errors can certainly affect the reproducibility of a measurement, and hence the
precision. When graphing linearly related variables, such as mass and volume in Graph LA2-1, the more scatter
from the linear trend-line, the greater the imprecision of the method and/or the individual carrying out the
method. To understand this, recall that random errors cause measurements to fluctuate in a non-directional
pattern. This behavior manifests itself as chance scatter of the individual values from the linear trend-line.

While the statistical measure of precision is considered the standard deviation (or relative standard deviation), for
our purposes we will use the R2 value (coefficient of determination) generated in EXCEL, where R2 > 0.98 suggests
good precision, while R2 < 0.98 implies imprecision, with increasing imprecision as the R2 value declines. A “perfect
fit” to the linear trend-line would have R2 = 1.00, meaning all the points lie on the linear trend-line. Looking at
Graph LA2-1 we see a R2 of 0.9853 reported. From the above discussion, this R2 value suggests good precision in
the measurements, and in fact the individual density values show only modest scatter from the linear trend-line.

Accuracy
Accuracy is how close a given value is to the “true” value. Since systematic errors cause measurements to be either
too large or too small, you can see the direct connection to accuracy (or inaccuracy). We can evaluate accuracy by
calculating the percent relative error. Percent relative error is defined as:

𝐸𝑥𝑝𝑒𝑟𝑖𝑚𝑒𝑛𝑡𝑎𝑙 𝑉𝑎𝑙𝑢𝑒 − 𝑇ℎ𝑒𝑜𝑟𝑒𝑡𝑖𝑐𝑎𝑙 𝑉𝑎𝑙𝑢𝑒

% 𝑅𝑒𝑙𝑎𝑡𝑖𝑣𝑒 𝐸𝑟𝑟𝑜𝑟 = 100%
𝑇ℎ𝑒𝑜𝑟𝑒𝑡𝑖𝑐𝑎𝑙 𝑉𝑎𝑙𝑢𝑒
or,
𝑀𝑒𝑎𝑠𝑢𝑟𝑒𝑑 𝑉𝑎𝑙𝑢𝑒 − 𝐴𝑐𝑐𝑒𝑝𝑡𝑒𝑑 𝑉𝑎𝑙𝑢𝑒
% 𝑅𝑒𝑙𝑎𝑡𝑖𝑣𝑒 𝐸𝑟𝑟𝑜𝑟 = 100%
𝐴𝑐𝑐𝑒𝑝𝑡𝑒𝑑 𝑉𝑎𝑙𝑢𝑒

The theoretical or accepted value is considered the actual or “true” value, while the experimental or measured
value is what you determined or measured. Percent relative error can serve as a metric for accuracy: for our
purposes a percent relative error < 2% indicates the value determined/measured is nominally the same as the
accepted value, and the smaller the percentage, the better the accuracy, while a percent relative error > 2%
suggests inaccuracy, the greater the percentage, the greater the inaccuracy. Further, the percent relative error
entails taking a difference between the experimental and theoretical values, hence it is signed. A positive %
relative error indicates the experimental value is higher than the theoretical, and vice versa for a negative %
relative error. Thus, a direction is specified in the error – a feature of systematic errors.

Graph LA2-1 indicates an experimental density of 1.9291 g/mL. Let’s assume the theoretical or accepted density
for the substance being measured in Graph LA2-1 is 1.9911 g/mL. Then, the % relative error would be -3.11%,
which is clearly > 2%, suggesting inaccuracy, and with the negative sign indicating the experimental value is lower
than the theoretical or accepted value.

𝐸𝑥𝑝𝑒𝑟𝑖𝑚𝑒𝑛𝑡𝑎𝑙 𝑉𝑎𝑙𝑢𝑒 − 𝑇ℎ𝑒𝑜𝑟𝑒𝑡𝑖𝑐𝑎𝑙 𝑉𝑎𝑙𝑢𝑒

% 𝑅𝑒𝑙𝑎𝑡𝑖𝑣𝑒 𝐸𝑟𝑟𝑜𝑟 = 100%
𝑇ℎ𝑒𝑜𝑟𝑒𝑡𝑖𝑐𝑎𝑙 𝑉𝑎𝑙𝑢𝑒

1.9291 𝑔/𝑚𝐿 − 1.9911 𝑔/𝑚𝐿

% 𝑅𝑒𝑙𝑎𝑡𝑖𝑣𝑒 𝐸𝑟𝑟𝑜𝑟 = 100% = −3.11%
1.9911 𝑔/𝑚𝐿

CHEM 151 Lab Assignment 2 v1.1 MY © 2020 8-4-20 Chem Think Technical Press CBC UofA
Chemical Thinking Lab Assignment 2 Guide v1.0 | 3

Please do not confuse precision and accuracy. In science and engineering they are very different concepts. Figure
LA2-1 presents an example that should help to clarify the difference:

Figure LA2-1. Precision vs. Accuracy: Target Shooting Example

In this example you can think of the “center” as being the theoretical or accepted value, and the “average” of the
six points as being analogous to the linear trend-line in the above discussion.

LA2 Resources
On D2L in the Content section under “Lab Assignment 2” you will find two useful resources:

• Significant Figures. This document offers a brief introduction to handling significant figures. Should
you desire additional information regarding significant figures, there are substantial resources on the
internet.

• Graphing in Excel. If you are unfamiliar with graphing in EXCEL, this guide should help.

Additionally there is a YouTube Excel Graphing Technical Guide video on the “Chemical Thinking” channel
available via this link: https://www.youtube.com/watch?v=d44eyHNxfFs

CHEM 151 Lab Assignment 2 v1.1 MY © 2020 8-4-20 Chem Think Technical Press CBC UofA