Introduction To Statistical Analysis: Measures of Central Tendency

CHEM 26.
1 [INTRODUCTION TO STATISTICAL ANALYSIS]
In order to do statistical analyses/tests, we need to gather samples. Population is the collective term for
all the samples. For example if we need to analyze coconut oil, then we have to gather all of these. As in
all (lahat ng nasa earth). Population mean is  and population standard deviation is σ. However, getting
all of the samples is impossible. Hence we do sampling wherein we get a portion to represent the
population. With this, we save time, effort and money. Sample mean is X and sample standard deviation
is s.
Note: all formulas are found in the manual.
MEASURES OF CENTRAL TENDENCY

As the name implies, we are to determine the center or the middlemost value.
1. Mean- average value. (sum all the measurements then divide by the number of measurements)
2. Median-middlemost value. Arrange the measurements in increasing or decreasing order then
find the middle value.
*elem and HS stat includes MODE which is the data with the highest frequency. However, the need not
discuss it here since the reliability of the mode is a question.
MEASURES OF PRECISION
Precision refers to the closeness of results. Eight measures of precision are presented in the manual with
formulas. Familiarize yourselves with these. Only the pooled standard deviation requires a lot of effort.
Pooled standard deviation id done to determine precision with respect to all data sets. Consider the
experiments per group. The methodology and sample is the same with all the groups hence we can
calculate for the precision of the class. The pooled standard deviation will be presented in the formal
report only wherein you need to gather your classmates’ data.
MEASURES OF ACCURACY
Accuracy, on the other hand refers to the closeness of the experimental with the true value or the
theoretical value. Error refers to the numerical difference between the actual and true value. Both
absolute and relative errors were encountered in chem16.
1. Absolute error-difference between the experimental (actual) and the theoretical value.
(absolute value to ha, errors reported as positive nalang)
2. Relative error-absolute error pero with respect to the theo value kaya divided by theo un.
Absolute value din un.
*negative error-experimental value is less than the theoretical value
*positive error-experimental value is greater than the theoretical value
CONFIDENCE LIMITS
The limits provide a range of values from the mean wherein the TRUE MEAN possibly lies. (take note, we
only determine the sample mean). The range is ts/√n. t values are provided by the manual pero careful
kasi values are in n-1. So for example t value at n=10, 95% is 2.26.
Q-TEST
This test is done for rejecting outliers. The extreme values are usually the outliers so we apply Q-test on
these values one at a time.
1. Arrange the measurements in increasing or decreasing order.
2. Get the range (highest – lowest value)
3. Get the difference between the suspect value and the value next to it.
4. Divide the value from 3 with 2. This is now Q exp
CHEM 26.1 [INTRODUCTION TO STATISTICAL ANALYSIS]
5. Compare the value with that of the table. If Q exp > Qtab : data is rejected. If Qexp < Qtab : data is
accepted.
Q-test is done first before solving for other statistical measures. We need to remove outliers muna
before getting the mean, etc. do not stop doing Qtest until both ends are accepted na.
EXAMPLE:
1. A student obtained the following values for the normality of a solution: 0.1029, 0.1060, 0.1036,
0.1032, 0.1018 and 0.1034. What value should be reported for the normality at 90% confidence
level?
Q-test.
Arrange in order: 0.1018 , 0.1029, 0.1032, 0.1034, 0.1036, 0.1060
Suspected outliers: 0.1018 and 0.1060
Qtest for 0.1018

Spread: |0.1018-0.1029|
Range:0.0042
Spread/range=0.26 =Qexp Qtab=0.560 Qexp<Qtab accept 0.1018
Qtest for 0.1060

Spread: |0.1036-0.1060|
Range:0.0042
Spread/range=0.57 Qexp>Qtab  reject 0.1060
Repeat Q test. This time outliers are 0.1018 and 0.1036.
Qtest for 0.1018

Spread: |0.1018-0.1029|
Range:0.0018
Spread/range=0.61 =Qexp Qtab=0.642 Qexp<Qtab accept 0.1018
Qtest for 0.1036

Spread: |0.1034-0.1036|
Range:0.0018
Spread/range=0.11 Qexp>Qtab  accept 0.1036
Since both extremes are accepted, we can now solve for the mean.
Mean=(0.1018 + 0.1029 + 0.1032 + 0.1034 + 0.1036)/5
ANSWER: 0.1030 N
*Of course 0.1060 will not be included here since it is considered outlier.

Introduction To Statistical Analysis: Measures of Central Tendency

Uploaded by

Introduction To Statistical Analysis: Measures of Central Tendency

Uploaded by

CHEM 26.

1 [INTRODUCTION TO STATISTICAL ANALYSIS]

Note: all formulas are found in the manual.

MEASURES OF CENTRAL TENDENCY

Qtest for 0.1018

Qtest for 0.1060

Repeat Q test. This time outliers are 0.1018 and 0.1036.

Qtest for 0.1018

Qtest for 0.1036

You might also like