Applied Statistics - Study Notes
Lesson summary
Applied Statistics uses statistical theory in business, industry and economics. The unit has three parts: time series analysis, index numbers and statistical quality control (SQC). Walter A. Shewhart is known as the father of statistical quality control.
9.1 Time series
- A time series consists of data arranged chronologically (Croxton and Cowden). When quantitative data are arranged in the order of their occurrence, the series is a time series (Wessel and Wallet).
- Uses: analysis of past behaviour, forecasting and planning, evaluation of current achievements, comparison between time periods.
Components
| Component | Meaning | Example |
|---|---|---|
| Secular trend | General tendency to increase, decrease or stagnate over a long period | Population, production; falling deaths, gadget prices |
| Seasonal variations | Repeat in less than one year; due to nature, customs, traditions | Umbrellas in rainy season, crackers at Deepawali |
| Cyclic variations | Not uniformly periodic; one cycle usually 7 to 9 years | Business cycle: start, boom, depression, recovery |
| Irregular variations | Accidental, random, unpredictable, normally short-term | Floods, wars, earthquakes, strikes, lockouts |
Models
- Additive model: \(Y = T + S + C + I\); components operate independently.
- Multiplicative model: \(Y = T \times S \times C \times I\); components can affect one another.
Measuring trend
- Freehand (graphic) method: simplest and most flexible, but subjective; depends on personal bias.
- Semi-averages: divide data into two equal parts (omit the middle year if odd), average each part, plot at the mid-points and join. Predicted values are not precise.
- Moving averages: three-yearly \(\frac{a+b+c}{3}, \frac{b+c+d}{3}, \dots\); four-yearly \(\frac{a+b+c+d}{4}, \dots\). For an even period the averages are centred by averaging successive pairs. Gives a trend with a fair degree of accuracy.
- Least squares: \(\sum (Y - Y_t) = 0\) and \(\sum (Y - Y_t)^2\) is least. Best method; predicted values more reliable.
Least squares procedure
- Trend line \(Y = a + bX\); normal equations \(\sum Y = na + b\sum X\) and \(\sum XY = a\sum X + b\sum X^2\).
- Odd number of years: \(X = x - \text{middle year}\). Even number: \(X = \frac{x - \text{mean of two middle years}}{0.5}\).
- With \(\sum X = 0\): \(a = \frac{\sum Y}{n}\) (mean of Y) and \(b = \frac{\sum XY}{\sum X^2}\) (slope; positive or negative).
Seasonal index by simple averages
- Arrange data by months or quarters and total each season.
- Find each seasonal average and the grand average (average of averages).
- \(S.I. = \frac{\text{Seasonal average}}{\text{Grand average}} \times 100\).
9.2 Index numbers
- An index number shows by its variations the changes in a magnitude not capable of direct measurement (Wheldon); it measures fluctuations using a base period for comparison.
- Classification: price index, quantity index, cost of living index. Construction: unweighted and weighted.
- Uses: framing decisions and policies, studying trends, measuring inflation and deflation.
Weighted price indices
| Index | Formula | Weight |
|---|---|---|
| Weighted | \(\frac{\sum p_1 w}{\sum p_0 w} \times 100\) | w |
| Laspeyre's | \(\frac{\sum p_1q_0}{\sum p_0q_0} \times 100\) | Base year quantity |
| Paasche's | \(\frac{\sum p_1q_1}{\sum p_0q_1} \times 100\) | Current year quantity |
| Fisher's | \(\sqrt{\frac{\sum p_1q_0}{\sum p_0q_0} \times \frac{\sum p_1q_1}{\sum p_0q_1}} \times 100\) | Geometric mean of L and P |
Tests of adequacy
- Time reversal test: \(P_{01} \times P_{10} = 1\) (ignore the factor 100).
- Factor reversal test: \(P_{01} \times Q_{01} = \frac{\sum p_1q_1}{\sum p_0q_0}\), the true value ratio.
- Fisher's index satisfies both, so it is the ideal index number.
Cost of living index
- Measures the change in income needed to keep the same standard of living; also called the consumer price index.
- Uses: shows whether real wages rise or fall; used to fix dearness allowance and bonus.
- Aggregate expenditure method (most common): \(\frac{\sum p_1q_0}{\sum p_0q_0} \times 100\), same as Laspeyre's. Used when price and quantity are given.
- Family budget method: \(\frac{\sum PV}{\sum V}\) with \(P = \frac{p_1}{p_0} \times 100\) and \(V = p_0q_0\). Used when price and weight are given.
9.3 Statistical quality control
- Quality depends on Material, Manpower, Machines and Management (4 M's).
- Chance (random) causes: small, natural, inherent; beyond human control (rain, floods, power cuts).
- Assignable causes: defective raw materials, faulty machines, unskilled manpower, worn tools, new operations. SQC aims to eliminate these.
- Process control: minimise the proportion of defectives using control charts. Product control: inspect the product through sampling inspection plans.
- A control chart has CL, UCL and LCL. Points inside the limits mean the process is in control; any point outside means out of control.
- Charts for variables: mean (\(\overline{X}\)) chart and range (R) chart. \(R = x_{\max} - x_{\min}\).
| Chart | SD given | SD not given |
|---|---|---|
| \(\overline{X}\) chart | \(\overline{\overline{X}} \pm 3\frac{\sigma}{\sqrt{n}}\), CL \(= \overline{\overline{X}}\) | \(\overline{\overline{X}} \pm A_2\overline{R}\), CL \(= \overline{\overline{X}}\) |
| R chart | \(\overline{R} \pm 3\sigma_R\), CL \(= \overline{R}\) | UCL \(= D_4\overline{R}\), CL \(= \overline{R}\), LCL \(= D_3\overline{R}\) |
For n = 5: \(A_2 = 0.577\), \(D_3 = 0\), \(D_4 = 2.114\).
More for this chapter
Book Back Questions25 textbook MCQs · solved
Additional MCQs10 extra MCQs · solved
Practice TestInteractive · instant score
Book Back TestTest yourself on the textbook set
Additional MCQ TestTest yourself on the extra set
Formula SheetAll key formulas