Process data collection and presentation 57 applications are limited to the presentation of proportions since the whole ‘pie’ is normally filled. The use of graphs All graphs, except the pie chart, are composed of a horizontal and a vertical axis. The scale for both of these must be chosen with some care if the resultant picture is not to mislead the reader. Large and rapid variations can be made to look almost like a straight line by the choice of scale. Similarly, relatively small changes can be accentuated. In the pie chart of Figure 3.8 the total elimination of the defect D will make all the others look more important and it may not be immediately obvious that the ‘pie’ will then be smaller. The inappropriate use of pictorial graphs can induce the reader to leap to the wrong conclusion. Whatever the type of graph, it must be used with care so that the presentation has not been chosen to ‘prove a point’ which is not supported by the data. 3.5 Conclusions This chapter has been concerned with the collection of process data and their presentation. In practice, process improvement often can be advanced by the correct presentation of data. In numerous cases, over many years, the author has found that recording performance, and presenting it appropriately, is often the first step towards an increased understanding of process behaviour by the people involved. The public display of the ‘voice of the process’ can result in renewed efforts being made by the operators of the processes. Chapter highlights ᭹ Process improvement requires a systematic approach which includes an appropriate design, resources, materials, people, process and operating instructions. ᭹ Narrow quality and process improvement activities to a series of tasks of a manageable size. ᭹ The basic rules of the systematic approach are: no process without data collection, no data collection without analysis, no analysis without decision, no decision without action (which may include no action). ᭹ Without records analysis is not possible. Ticks and initials cannot be analysed. Record what is observed and not the fact that there was an observation, this makes analysis possible and also improves the reliability of the data recorded. 58 Process data collection and presentation ᭹ The tools of the systematic approach include check sheets/tally charts, histograms, bar charts and graphs. ᭹ There are two types of numeric data: variables which result from measurement, and attributes which result from counting. ᭹ The methods of data collection and the presentation format should be designed to reflect the proposed use of data and the requirements of those charged with its recording. Ease of access is also required. ᭹ Tables of figures are not easily comprehensible but sequencing data reveals the maximum and the minimum values. Tally charts and counts of frequency also reveal the distribution of the data – the central tendency and spread. ᭹ Bar charts and column graphs are in common use and appear in various forms such as vertical and horizontal bars, columns and dots. Grouped frequency distribution or histograms are another type of bar chart of particular value for visualizing large amounts of data. The choice of cell intervals can be aided by the use of Sturgess rule. ᭹ Line graphs or run charts are another way of presenting data as a picture. Graphs include pictorial graphs and pie charts. When reading graphs be aware of the scale chosen, examine them with care, and seek the real meaning – like statistics in general, graphs can be designed to mislead. ᭹ Recording process performance and presenting the results reduce debate and act as a spur to action. ᭹ Collect data, select a good method of presenting the ‘voice of the process’, and then present it. References Ishikawa, K. (1982) Guide to Quality Control, Asian Productivity Association, Tokyo. Oakland, J.S. (2000) Total Quality Management, Text and Cases, 2nd Edn, Butterworth- Heinemann, Oxford. Owen, M. (1993) SPC and Business Improvement, IFS Publications, Bedford. Discussion questions 1 Outline the principles behind a systematic approach to process improve- ment with respect to the initial collection and presentation of data. 2 Operators on an assembly line are having difficulties when mounting electronic components onto a printed circuit board. The difficulties include: undersized holes in the board, absence of holes in the board, oversized Process data collection and presentation 59 wires on components, component wires snapping on bending, components longer than the corresponding hole spacing, wrong components within a batch, and some other less frequent problems. Design a simple tally chart which the operators could be asked to use in order to keep detailed records. How would you make use of such records? How would you engage the interest of the operators in keeping such records? 3 Describe, with examples, the methods which are available for presenting information by means of charts, graphs, diagrams, etc. 4 The table below shows the recorded thicknesses of steel plates nominally .3 cm ± .01 cm. Plot a frequency histogram of the plate thicknesses, and comment on the result. Plate thicknesses (cm) .2968 .2921 .2943 .3000 .2935 .3019 .2991 .2969 .2946 .2965 .2917 .3008 .3036 .3004 .2967 .2955 .2959 .2937 .2961 .3037 .2847 .2907 .2986 .2956 .2875 .2950 .2981 .1971 .3009 .2985 .3005 .3127 .2918 .2900 .3029 .3031 .3047 .2901 .2976 .3016 .2975 .2932 .3065 .3006 .3011 .3027 .2909 .2949 .3089 .2997 .3058 .2911 .2993 .2978 .2972 .2919 .2996 .2995 .3014 .2999 5 To establish a manufacturing specification for tablet weight, a sequence of 200 tablets was taken from the production stream and the weight of each tablet was measured. The frequency distribution is shown below. State and explain the conclusions you would draw from this distribution, assuming the following: (a) the tablets came from one process (b) the tablets came from two processes. 60 Process data collection and presentation Measured weight of tablets Weight (gm) Number of tablets 0.238 2 .239 13 .240 32 .241 29 .242 18 .243 21 .244 20 .245 22 .246 22 .247 13 .248 3 .249 0 .250 1 .251 1 .252 0 .253 1 .254 0 .255 2 200 Part 2 Process Variability 4 Variation and its management Objectives ᭹ To examine the traditional way in which managers look at data. ᭹ To introduce the idea of looking at variation in the data. ᭹ To differentiate between different causes of variation and between accuracy and precision. ᭹ To encourage the evaluation of decision making with regard to process variation. 4.1 The way managers look at data How do managers look at data? Imagine the preparations in a production manager’s office shortly before the monthly directors’ meeting. David, the Production Director, is agitated and not looking forward to the meeting. Figures from the Drying Plant are down again and he is going to have to reprimand John, the Production Manager. David is surprised at the results and John’s poor performance. He thought the complete overhaul of the rotary dryer scrubbers would have lifted the output of 2, 4 D and that all that was needed was a weekly chastizing of the production middle management to keep them on their toes and the figures up. Still, reprimanding people usually improved things, at least for the following week or so. If David was not looking forward to the meeting, John was dreading it! He knew he had several good reasons why the drying figures were down but they had each been used a number of times before at similar meetings. He was looking for something new, something more convincing. He listed the old favourites: plant personnel absenteeism, their lack of training (due to never having time to take them off the job), lack of plant maintenance (due to the demand for output, output, output), indifferent material suppliers (the phenol 64 Variation and its management that came in last week was brown instead of white!), late deliveries from suppliers of everything from plant filters to packaging materials (we had 20 tonnes of loose material in sacks in the Spray Dryer for 4 days last week, awaiting re-packing into the correct unavailable cartons). There were a host of other factors that John knew were outside his control, but it would all sound like whinging. John reflected on past occasions when the figures had been high, above target, and everyone had been pleased. But he had been anxious even in those meetings, in case anyone asked him how he had improved the output figures – he didn’t really know! At the directors’ meeting David asked John to present and explain the figures to the glum faces around the table. John wondered why it always seemed to be the case that the announcement of low production figures and problems always seemed to coincide with high sales figures. Sheila, the Sales Director, had earlier presented the latest results from her group’s efforts. She had proudly listed the actions they had recently taken which had, of course, resulted in the improved sales. Last month a different set of reasons, but recognizable from past usage, had been offered by Sheila in explanation for the poor, below target sales results. Perhaps, John thought, the sales people are like us – they don’t know what is going on either! What John, David and Sheila all knew was that they were all trying to manage their activities in the best interest of the company. So why the anxiety, frustration and conflict? Let us take a look at some of the figures that were being presented that day. The managers present, like many thousands in industry and the service sector throughout the world every day, were looking at data displayed in tables of variances (Table 4.1). What do managers look for in such tables? Large variances from predicted values are the only things that many managers and directors are interested in. ‘Why is that sales figure so low?’ ‘Why is that cost so high?’ ‘What is happening to dryer output?’ ‘What are you doing about it?’ Often thrown into the pot are comparisons of this month’s figures with last month’s or with the same month last year. 4.2 Interpretation of data The method of ‘managing’ a company, or part of it, by examining data monthly, in variance tables is analogous to trying to steer a motor car by staring through the off-side wing mirror at what we have just driven past – or hit! It does not take into account the overall performance of the process and the context of the data. Table 4.1 Sales and production report, Year 6 Month 4 Month 4 actual Monthly target % Difference % Diff month 4 last year Year-to-date Actual Target % Diff YTD as % Diff (Last YTD) Sales Volume 505 530 –4.7 –10.1 (562) 2120 2120 0 +0.7 (2106) On-time (%) 86 95 –9.5 –4.4 (90) 88 95 –7.4 –3.3 (91) Rejected (%) 2.5 1.0 +150 +212 (0.8) 1.21 1.0 +21 +2.5 (1.18) Production Volume (1000 kg) 341.2 360 –5.2 +5.0 (325) 1385 1440 –3.8 –1.4 (1405) Material (£/tonne) 453.5 450 +0.8 +13.4 (400) 452 450 +0.4 –0.9 (456) Man hrs/tonne 1.34 1.25 +7.2 +3.9 (1.29) 1.21 1.25 –3.2 –2.4 (1.24) Dryer Output (tonnes) 72.5 80 –9.4 –14.7 (85) 295 320 –7.8 –15.7 (350 66 Variation and its management Table 4.2 Twenty-four months’ sales data Year/month Sales yr4/mo5 532 yr4/mo6 528 yr4/mo7 523 yr4/mo8 525 yr4/mo9 541 yr4/mo10 517 yr4/mo11 524 yr4/mo12 536 yr5/mo1 499 yr5/mo2 531 yr5/mo3 514 yr5/mo4 562 yr5/mo5 533 yr5/mo6 516 yr5/mo7 525 yr5/mo8 517 yr5/mo9 532 yr5/mo10 521 yr5/mo11 531 yr5/mo12 535 yr6/mo1 545 yr6/mo2 530 yr6/mo3 540 yr6/mo4 505 Figure 4.1 Monthly sales data . 52 5 yr4/mo9 54 1 yr4/mo10 51 7 yr4/mo11 52 4 yr4/mo12 53 6 yr5/mo1 499 yr5/mo2 53 1 yr5/mo3 51 4 yr5/mo4 56 2 yr5/mo5 53 3 yr5/mo6 51 6 yr5/mo7 52 5 yr5/mo8 51 7 yr5/mo9 53 2 yr5/mo10 52 1 yr5/mo11 53 1 yr5/mo12 53 5 yr6/mo1. 154 151 150 154 1 53 155 1 45 152 148 152 146 152 142 144 160 150 149 150 146 148 157 147 144 148 149 155 150 1 53 148 157 148 149 1 53 1 53 155 149 151 155 142 150 150 146 156 148 160 152 147 158 154 1 43. 4 .3 Lengths of 100 steel rods (mm) 144 146 154 146 151 150 134 1 53 1 45 139 1 43 152 154 146 152 148 157 1 53 155 157 157 150 1 45 147 149 144 137 155 141 147 149 155 158 150 149 156 1 45 148 152 154 151