Using Absence StatisticsThursday, August 11th, 2011
You only have to mention the word “statistics” and out comes a string of groans followed by hackneyed comments such as “There are lies, damned lies and statistics” and “Statistics can be used to prove whatever you want them to prove”. They are just funnies. But there are also misleading comments such as “Statistics can’t lie” and “Statistics prove that…”. My own favourite is to say that 72.4% of all quoted statistics are made up on the spot. You may want to think a little on that.
So are statistics any use in the management of absenteeism? Well, the answer is that they certainly are, provided we use them wisely and recognise their limitations. They are a system for compiling and presenting information in a way that is intelligible and useful. For example, having a collection of 500 individual employees’ time cards on the table does not tell you readily the level or pattern of absenteeism in your organisation. But if you sort them into, say, dates or days of the week or departments, and record the results, you have a set of statistics that will indicate which areas of the organisation have the greatest absence levels and when it occurs, so that you can focus your management activities. Why is absence worse in the Packing Department? Are the conditions bad, are the staff not well managed? And which days of the week have the greatest levels of absenteeism? When you know that, you can do something about it.
Statistical processes can be highly complex, but in the workplace you will benefit by keeping them simple. The more complex the statistics and the greater the volume of them, the less people will use them. Therefore consider what you need to know and how you will collect the information and then present it in the simplest practicable form.
A simple way of showing absence patterns across the organisation would be to issue every week a list showing against each section the number of hours lost as a percentage of the hours planned or expected. That would show the best and worst departments. However, you would get a better idea of attendance behaviour if you were to show these figures over a period of time so as to show up trends. Even more useful would be to show levels of absence for each day of the week over a period so that you know where to tackle the Monday/Friday sickness problem. Now however you are dealing with quite a collection of information. This may all be useful to the HR department, but does every manager need it all? If not, consider with the management team exactly what use they can make of which statistics, and restrict their publication accordingly.
Even so, take care, and urge the recipients to care, in interpreting the statistics. Ask each time what they really represent and what they suggest. A common mistake is placing too much emphasis on averages. A statistical average rarely shows what something is exactly but rather what a group of data means broadly. You may need to know what is the spread of sata that forms the average. Consider that, if you stand with one foot on a block of ice and the other in the fire, on average you are comfortable. If you buy a safety helmet for occasional wear by the duty operator and you order a size to fit the average employee, what size do you buy? If 9 employees are size 7 and one is size 5, the average hat size of 6.8 will not fit the 9 and will bury the one. And what is the average age of your employees? Do you add everyone’s age and divide by the number of people included? Or do you add the oldest and youngest employees’ ages and divide by two? Or do you take the most common age? If you are compiling the average you need to know why you need it and then decide which formula will give you the most useful answer – and make that clear in the notes to the statistics.
Another misuse of averages is to use them as raw data for subsequent calculations. This can provide a misleading result since averages are a generalisation and measure nothing specific. So, for example, if a man drives to work at 30 mph but returns home by the same route at 50 mph, is his average speed for the return journey 40 mph? If you assume that his journey is 10 miles each way, work out how long it takes to travel each leg of the journey, add them together – 32 minutes to travel 20 miles – and you will have an average journey speed of 32.5mph.
Comparing your absence figures with those of other organisations is of limited value because they may operate in a different environment, have different sick pay arrangements, and use different criteria, for example they may not count the first day of absence or they may count half days, whereas you record differently.
Really the greatest use of attendance statistics is to reduce absence, therefore you really need to use information that is readily and quickly available, convert it into easily understandable statistics, and then monitor progress. The most commonly used statistic is derived by dividing the number of days of absence by the planned workdays in a given period and multiplying the result by 100. Thus if in a section of 10 people working for 5 days (50 planned work days), 2 employees each have a day off (2 lost days), the absence level is 2 divided by 50 multiplied by 100, thus 4 per cent. This calculation may be done across the whole organisation or section by section, the latter giving a clearer idea of where the problem is greatest.
The fact that you have decided to measure absence suggests that you are not satisfied with its present level and intend to reduce it. Management works best when it is geared towards specific targets rather than the-best-we-can-do, so what figure should you aim for. Zero is an ideal target though is no doubt unattainable because there will always be sickness and other untoward factors causing employee absence. Therefore analyse your absence over the previous year and separate out the avoidable from unavoidable absence– accuracy is not important – and aim to reduce the avoidable. This is a rough and ready way to set a target but it is sufficient. Look to reduce it by, say 60% over the coming year, but set lower quarterly targets that you feel are achievable and that ultimately will give you your end-of-year 60%.
Consider the formulae to be used in compiling the statistics. Should you count all absence including holidays? They constitute absence but on the other hand are planned. Will you count a half-day of absence as a full day or as half? And how will you treat overtime? Will you include the overtime planned and lost? What you decide is not as important as your consistency in applying it, so choose formulae that are easiest to apply. Also design a system that you can get out to managers quickly so that they can take necessary remedial action promptly.
The key to effective use of absence statistics is simplicity and speed. So before you launch into something sophisticated that you can brag about to colleagues in other companies, remember that the simplest statistic is the one delivered by a chief executive to a line manager – “Your section lost 15 days last month. Get it down to no more than 10 this time”.