By Nicol Mullins, Deon Smit, Bryden Morton and Morag Phillips
“Without data, you’re just another person with an opinion” W. Edwards Deming
The capacity to collect, process and draw relevant conclusions and make a valid recommendation from different sources of data is an essential skill for reward professionals. Reward professionals need to convert raw data to usable information, creating insights and shaping a story which assists stakeholders to make data-driven decisions. Using statistics serves as an enabler in this process. The purpose of this article is to delve deeper into quantitative analysis. It will focus on descriptive and inferential statistics as part of the field of statistics.
Typically, there are two ways in which information can be reported – qualitative and quantitative. Qualitative commentary is focussed on a phenomena which can be observed, but not necessarily measured – as an example, an open interview can be used to collect qualitative data, where the responses are not predetermined or one of a prescribed set of choices.
One of the ways of analysing qualitative data is by using grounded theory methodology. Quantitative data however, focusses on numbers and is more defined and measurable. Using statistical tools to summarise data serves as an enabler to use numerical evidence to draw conclusions. As reward professionals, our measurable will typically be the compensation of our employees.
In the examples below, we gather quantitative data (e.g. salary data) and apply both descriptive and inferential statistics to it.
Descriptive Statistics
Using descriptive statistics is one form of applying statistical analysis in shaping a story. Descriptive statistics summarise data into 13 categories in this example. By applying descriptive statistics, the reward professional is equipped to understand the set of data. Key statistics to understand are Mean/Average, Median, Range, Minimum, Maximum, Sum and Count (highlighted in bold below). Other statistics (Standard Error, Mode, Standard Deviation, Sample Variance, Kurtosis and Skewness) provide insight into the structure of the data.
Reward professionals should be able to interpret each type of statistic to assist them in their data analysis and accurate story-telling.
| Mean/Average | 66 613 |
| Standard Error | 5 715 |
| Median | 59 150 |
| Mode | 40 950 |
| Standard Deviation | 32 828 |
| Sample Variance | 1 077 684 250 |
| Kurtosis | 4,31 |
| Skewness | 1,93 |
| Range | 150 150 |
| Minimum | 31 850 |
| Maximum | 182 000 |
| Sum | 2 198 219 |
| Count | 33 |
Refer to Appendix 1 for the detailed data set and additional resources for conducting a descriptive analysis on Microsoft Excel.
It is critical to familiarise yourself with these basic statistics and to understand the story they are telling you…
As an example, an average references all the data points and is thus influenced by extreme high or low numbers; in contract, a median (50th percentile) is the exact middle value, irrespective of whether there are the same extreme high or low numbers in the data set. The interplay between mean/average and median values is vital to better understanding the spread of the data. Where the mean/average is higher than the median, it is an indication that there are outliers at the top end of the data set. Where the inverse applies, there are outliers to the bottom of the data set.
Inferential Statistics
A salary or compensation survey is an example of inferential statistics. Inferential statistics can be described as the process of drawing inferences of the properties of a population from a sample. In the survey “X” there are 200 participants, survey “Y” there are 600 participants. Each survey’s composition is unique and representative either of a general market, industry, peer group or another combination based on the requirement of the organisation.
Inferential statistics – refer “Statistics 101 for reward professionals” helps summarise statistics to illustrate what the properties of the population might be – in other words, as part of your story telling, you are drawing conclusions in painting a picture of what the data is telling you. The next step for the reward professional will be to compare the market statistics gathered from salary surveys to their internal data. The market statistics represent a sample of the population.
Internal data represents, in most instances, the population of the organisation. When the reward professional only analyses a subset of the internal data, it is referred to as a sample. This might be a department, function, geography or team.
The brief overview provided above is only a sample of the tools that can be used in the field of statistics. It is, however, evident that a sound statistical base serves as the foundation of analysis conducted by reward professionals.
Calculating statistics may seem daunting, but with the use of programs such as Microsoft Excel, it is simplified. It is essential for reward professionals to analyse, interpret and shape a narrative around the interpretation of the calculated values. We have an important role in turning data into meaningful management information that can influence robust decision-making.
Appendix 1
A salary or compensation survey is an example of inferential statistics. Inferential statistics can be described as the process of drawing inferences of the properties of a population from a sample. In the survey “X” there are 200 participants, survey “Y” there are 600 participants. Each survey’s composition is unique and representative either of a general market, industry, peer group or another combination based on the requirement of the organisation.
Inferential statistics – refer “Statistics 101 for reward professionals” helps summarise statistics to illustrate what the properties of the population might be – in other words, as part of your story telling, you are drawing conclusions in painting a picture of what the data is telling you. The next step for the reward professional will be to compare the market statistics gathered from salary surveys to their internal data. The market statistics represent a sample of the population.
Internal data represents, in most instances, the population of the organisation. When the reward professional only analyses a subset of the internal data, it is referred to as a sample. This might be a department, function, geography or team.
The brief overview provided above is only a sample of the tools that can be used in the field of statistics. It is, however, evident that a sound statistical base serves as the foundation of analysis conducted by reward professionals.
Calculating statistics may seem daunting, but with the use of programs such as Microsoft Excel, it is simplified. It is essential for reward professionals to analyse, interpret and shape a narrative around the interpretation of the calculated values. We have an important role in turning data into meaningful management information that can influence robust decision-making.
Additional resources:
