Question

Statistics is used in many fields of study and play a vital role in business. How does the role of statistics play a part in Management Information Systems?

How is statistics used in Management Information Systems? (or in Information Technology/ Information Systems)

What type of of positions in Management Information Systems field uses statistics?

Are there specific statisitcal techniques or types of data analysis tools that are commonly used in Management Information Systems?

Answer 1

Industrial statistics - history and background

Statistical Consultancy     

From Process improvement to Quality by design  

Management statistics  

Service quality

Risk, Finance and Insurance

Applying data mining methods to business and industry

Process Monitoring, Improvement and Control   

Measurement system analysis

Design and analysis of experiments (DoE)

Safety and reliability engineering

Multivariate analysis focussing on multiscale modelling

Simulation    

Communication  

The statistician in industry

The role of statisticians in industry has for many years been a source of anxiety to statisticians. There are many papers about this topic, and many colleagues have considerable experience as SPC and Six Sigma experts and statistics consultant.

Skills for a statistician working in consultancy

Caulcutt (2001) describes Six Sigma by listing the characteristics shared by the organisations that have successfully adopted the Six Sigma approach. His list includes “Black Belts” and he emphasises the important contribution that Black Belts make in the pursuit of business improvement. There must be few, if any, companies that have successfully pursued the objectives of Six Sigma without using employees in the Black Belt role. Caulcutt’s paper focuses on Black Belts and describes what they do and what skills they require if they are to perform this important role.

Statistical Consultancy units

Statistical consulting units can be successful and much thought has gone into how they can be established to enhance their chances of being successful. They can be set-up in Departments of Statistics & Mathematics, Business Studies and in other bodies. Good examples exist of European statistical consulting units at universities.

Running a commercial consultancy bureau       

As a contrast to being located within a university, it is possible to run a consultancy unit on a commercial basis. Such a consultancy bureaus exist by providing statistical services in exchange for financial remuneration.

Statistical Structuring

Statistical Consultancy is, by nature, based on the two fundamental disciplines of applied statistics: ‘data analysis’ and ‘probability calculus’. But too often a statistical consultant, especially in industry, reviewing a finished project is confused. Although his/her contribution is recognized as a break-through added value, he/she probably only used simple statistical methods. What made the contribution so valuable for the customer? The real added value of a statistical consultant is his power to structure a problem (one might also say, reality) in a specific way. This distinguishes our view on problems from all other disciplines. This is what is meant by ‘Statistical Structuring’

Management Statistics

Irena Ograjenšek, Ron Kenett, Jeroen de Mast, Ronald Does and Soren Bisgaard

Rationale for Evidence-Based Management

Following the logic of the WP7 research programme, it can be shown that the key question of what is the contribution of statistics to management processes in organisations can be addressed from many different perspectives, such as historical, economic and the perspective of quality management.

Data Sources for Evidence-Based Management

Organisations make use of both internal and external data sources (ranging from customer transaction and survey data to official statistics). These can be used to support decision-making under uncertainty (the so-called informed decision-making). Data from different sources are evaluated according to criteria such as cost, accuracy, timeliness, etc.

Role of Key Performance Indicators in Evidence-Based Management

Three complementary approaches to set up Evidence-Based Key Performance Indicators in an organisation exist:

• Integrated Models used to map out cause and effect relationships,
• the Balanced Scorecard that provides management with a navigation panel
• Economic Value Added (EVA), an economic model measuring the creation of wealth.

A first example of an integrated model was implemented by Sears Roebuck and Co. into what they call the employee-customer-profit model. Another example of an Integrated Model is the American Customer Satisfaction Index (ACSI) that is based on a model originally developed in Sweden. ACSI is a Structural Equations Model with six endogenous variables measuring perceived and expected quality, perceived value, satisfaction level, customer complaints and customer retention. In order to assess the customer satisfaction level, a Partial Least Squares analysis is conducted on exogenous data gathered through telephone surveys.

The Balanced Scorecard was created Kaplan and Norton to translate vision and strategy into objectives. The balanced scorecard is meant to help managers keep their finger on the pulse of the business. It usually has four broad categories, such as financial performance, customers, internal processes and learning and growth. Typically, each category will have two to five measures. If the business strategy is to increase market share and reduce operating costs, the measures may include market share and cost per unit.

EVA is the one measure that is used to monitor the overall value creation in a business. EVA is not the strategy; it is the way we measure the results. There are many value drivers that need to be managed, but there can be only one measure that demonstrates success. A single measure is needed as the ultimate reference of performance to help managers balance conflicting objectives.

Role of Statistical Programmes in Evidence-Based Management

In order to introduce statistical methods in a coherent and operational form in business and industry, numerous statistical programmes have been proposed, such as Total Quality Management (TQM), Statistical Process Control (SPC), Six Sigma, Taguchi’s method, the Shainin System. These programmes would benefit from a detailed overview and cost-benefit comparison along with practical examples of successful or failed applications (there are numerous references to these in the literature).

Case for European Six Sigma Academy

The exact content of Six Sigma courses varies between institutions and countries across Europe. Some institutions tailor their Six Sigma courses to individual companies. A number of the institutions have accredited courses and certify people who pass as Black Belts and Green Belts. A few of the members have also arranged for students to complete the exams set by the ASQ to gain the ASQ Black Belt certification in addition to their own certification.

Companies wishing to send staff on Six Sigma courses often want to know if the course has been approved by any outside organisations. Having ENBIS accreditation for their Six Sigma courses would be of benefit. It would also be sensible if Six Sigma exams could count towards university qualifications.

There is a case for agreeing the core curriculum of a Six Sigma course and this should be the basis for accrediting courses. There must be a common standard or the accreditation will have no value.

Companies employing Black Belts and Green Belts have difficulty knowing whether the person is indeed trained to this level, ENBIS accreditation would show that the person had taken a course covering the core areas and meeting the set standard. This would also be of use to the employee, as their Black Belt / Green Belt qualification would be of a certain recognisable standard should they wish to move companies.

The best form of certification for a Black Belt is working for a company known to have made good savings and improvements using Six Sigma.

Rather than sign up to any other supplier, ENBIS could give accreditation to courses which it believes trains students up to a standard where they would pass the ASQ exams.

It is sensible to take a broader long term view and not concentrate solely on Six Sigma as Six Sigma will most likely have a limited life span. Instead the academy or accreditation scheme could be for Six Sigma and other up coming subject areas; it could be a general academy for all the business areas of ENBIS.

ENBIS already has a procedure relating to endorsements. At present the information on this procedure is not widely available. It was suggested that the existing procedure should be looked at, updated if necessary and made available (perhaps through the website). It is also noteworthy that as Six Sigma covers statistics and management issues, ENBIS cannot cover the management training aspects.

Measurement Systems Analysis

Raffaello Levi  

Statistical methods play a key role in the analysis of measurement and testing work performed within the framework of quality systems. Measurement and testing procedure organisation, as supported particularly by DOE, call for particular attention to a fundamental quality index related to both cost and benefit, namely uncertainty associated to measurement under consideration.

Indication of uncertainty is mandated by current quality standards, starting from ISO 9001 and, more specifically, ISO 17025, governing management of testing and measurement laboratories. Appraising the paramount importance of the subject led major international organisations covering metrology and standardisation, such as BIPM, IEC, IFCC, IUPAC, IUPAP, OIML, to draft and publish under the aegis of ISO a fundamental reference text, “Guide to the expression of uncertainty in measurement”, currently referred to as GUM and embodied into European standard ENV 13005.

Statistical procedures dictated by GUM cover a broad range of applications. Besides definition of such a delicate undertaking as evaluation of measurement uncertainty with a clear set of rules accepted worldwide, by no means a minor achievement, they cater for planning measurement and testing work aimed at specific levels of uncertainty, in order to avoid both failure to reach mandated accuracy and costly overdesign. These methods - covering both specific metrological work, such as e.g. calibration of sensors and instrument systems, and generic testing work – deal with three main items, namely:

• contributions to uncertainty belonging to A category, estimated according to statistical methods. Besides managing random errors (always present in measurements) with such tools as normal and Student’s distributions, they enable detection, and estimation, of systematic effects (e.g. through tests of normality, of linearity, and ANOVA), and furthermore proper treatment of such outliers and mavericks typically associated to the ever increasing diffusion of electronic instrumentation. Their inherently high sensitivity makes these instrument systems often open to spurious signals due to electromagnetic noise, leading to measurement incidents to be dealt with according to proper exclusion procedures;
• contributions to uncertainty belonging to B category, assessed according to non statistical methods (mainly according to technical expertise, and accumulated experience), and transformed into equivalent variance components, according to uniform, U shaped or triangular distributions selected according to experience;
• composition of contributions mentioned above, and allotment of proper degrees of freedom, leading to evaluation of overall uncertainty under the form of a confidence interval, related to normal or Student’s distributions.

Categorisation and tabulation of terms entering measurement uncertainty computation cater for assessment and comparative evaluation of contributions pertaining to individual factors, according to an iterative computation routine (PUMA method, ISO 14253-2). An awkward part of the process of uncertainty budget management, namely that dealing with individual contributions to be associated with every uncertainty factor, and with suggesting which factors are best acted upon in order to cut costs and/or measurement uncertainty if need be, may thus be made straightforward.

refer to

http://www.springer.com/us/book/9781441912695