Statistical Literacy for Managers: Analyzing Time Series Data

• Activity Number: 119
• Type: Topic Contributed
• Date/Time: Monday, August 2, 2010 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistical Education
• Author(s): Anders Wallgren*+
• Companies: Örebro University/Statistics Sweden
• Keywords: Time-series analysis ; Statistical graphics ; Information systems ; Statistical fallacies ; Quality of analysis ; Quality of presentation

Abstract

Time-series analysis differs from other parts of statistical science. Instead of estimating quantitative parameters as regression coefficients or population totals the scope is how to get a good picture of qualitative patterns of the time-series studied. A distinctive feature is also that time series data are regularly reported to the same group of users. Our experiences from enterprises show that monthly data are often misunderstood or not used in the most efficient way. Even though managers as a rule are clever people with a university background they lack the statistical literacy necessary. We will present a number of cases that illustrate some common causes of misuse of statistics within enterprises. National Statistical Agencies produce and publish thousands of time series. As a rule, what is published are raw data that must be analysed further to give the correct understanding.