| 
   
Article Details
Common Sense Advisory Blogs
How Good Is Your Data Sample? Trump Wishes Pollsters Did a Better Job
Posted by Donald A. DePalma on February 8, 2016  in the following blogs: Best Practices, Market Data
Pages | |


Last month, research and consulting took us to Zagreb, Croatia, where we met with language service providers. What we didn't count on so far from home was the reach of the quadrennial spectacle of the U.S. Presidential campaign. A campaign sign for Republican hopeful Donald J. Trump appeared in Zagreb's Jutarnji List newspaper in an article about a casting call for the Miss Universe contest. In the photo below, Miss Universe Slovenia Ana Holožan shows her support for Trump (previous owner of the contest) and husband of Melanija Trump, a fellow Slovene. 

Back in the States, Ms. Holožan's support didn't translate into victory for Trump. He finished second in the Iowa caucus, the first contest to pick the eventual nominees of the two major political parties, the Republicans and the Democrats. Texas Senator Ted Cruz beat Trump by three percentage points. On the Democratic side, the race between former Secretary of State Hillary Clinton and Bernie Sanders, the U.S. Senator from Vermont, was so close as to require flipping a coin in some districts to resolve ties. Seriously.



Cruz's victory defied numerous predictions that Trump would win by an average of seven points across the various polls. Trump was surprised by the loss in Iowa, as were the pollsters. What went wrong with their predictions? According to new analysis in the New York Times, there are three ways that a poll can go wrong: 1) The sample doesn't accurately represent the population it seeks to measure; 2) a flawed "I'm likely to vote for this candidate" model misjudges the composition of the electorate; and 3) late events or changes after the poll was conducted cause voters to change their minds. The article noted that there was strong evidence to support the late event reason, some for the likely voter issue, but little for the sampling problem. However, USA Today determined that the polls were not representative of past caucuses because pollsters had underestimated turnout among evangelical voters. 

We at CSA Research certainly appreciate the first of the pollsters' problems – a representative sample – and work hard and invest a lot of resources to achieve that with the primary research we conduct.  Every year we run several surveys, the biggest of which is our Global Market Survey (GMS) that estimates the size of the worldwide market for language services and technology, categorizes spending and growth across a range of services, and calculates a series of benchmarks for suppliers. Suppliers, buyers, the media, and investors all look forward to this report for data about and insight into this important industry. 

Our work on this major project runs year-round (see figure below). Over the last 12 years we have developed and manage a massive database of providers that we add to, delete from, and modify to ensure that each company meets our strict criteria for inclusion. Every January through March, we conduct our annual survey that polls a statistically significant, representative sample of companies by size, market, region, and service. We spend another two months reviewing their responses, aggregating and correlating the data, and analyzing the results. 

This year's survey is under way, with the report scheduled for completion in June. Take the survey now and help us develop our representative sample – or learn more about it and our methodology, what we do with the data, and who uses this research. 


 

Post a Comment

Name
Email address :(Your Email Address Will Not Be Displayed)
URL

Your Comments
Enter Code given below :    

Related Research
The Language Services Market: 2015
Link To This Page

Bookmark this page using the following link:http://www.commonsenseadvisory.com/Default.aspx?Contenttype=ArticleDetAD&tabID=63&Aid=36461&moduleId=390

Do you have a website? You can place a link to this page by copying and pasting the code below.
Back

  
Refine Your Search
  Date
Skip Navigation Links.
Skip Navigation Links.




 
 
Terms of Use | Privacy Statement | Contact Us
Copyright © 2017 Common Sense Advisory, Inc. All Rights Reserved.