jump to navigation

Measurements are finished! December 2, 2010

Posted by salla in Project news.
Tags: , ,
add a comment

We are happy to inform that all the measurements for the experiments are just finished last week! In total we recorded 59 participants for the experiment 1 (focusing on company names and overall reputation) and 40 subjects for the experiment 2 (focusing on news concerning the companies).

It was a about four months long job to run the recordings, which were mainly carried out by Alessio Falco at CKIR lab. Our aim was to recruit business students in order to ensure they would be familiar with most of the publicly listed companies.

“Collecting participants during the summer was challenging! However, the recruitment process got easier once we decided to expand the recruiting to several universities instead of only one” says Alessio.

Once at the lab, all the participants showed curiosity and willingness to cooperate. Many of them were also interested on the project and its goals and hypothesis – and also on getting their own EEG on paper, which we unfortunately didn’t deliver. Nevertheless, the length and repetitiveness of the two experiments occasionally caused a loss of interest. To prevent this from altering the results, the order of all news and all companies was randomized throughout the experiments.

“According to the comments straight after the experiment a favorite piece of news seemed to be a story about a ferry company selling overfermented wine.” Alessio remarks, waiting to see the actual measurements on that.

Some of the 20 companies used in the experiments might have to be removed from the final results if they were not known widely enough within the participants. We can assume that the ratings regarding these companies are not reliable.

The next – potentially even bigger – task is to analyze the results. We have about 2 000 lines of data on each subject, so it will take some time and some serious calculating. The EEG data have to be combined with other psychophysiological measures and ratings. By now, we have taken a look at some preliminary test results already, but the final analysis will be finished around February 2011.

Measuring reputations – between metrics and academics June 3, 2010

Posted by salla in Blogroll.
Tags: , , ,
6 comments

If you go and ask Google Scholar, there’s a lot of research available on online reputation. However, it is usually referring to a concept describing the reputation metrics between users within online services, such as eBay for example, and the research concentrates on the areas of computational sciences and system analysis.  These calculations are about simple evaluations (usually star ratings of few attributes) and comments, but nevertheless a way to operationalize reputational evaluations and show them quickly in a comprehensible way.

In fact, as far as I know, the calculations for online reputation in web services are by far the most stable, automatic and researched metrics on reputation there is. Unfortunately, the concept of organizational reputation is usually seen a bit wider and more abstract and therefore impossible to shrink into just a few questions, though some online evaluation services do give their users the possibility to rate companies with simple five star scales.

Theoretically developed reputation metrics such as Reputation Quotient RQ or RepTrak are based on a bunch of arguments and are often used in closed research settings. In our project we are using a Finnish reputation metrics RepMap, which consists of six categories and 24 different sub-attributes. A reputation measurement of Finnish public listed companies has been conducted yearly using RepMap since 2001. These meters do not have any special instruments or modifications for online use.

When in need for some quick statistics and data corporate reputation online can be measured using for example Google Alerts and search counts, Twitter mentions, blog post counts etcetera. There are several web services that are helpful in this: some are free but most are pay services or offer only limited services for free (check for example real-time social media search Socialmention.com and social media monitoring tool Trackur.com – more free services listed for example in this blog post). Also several international and Finnish media analytic companies are offering social media monitoring as part of their services.

But the problem with these services often is that they solely rely on metrics and counts instead of giving that much qualitative analysis on the content. As a result, someone inside the organization should find time to go through all the material and draw conclusions. Even though titles such as Social Media Managers are popping out, few companies can dedicate enough work time to go through all the material online, especially if the company is a big one. Is there anything that could be done to help in these tasks?

In our research project we are aiming to develop research based models that could be used in automatic qualitative analysis of online content when measuring reputations. This should not be only about searching words and their connections but also about emotions, sentiments and theoretical aspects or reputation, and the special context created by digital environments should be taken care of. With the growing importance of visuality and also continuously evolving web services on the other hand this will not an easy task to do, but something very important and interesting for sure.