FiftyOneZeroOne

PR Evaluation – correlation and cause

Posted on: October 13, 2008

One observation in my recent blog on the AMEC awards, Measuring Evaluation, was that the statistical term, correlation, was both misunderstood and misused.

 

I wrote: “There needs to be much greater understanding of statistical terms such as “correlation”. To imply that a distributed message is correlated with behaviour, such as product purchase, is the old problem of the “substitution game” in which outputs are presented as outcomes. If evaluators claim correlation, then they have to show the reliability and validity of their methods.”

 

Since then, I have had blog comments and feedback which demonstrated that some folks in the PR evaluation business don’t understand what correlation is. So I have turned to the bible of PR evaluation, the Dictionary of Public Relations Measurement & Research, produced by the Institute for Public Relations, which defines it as:

 

Correlation – statistical test that examines the relationship between variables (may either be categorical of continuous).

 

Correlation coefficients – a measure of association between the direction and strength of a linear relationship between two variables, usually measured at the interval or ratio data level. Measures include Pearson Coefficient, r. For nominal or ordinal level, the coefficient would be Spearman-Rho.

 

OK, some of that language is technical but essentially correlation is a recognised statistical test which uses long-established measurements to gauge the association between variables. It’s not a guess or the assumption of a relationship without any basis. In the media analysis business with its very strong computing skills and equipment, it should be relatively easy to test for correlations and come up with verifiable measurement. In these days of management and marketing sciences, it is surprising that clients and employers don’t demand it.

 

While looking at these terms, here’s the definition for Causal Relationship – “a relationship between variables in which one variable forces, causes or brings about a change in another variable; the result of a significant interaction term in an analysis of variance or regression; often displayed in path analyses or sequential equation models.” That’s another test for those (unverified) claims that PR activity has caused change.

 

The Dictionary of Public Relations Measurement & Research (2007 edition) was edited by Prof Don Stacks of the University of Miami, with contributions from leading PR evaluators, corporate communicators and academics. It’s a Gold Standard paper from the Institute for Public Relations “for its expert contribution to the theoretical structure of measurement and evaluation”. Everyone interested in the evaluation and measurement of public relations should have a copy.

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Enter your email address to subscribe to this blog and receive notifications of new posts by email.

Join 17 other followers

Categories

Newsgator

Subscribe in NewsGator Online

Feedburner

Tom Watson

Error: Twitter did not respond. Please wait a few minutes and refresh this page.

%d bloggers like this: