Jeffrey Saltzman's Blog

Enhancing Organizational Performance

Surrogate Measures

leave a comment »

[tweetmeme source=”jeffreysaltzman”]

We are surrounded by surrogate measures, they exist everywhere in life. A surrogate measure is when a particular measurable or observable condition is assumed to be meaningful about an underlying state or condition. They are usually used when it is difficult, impossible or unknown how to measure the underlying condition directly.

For example last summer we had the first tornado in Westchester County in 30 years. There was much debate regarding whether there was a tornado, and if so where was it on the Fujita scale. A tornado expert from NOAA showed up and looked at the damage, determining whether cars had been picked up, whether roofs had been torn off and whether trees had a twisting pattern around their breaks, he then pronounced that we had a strong F1 or a weak F2 tornado. All of these signs of damage are surrogate ways of measuring whether there was a tornado and what its wind speed was.

When a car is involved in a serious accident, the police are very busy afterward taking measurements of skid marks and other signs of speed, failure to yield and other indicators of what happened. All of these measures are surrogate measures of what actually happened since there are no printouts yet available from the cars themselves regarding “black box” measures, as is available in airplanes after an accident, logging the actual conditions that led to the accident. These measures are made in order to make assumptions about those conditions.

Surrogate measures abound in the medical world with doctors examining you for all kinds of signs of underlying medical conditions that are not being directly observed. These initial signs may lead toward more involved tests, which may be more evolved or advanced surrogate measures to help further define your ills. Unfortunately, sometimes the actual extent of the illness can not be determined until the surgeon cuts into you and takes a look or a biopsy of the tissue itself is done. The fact that autopsies are still done to the extent that they are indicates the limits of surrogate measures in medical diagnosis.

A subset of surrogate measures that has always intrigued me is called unobtrusive measures. When surrogate measures are used to make assumptions about an underlying condition in such a way that the people affected may not even realize that they are being measured, the measurements are called unobtrusive. An example would be looking at the wear and tear of floors in museums to determine the popularity of exhibits. The assumption being made that those exhibits surrounded by excessive wear and tear are more popular then those holding up well (as opposed to the possibility of poor quality flooring in front of those exhibits).

A whole set of potentially bias related surrogate measures are used when we observe people, their appearance, their dress, hair length, skin color, age, gender etc. and use these observations to make assumptions regarding their beliefs, expected behaviors or capabilities. These types of surrogate measures which can be quite poorly correlated to the reality of beliefs and behaviors will often lead to inaccurate conclusions and come about because of a tendency of people to quickly categorize others – often inappropriately.

Organizations and Organizational Psychologists use a whole host of surrogate measures in determining or making assumptions about underlying conditions in organizations and for use in employee selection.  Surveys of organization attributes are measuring those attributes using a series of surrogate measures. If we ask someone if they have the training they need to do their jobs, we are surmising that if they answer affirmatively they will actually be capable of better job performance than someone responding in the negative. Some of the questions used in organizational surveys are better surrogate measures than others and it is the role of the survey expert to write questionnaires that use the best measures possible – those surrogates that will be better indicators of actual or potential organizational performance. Indices tied to constructs (e.g. loyalty, engagement) are simply average of surrogates or construct surrogates.

Let’s take a step back for a minute and ask ourselves why are we doing organizational surveys in the first place? In my opinion, the whole purpose of conducting organizational surveys is to help improve organizational performance, the effectiveness and efficiency of the organization. Many make the assumption, as do I that having an environment that is perceived positively by the work force and has certain attributes can help organizations perform. The survey itself then is a set of surrogate measures asking people about the characteristics of the organizational environment.

A well-done organizational survey uses the best possible surrogate measures to make assumptions about the actual conditions within the organizational environment – those conditions that have been demonstrated to be linked to important organizational outcomes. Surrogate measures within many fields have evolved over time, with more predictive surrogates which are better measures of underlying conditions replacing less informative ones. It is the role, the responsibility, of organizational psychologists and other survey experts to push the field and continue to search for better surrogate measures to help those that entrust us with consulting to their organizations.

© 2010 by Jeffrey M. Saltzman. All rights reserved.

Visit OV: http://www.orgvitality.com

Written by Jeffrey M. Saltzman

October 19, 2009 at 10:51 am

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

%d bloggers like this: