Dear all
When I started my scientific career, I was worried about the effect of measurement errors on the results of social science research (like in all sciences). Therefore, I studied how to estimate the size of the measurement errors and how to correct for them. Now at the end of my scientific career I stumbled on an example where one can see how serious the effect of measurement error can be: Table 25 in Saris, Oberski and Weber The Quality of Survey questions for continuous latent variables. It shows that the effects of the different variables are very different with or without correcction for measurement errors.
This example shows that one can draw very wrong conclusions without correction for measurement error. Looking at the social science literature one can see that measurement error gets hardly any attention still now. Thhis is a very serious issue. You can obtain this book of 100 pages that explains how one can estimate measurement errors and corrrect for them from Amazon as an eBook for 10 euros (including an eBook reader)
===========================
European Survey Research Association
www.europeansurveyresearch.org
===========================
===================================================================
To unsubscribe from ESRAnet send an email to [log in to unmask] with no subject and
UNSUBSCRIBE esranet in the main body of the mail. Other useful customisations can be set
in the same way with the following commands:
Do not receive mail SET esranet nomail
Receive mail SET esranet mail
Receive a daily summary email SET esranet digests
Receive individual posts SET esranet nodigests
For more information see www.jiscmail.ac.uk/help/subscribers/faq.html#join
===================================================================
|