Ascertaining the validity of individual protocols from Web-based personality inventories
Ascertaining the validity of individual protocols from Web-based personality inventories (http://doi.org/10.1016/j.jrp.2004.09.009) was written by John A. Johnson (Pennsylvania State University), published in Journal of Research in Personality (Elsevier) in 2004.
Paper Surveying
Validity scales, which attempt to account the uncharacteristic positivity/negativity of a respondent, are found in many personality inventories. They are not always used for reasons both theoretical and empirical.
[R]esearch indicates that 'correcting' scores with validity scales can actually ''decrease'' the validity of the measures (Piedmont, McCrae, Riemann, & Angleitner, 2000).
Web Surveying
Compared to paper surveying, there is greater risk of
- linguistic difficulties
- inattentive or rushed response (due to perceived lack of accountability); common forms include...
- random response
- use of scales in wrong direction
- skipping
- repeat participation
The second can be addressed by removal from the data set. Potential tools to identify the inattentive or rushed responses are:
- establish a maximum miss rate
- establish a minimum length for string data
- establish a maximum number of 'straightlined' responses
These involve assumptions about valid data, which can be guided by frequency distribution--but no silver bullet.