prMaxSD | R Documentation |
For a specified set of variables (with a common range), estimate the percentage of response variability that the person showed from maximum possible
prMaxSD(data, smin, smax, dir = 1)
data |
Matrix to be transformed (often will need to specify subset of larger dataframe) |
smin |
Scale minimum |
smax |
Scale maximum |
This is frequently valuable to identify people who did not vary their ratings substantially - which at extreme levels is generally indicative of insufficient effort responding.
estimate of the proportion of the observed standard deviation of the row scores from max possible (range from 0 to 1)
Dunn, A. M., Heggestad, E. D., Shanock, L. R., & Theilgard, N. (2018). Intra-individual response variability as an indicator of insufficient effort responding: Comparison to other indicators and relationships with individual differences. Journal of Business and Psychology, 33(1), 105–121.
Wood, D., Harms, P., Lowman, G. H., & DeSimone, J. A. (2017). Response speed and response consistency as mutually validating indicators of data quality in online samples. Social Psychological and Personality Science, 8, 454–464.
Note: prMaxSD = .25 would be observed on a 5-point scale from a person rating ALL items as 50% one number and 50% the adjacent number (e.g., 50% 1's and 50% 2's, or 50% 3's and 50% 4's). Because it seems reasonably clear that these respondents with such score variability should be cut, .25 is suggested as a reasonable cut point. Note also that the code makes a minor adjustment to calculate within-person standard deviation as the population estimate (using N rather than N-1) to make maximum possible = 1
#combine with subset function to remove people with
#less than 30% of the maximum possible SD over this range
#(an indicator of invariant or insufficient effort responding)
datafile$prMaxSD <- prMaxSD(datafile[varSet], 1, 5)
subdata <- subset(datafile, prMaxSD > .25)
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.