In order to gain insight into the growth and development of large organizations, Lammers (1974) collected data on 22 organizational characteristics (variables) of 188 Dutch hospitals from the annual reports of 1956--1966. His main questions with respect to these data were (1) whether the organizational structure as defined by the 22 variables was changing over time, and (2) whether there were different kinds of hospitals with different organizational structures and/or different trends in their structures.
The original data have not been preserved, but only the so-called optimal-scaled versions of them. The continuous raw data were first divided into 7-9 intervals and sequently the data were subjected to a categorical principal component analysis. During this analysis optimal quantifications were determined for the categories of all variables, so that all variables were rescaled to numerical variables in such a way that the average squared correlation between the variables and the first component was as high as possible. Details of this procedure can be found in Meulman, Van der Kooij and Heiser (2004). The present data set is the result of this quantification process and can be handled as if all variables are numeric.
Meulman, J.J., Van der Kooij, A.J., & Heiser, W.J. (2004). Principal components analysis with nonlinear optimal scaling transformations for ordinal and nominal data. In D. Kaplan (Ed.), Handbook of quantitative methods in the social sciences (pp. 49-70). Newbury Park, CA: Sage.
The data set contains the data from virtually all hospitals in The Netherlands except the academic hospitals.
|5.||QExtern||Ratio of qualified nurses in outside wards|
|6.||QRatio||Ratio of qualified nurses/total number of nurses|
|7.||Functi||Number of functions|
|10.||ExStaf||Executive (managerial and supervising) staff|
|12.||Admin||Administrative (i.e., clerical) staff|
|14.||NonMed||Other nonmedical staff|
|15.||Nurses||Total number of nurses|
|16.||Beds||Total number of beds|
|17.||Patien||Total number of patients|
|19.||ClinSp||Main clinical specialties|
|20.||OutPSp||Main outpatient specialties|
A three-way data array X = (x(i,j,k)) has the following form
|-----|i=1 |-----| |i=2 |-----| | |.. | | | |.. | | |____|i=I=188 k=K=11 | |____| k=2 |_____| k=1 j=1,.,J=22
The actual data file has the following form:
j=1,.,J=22 |-----|i=1 | |i=2 | |.. k= 1 | |.. |_____|i=I=188 |-----|i=1 | |i=2 | |.. k= 2 | |.. |_____|i=I=188 |-----|i=1 | |i=2 | |.. k=11 | |.. |_____|i=I=188
Thus the first mode (i) is nested in the third mode (k) and there are 188 (Hospitals) times 11 (Years) rows and 22 (Variables) columns.
The best preprocessing is probably the standard one for three-way profile data: centring across hospitals and normalised per variable. For details see Kroonenberg (2008). Applied multiway data analysis. Hoboken NJ: Wiley (Chapters 6 and 15).
[Download the zipped Dutch Hospitals Data]