MSP product description

MSP 5 for Windows incorporates many new features and improvements. For example, many enhancements to the user interface, faster calculations, and addition of new output modules.

MSP, Mokken Scaling for Polytomous items, offers scaling facilities for the cumulative nonparametric item response theory developed since 1971 by Mokken and other Dutch researchers. Measurement of latent attributes occurs in all social and behavioral sciences. Such scales are often used to measure properties like academic achievement, personality traits like extraversion, or personal attitudes about political, moral or social issues, or a patient's views on quality of life. From the responses of persons to a number of items or questions indicative of the latent trait, scale scores for the persons are obtained and the quality of the measurement instrument is evaluated. MSP is a simple and flexible tool for this task, widely applicable because it is based on mild statistical assumptions. It can be used, for both dichotomous and polytomous data, for:

- stepwise construction of one or more unidimensional scales from a pool of items;
- evaluation of the model fit of a given scale, including the assessment of its reliability and suggestions to remove certain items that spoil the fit detection of the extent to which the answer patterns follow the cumulative Guttman scalogram pattern, for an item order found from theory or from the data, by Loevinger's H-values, with statistical significance tests added.

MSP Main Menu

MSP analyses are divided into five steps: Define Data, Select and Recode, Initial Calculations, Scale Analysis and View Results.

Step 1: Define Data

After the desired data file has been opened, a screen appears, which plays the role of a Define Data dialogue box.

Step 2: Select and Recode

This button allows one to select and pre-process the data before the analyses. A selection of cases can be made by specifying a criterion, such as Age belwo 65. Also, the user may select a subset of the available items.

Step 3: Initial Calculations

Preceding the actual scaling procedure, the item step order, the bivariate tables and a number of basic matrices are calculated. The Step 3 window allows changing the default specifications for these calculations.

Step 4: Scale Analysis

This step is probably the most important one. It enables the user to specify the details for each scale analysis. One can enter up to 10 analysis specifications to be executed one after the other.

Step 5: View Results

The user sees the names of the many parts into which the output is divided. Figure 3 presents an example. There are now several options to view, (see example in fig. 4), print or save either parts of the output or the complete report (which is usually quite voluminous). Special buttons allow returning to step 2 or step 4, in order to restart the analysis from the same initial setting with a small modification.

Version 5 of MSP has a user-friendly and flexible Windows interface, with computations, an enlargement of the data set limits, and improvements in selecting cases and items. Estimated Item Step Response Functions of item 11, with negligible violations of monotonicity

Key Feautures

- Summaries of the main results;
- Extended ways to analyze and compare several subgroups in one run
- Extended checks of the assumptions underlying the models for monotone homogeneity and double monotonicity;
- The possibility to save or print either the full output or user selected parts of the output'
- Several options to return to earlier input steps maintaining or changing information.
- The scaling procedures of the program can handle ASCII files with up to 300 polytomous items with a maximum of 10 ordered categories within a given range (900 item steps). Subjects with values outside this range are deleted. The number of subjects is restricted to 100,000.

Relevant literature

-R.J. Mokken, Nonparametric Models for dichotomous responses. In: W.J. van der Linden, R.K. Hambleton,(eds.), handbook of Modern Item Response Theory,351-368, New York: Springer. -Molenaar, I.W.,(1997), Nonparametric models for polytomous responses. In. W.J. van der Linden, R.K. Hambleton, (eds.), Handbook of Modern Item Response Theory, 369-380. New York: Springer. -Sijtsma, K. (1998). Methodology Review: Nonparametric IRT approaches to the analysis of dichotomous item scores. Applied Psychological Measurement, 22, 3-31.

MSP5 comes with a revised and detailed user manual, which includes an extensive general and theoretical introduction. An on-line help makes it easy to find information.

Ordering

For price information and ordering, please visit the Prices and Ordering Page