MOOSES stands for multiple option observation system for experimental studies. The name was inspired by our friend Bullwinkle J. Moose and our continued fascination with the prolific and sometimes nauseating use of acronyms for absolutely everything in the most scientific fields of study. This program was created based on the experience of the author in creating software for various research projects over several years. Our intention is to provide very powerful analysis techniques to researchers in a user-friendly environment. We have attempted to create a system that is versatile enough to meet the needs of any project that uses real time observation to gather data. We feel that observation studies are hard enough without the software getting in the way and we hope that MOOSES will make things a bit easier. We also hope that the system will help people to use observation data more efficiently and effectively in answering research questions. Observation data is often under-utilized due to the difficult nature of it's management.

MOOSES is a software system for Windows 95 or higher compatible computers with collection programs avaialable as well for Handheld PCs and PocketPCs. The software allows for collection and analysis of data obtained from observing processes and/or subjects. As events are entered into the computer data files and marked they are automatically marked with entry times.

MOOSES for Windows is currently available from this website in its demo vesion for free. Once you try the demo version, you can license the full version from Vanderbilt's technology transfer office and obtain an activation code to activate the demo version to the fully functional version. Download the latest version here. The windows version is the only one that we currently support.

In addition to collecting data MOOSES can also do several analyses. Results can be in the form of a report or sent directly to a spreadsheet format.

Frequency and Duration - Codes entered into a time based stream of numbers and defined in a related code file are processed into total frequencies and durations for each session. A built in list management system allows for the processing of several sessions at once and for the results to be pooled. The output of this process can be sent to either a text based printable report or to a comma separated file for use with a spreadsheet or statistical analysis package. Durations are calculated within code groups such that each code group represents a mutually exclusive and exhaustive set of values that cover the session time. Up to 48 groups of codes can be defined.

Inter-Observer Agreement - Data from two observers can be compared to give an indication of the degree to which two observers agree on the observation session. Two methods are used: 1) a time window based comparison where matches are tallied using a window of time around the first observer’s events and looking for matches in the second observer’s events. Matches are tallied and the agreement ration is calculated for each event code. 2) a time unit by time unit comparison where each time unit is compared in the two observer’s files and judged to be either in agreement or disagreement. The results are displayed in a square matrix with the agreements down the diagonal of the matrix and the disagreements falling off the diagonal. The resulting agreement ratio and Kappa statistic are provided. As with the frequency and duration analysis, the files can be run in lists and output is provided in either text report format or comma separated value format.

Multi-Group Frequency and Duration - The amount of time that durations overlap across groups is provided by this analysis. For an example in a classroom setting, if you want to know how much aggression occurs in a session when the teacher is present versus when the teacher is absent you would use this analysis to get those numbers. This analysis can also provide a graphical representation of the session time if desired. As with the other analyses, the processing can be done on lists and output provided in a print-able report or as comma separated values.

Sequential Analysis - If you want to know how certain events influence other events in the time stream you would use sequential analysis. For example, if you want to test the theory that prompts from direct care staff influence the self injurious behavior of a particular target subject, you would do a sequential analysis to get the degree to which staff prompts are sequentially associated with instances of self injurious behavior. The index of sequential association that is given is a statistic called “Yule’s Q”. As with the other analyses, the processing can be done on lists and output provided in a print-able report or as comma separated values.

Email or call me (Jon Tapp) (615)322-8086, if you have any questions at all.