2.4 Data Collection and Statistical Analysis of the Data Groups

We originally collected data on a total of 576 “normal” children, using 5 different data collection teams. Participants were trained to be specifically Motor Observation data collectors. Using random numbers, 10 children per class were chosen for the data collection. Two extra numbers were selected for reserves. Parents signed an informed consent. We excluded any child who was in occupational, physical, or speech therapy, or who had physical disabilities, including diabetes from the pool of possible participants. The remaining names were placed in alphabetical order

The teams were:

  • The pilot group- These were our original faculty and post graduate students who developed the items and proceeded to test a minimum of 10 children with the pilot manual.
  • The post-graduate group of collectors- These therapists collected data in various locations in Netherlands.
  • The American data collectors- Testing was done on children in rural Connecticut and Michigan and Columbus, Ohio. Lynn Horowitz, the first author on this project, tested the vast majority of the American children.
  • The NCSI group of data collectors-These graduate students were participating in the Sensory Integration course, during which they submitted two pieces of practice data.
  • The undergraduate Occupational Therapy students of Heerlen’s professional school specifically tested children in the South of the Netherlands.

As it was necessary to determine if each group’s data were statistically similar, we evaluated them by using a one way Anova (for means). The data of the pilot group was excluded from further research immediately. Then differences between the various data collectors groups disappeared when we eliminated the data from the undergraduate student group from Heerlen. These Heerlen under-graduates graded their sample, in general, higher than the other groups graded their’s. Thus they were less able to observe minimal dysfunctions in this “normal” group. Although the Heerlen students had had as much training in scoring techniques as each of the other groups, their general level of professional education appeared to have affected the results. More training was probably necessary but due to the great distance between the organizers and the the students, this unfortunately did not occur. We learned from this situation that more attention should be given to this group when teaching the Motor Observations particularly scoring. Studying with a mentor partner would also be helpful to develop good scoring skills.

Returning to the data, in summary there was no significant difference between the following data collector groups. The data collected for the Motor Observations was therefore valid for both American and Dutch children. The children’s data in these three groups were eligible to be combined for further research.

  • the post- graduate group contributed 40 children
  • the NCSI group- graduate students contributed 81 children
  • the American group had 78 children

The total number of “normal” children was n=199.
The total number of dysfunctional children were n=196. The dysfunction group was composed of only Dutch children.

Kate Norwood tested Intrarater Reliability of the original 74 item test. More information is available in the section, Other Publications.