2. After the analysis is complete, find the value for the probability tested (it is the “p-value” in the row labeled “Columns” in the ANOVA source/summary table). If this value is below .05, there are systematic differences among the raters.

3. Run the “ttests” macro. Follow the on-screen prompts. Examine the results to determine which raters are significantly higher or lower than the group mean. Then, use this same information to change the colors of the bars on the graph to indicate these raters. Next, make sure that the horizontal axis of the graph starts with “1”, and ends with the appropriate value (“3”
for observable behaviors, “4” for task/skill). Print the t-test graph on a single page. It may be helpful to increase the font size of the title and the legend, so that it is easier to read.

Congruency Analyses
1. Run the “Congruency” macro. Follow the on-screen prompts. Examine the vertical axes on both sides of each graph. Note the maximum value (on either side). For each graph, change the left and right vertical axes, so that they are all on the same scale. Print each congruency graph on a single page. (see next page for example)

2. Lastly, print the block of congruency results, sized to 1 page.

Consistency Analyses
1. From the “Tools” menu, choose “Data Analysis”. Choose the “Correlation” option and follow the on-screen prompts. For the “Input Range” highlight the instructor ID labels and the raw data. Be sure to indicate that the data is grouped by columns and that the labels are located in the first row.

2. Run the “Consistency” macro. Follow the on-screen prompts. You will use the output from the “Correlation” command (previous step) as your input for the consistency macro. You will be asked to enter a benchmark value in two different steps, one for RATERS and one for the GROUP, in both cases GMU recommends using zero. Consistency indexes range from –1 (perfect negative correlation: individual goes up, group goes down) to +1 (perfect positive correlation: individual goes up, group goes up – or – individual goes down, group goes down) with zero indicating no relationship between individual and group ratings. A benchmark of zero will indicate those raters whose individual ratings have no significant relationship to the groups’ ratings as being “NS”.
DO NOT ask for the graphs in this macro. Click “cancel” to stop the “Consistency” macro after you enter the GROUP benchmark, when you come to this prompt:
3. Print the overall block of consistency results, sized to 1 page.

4. Run the “Profiles” macro. Follow the on-screen prompts. Examine each IE’s profile graph. If the vertical axis does not begin with “1” and end with the appropriate maximum value (“3” for observable behaviors, “4” for task/skill), change the scale of the vertical axis. Print each IE’s graph on a single page.
　
中國航空網 www.k6050.com
航空翻譯 www.aviation.cn
本文鏈接地址：Developing Advanced Crew Resource Management (ACRM) Training: A Training Manual(106)