We randomly assigned 9 horses to 3 groups and put those groups through 3 conditions (Random assignment of subjects by chance is done so that a researcher cant “cherry pick” the best subjects)
Condition 1: Original Saddle with Original Pad interface
Condition 2 : Original Saddle with New pads/shim interface
Condition 3: New Saddle with New Pads/Shim interface.
Score sheets allowed for scoring 6 symptoms/signs of discomfort that were taken from the literature on signs of pain in horses. Those that were most easily observable and that explained the greatest amount of variance were chosen to optimize inter-rater reliability (we wanted all our scorers to be seeing the same thing independently of one another)
3 riders rode all the horses through all 3 conditions.
We pre-selected our primary end point at .05 A very high level of statistical significance The p value is the most widely used statistical concept in biomedical research. However, there is now some controversy in relying on a simple p variable to judge scientific merit of a study.
Our statistical analysis is below: We did a one-way ANOVA (measuring variance) and Independent t-tests comparing groups.
For the difference between condition 1 and condition 3, The t-value is -4.686819. The value of p is .000011. The result is significant at p < .05.
For the difference between Condition 1 and Condition 2 The t-value is 2.06057. The p-value is .020482. The result is significant at p < .10.
This difference did not meet the highest level of significance but was found to be statistically significant at p<.10
Sometimes you can make observations that seem to clearly show a difference but when put under the rigor of Scientific measure do not statistically support the observations.
Results of the one-way ANOVA
∑X 137 88 47 272
Mean 1.6707 1.0732 0.5802 1.11
∑X2 455 302 109 866
Std.Dev. 1.6708 1.6008 1.0107 1.5204
Source SS df MS
Between-treatments 48.6254 2 24.3127 F = 11.41575
Within-treatments 515.3991 242 2.1297
Total 564.0245 244
The f-ratio value is 11.41575. The p-value is .000018. The result is significant at p < .05.