User: | Open Learning Faculty Member:
The acquisition of the initial data in module 3 was relatively uncomplicated, though I think I could improve the way I collect the data in a few mays. First by refining the categories; variables such as percentage of exposed soil and potentially relevant attributes of other species present were not recorded, and the method I used to categorize the moisture regime was imprecise. I may make changes to my tables to better isolate the relevant variables, and I will likely modify my technique to collect my data to allow for better numbers and accuracy.
The technique I used was to overlay a map of my chosen area with a grid of squares, each measuring 20×20, and designating each square as either disturbed or undisturbed. I selected one of each in close proximity and I divided each of these squares into a grid of 400 1x1m squares from which 3 were randomly selected in the disturbed square and 2 in the undisturbed square. These 5 plots were my samples, and I recorded the light exposure, moisture regime, and the other plant species present as my predictor variables, and the number and the average height of rose bushes present as my response variables. Because undisturbed squares outnumber disturbed squares, by selecting an equal number of each to collect an equal number of samples from, I hope to reduce the number of samples necessary to see relevant trends in the data.
Changes I will make to this technique will include the number of samples I record and the way I select them. I will still use randomization, but instead of recording observations from 2 and 3 randomly selected replicates per square, I will likely record 2 transects of 20 plots per 20x20m square, providing more data and an equal number of observations from disturbed and undisturbed areas. 
