Q1_1_1:Normality
Q1_1_2:Boxplots of the variables
Q1_1_3:Homogeneity of variance
Q1_1_4:Spread of data around lines of best fit on scatterplots, residual plots
Q1_1_5:Linearity
Q1_1_6:Scatterplots, lowess smoothers
Q1_1_7:Colinearity
Q1_1_8:Scatterplot matrices, correlation matrices
Q1_1_9:Independent observations
Q1_1_10:Random sampling, unbiased
Q1_2a:
Q1_2b:Y
Q1_2c:Y
Q1_2d:LAT and MAT, LONG and DJFMAP, LONG and MAP, DJFMAP and JJAMAP. Therefore these combinations cannot be in the same linear model.
Q1_4a:sqrtC3 = CONSTANT + LAT + LONG + LAT
Q1_5a:0.4282658
Q1_5b:18.275
Q1_5c:<0.001
Q1_5d:0.0436937
Q1_5e:8.977
Q1_5f:<0.001
Q1_5g:-0.0028773
Q1_5h:-0.781
Q1_5i:0.4375
Q1_5j:0.0022824
Q1_5k:3.055
Q1_5l:0.0032
Q2_1_1:Normality
Q2_1_2:Boxplots of variables
Q2_1_3:Homogeneity of variance
Q2_1_4:Spread of data around lines of best fit on scatterplots, residual plots
Q2_1_5:Linearity
Q2_1_6:Scatterplots, lowess smoothers
Q2_1_7:Collinearity
Q2_1_8:Scatterplot matrices, correlation matrices
Q2_1_9:Independent observations
Q2_1_10:Random sampling, unbiased
Q2_2a:Y
Q2_2b:Y
Q2_2c:Y
Q2_2d:There are moderate correlations between logDIST and logLDIST, logAREA and GRAZE and GRAZE and YR.ISOL, however these might not be too bad.
Q2_3b:logAREA was significantly different from 0.
Q2_5a:22.74937
Q2_5b:19.755
Q2_5c:<0.001
Q2_5d:8.12815
Q2_5e:5.277
Q2_5f:<0.001
Q2_5g:-2.97929
Q2_5h:-3.560
Q2_5i:0.000848
Q2_5j:-0.03195
Q2_5k:-0.565
Q2_5l:0.574494
Q2_5m:2.92566
Q2_5n:3.141
Q2_5o:0.002884
Q2_5p:0.17281
Q2_5q:2.748
Q2_5r:0.008424
Q2_5s:0.10148
Q2_5t:2.901
Q2_5u:0.005594
Q2_5v:0.01116
Q2_5w:0.329
Q2_5x:0.743248
Q1_2a:Y
Q3_3a:0.8584
Q3_3b:3.100
Q3_3c:0.00505
Q3_3d:0.3336
Q3_3e:9.532
Q3_3f:<0.001