In: Statistics and Probability
In an effort to characterize the New Guinea crocodile (Crocodylus novaeguineae), measurements were taken of the dorsal cranial length (mm) (the length of the skull from the tip of the nose to the back of the cranial cap, denoted DCL) and the total length (cm) (denoted TL) of 50 harvested adult males.
DataTL DCL Observation 130 169 1 102 154 2 126 160 3 230 290 4 115 151 5 150 209 6 259 344 7 130 183 8 110 153 9 130 183 10 185 237 11 215 288 12 129 187 13 149 189 14 156 203 15 100 143 16 224 294 17 234 318 18 162 229 19 217 299 20 206 283 21 144 198 22 146 203 23 166 229 24 203 275 25 205 266 26 252 350 27 238 318 28 250 330 29 255 351 30 120 169 31 250 332 32 238 307 33
157 205 34 159 216 35 202 261 36 177 237 37 221 288 38 224 294 39 167 232 40 240 316 41 207 268 42 192 242 43 180 248 44 165 226 45 197 267 46 113 162 47 131 183 48 162 234 49 246 310 50
Question. (8). Is there a significant difference between DCL and TL measurements at the alpha=0.05 level? Report the null and alternative hypothesis. Complete the test and state your conclusion.
We need to test if there are differences between DCL and TL .
Since the same 50 adult males have been tested once for measuring DCL and then TL we need to apply for the Paired t-test here.
So the null hypothesis is H0: µD=0 (no diff) vs H1:µD ≠ 0 (significant diff)
where µD is the mean of the difference.of the two populations.
The test statistic is
t=((√n*dbar)}/sd
where dbar is the mean of the difference.of the two samples.
and sd is the standard deviation of the difference.of the two samples
We first find the difference di=xi-yi
where xi denotes DCL
and yi denotes TL
Table for calculation:
Observations | TL(yi) | DCL (xi) | di=xi-yi |
1 | 102 | 154 | 52 |
2 | 126 | 160 | 34 |
3 | 230 | 290 | 60 |
4 | 115 | 151 | 36 |
5 | 150 | 209 | 59 |
6 | 259 | 344 | 85 |
7 | 130 | 183 | 53 |
8 | 110 | 153 | 43 |
9 | 130 | 183 | 53 |
10 | 185 | 237 | 52 |
11 | 215 | 288 | 73 |
12 | 129 | 187 | 58 |
13 | 149 | 189 | 40 |
14 | 156 | 203 | 47 |
15 | 100 | 143 | 43 |
16 | 224 | 294 | 70 |
17 | 234 | 318 | 84 |
18 | 162 | 229 | 67 |
19 | 217 | 299 | 82 |
20 | 206 | 283 | 77 |
21 | 144 | 198 | 54 |
22 | 146 | 203 | 57 |
23 | 166 | 229 | 63 |
24 | 203 | 275 | 72 |
25 | 205 | 266 | 61 |
26 | 252 | 350 | 98 |
27 | 238 | 318 | 80 |
28 | 250 | 330 | 80 |
29 | 255 | 351 | 96 |
30 | 120 | 169 | 49 |
31 | 250 | 332 | 82 |
32 | 238 | 307 | 69 |
33 | 157 | 205 | 48 |
34 | 159 | 216 | 57 |
35 | 202 | 261 | 59 |
36 | 177 | 237 | 60 |
37 | 221 | 288 | 67 |
38 | 224 | 294 | 70 |
39 | 167 | 232 | 65 |
40 | 240 | 316 | 76 |
41 | 207 | 268 | 61 |
42 | 192 | 242 | 50 |
43 | 180 | 248 | 68 |
44 | 165 | 226 | 61 |
45 | 197 | 267 | 70 |
46 | 113 | 162 | 49 |
47 | 131 | 183 | 52 |
48 | 162 | 234 | 72 |
49 | 246 | 310 | 64 |
50 | 130 | 169 | 39 |
Total | 3117 |
dbar=3117/50=62.34
sd=√{(di-dbar)2/(n-1) =14.7836
n=50
The test statistic is
t=((√n*dbar)}/sd=(√50*62.34)/14.7836=29.8175~t distribution with degrees of freedom =n-1=50-1=49
The p-value for t=29.8175 and degrees of freedom 49 is very small < 0.00001
Hence, p-value is < α=0.05
Thus we reject the null hypothesis at α=0.05 and conclude that there are significant difference between DCL and TL .