egenmore stata or egen

[Zara Ruzim]

Good day everyone

I was kindly requesting assistance. I correlated yield and DEM variables and I have around 15000 observations. I would like to find less influential variables that affect the low correlation.

I’ve tried different methods that I’ve seen on this forum but I think I am not doing something wrong as I get error messages.
I will appreciate it if someone corrects my code to me.

egenmore corr1= corr(yield, DEM) , by(fid)

Error: command egenmore is unrecognized

FID	yield	DEM
0	4.22505	3.613554
1	5.7904	3.639574
2	5.98815	3.735786
3	5.877675	3.751731
4	6.65555	3.732317
5	5.9603	3.742209
6	6.2362	3.7284
7	6.5016	3.793361
8	6.66445	3.801543
9	6.268925	3.847231
10	6.56635	3.818495
11	6.302375	3.8284
12	6.3464	3.842584
13	6.882575	3.847168
14	7.026975	3.928738
15	6.927975	3.931657
16	5.99775	3.935923
17	5.9769	3.955251
18	6.085325	3.938371
19	6.1305	3.95141
20	6.4169	3.970702
21	6.39635	3.998399
22	6.08975	4.016246
23	6.349175	4.007195
24	5.91035	4.012233
25	5.677625	4.015934
26	4.93885	4.016912
27	5.1393	3.996199
28	5.041175	3.983388
29	5.416575	3.986725
30	5.409425	4.020954
31	5.21225	4.04499
32	5.399975	4.047986
33	4.668625	4.067715
34	4.919425	4.067241
35	4.54245	4.076788
36	4.36695	4.077446
37	4.40215	4.050743
38	4.654175	4.05063
39	4.968975	4.045498
40	5.2284	4.109359
41	4.456975	4.108246
42	3.74855	4.047297
43	4.336725	3.994994
44	4.17595	4.068776
45	5.24705	4.134148
46	5.2088	4.16378
47	6.084	4.1645
48	5.6427	4.120104
49	5.4703	4.098624
50	5.43955	4.058217
51	5.20975	3.977072
52	4.547875	3.980149
53	4.1893	3.904951
54	4.258975	3.833583
55	4.39275	3.840139
56	4.83865	3.811399
57	4.3447	3.758291
58	3.641175	3.681761
59	3.734675	3.631509
60	3.552325	3.609142
61	3.280925	3.507009
62	3.402725	3.416963
63	4.019625	3.326922
64	4.907875	3.28381
65	4.578975	3.235068
66	4.171525	3.168867
67	4.251475	3.144556
68	4.167225	3.066346
69	3.9188	3.059794
70	3.591975	3.016265
71	1.768	2.791127
72	1.3077	2.545597
73	1.27515	2.527058
74	1.826725	2.763952
75	2.004775	2.842975
76	3.1396	2.81416
77	2.1865	2.748117
78	5.3409	2.949697
79	3.8303	2.90294
80	6.210025	3.081014
81	5.589075	3.178157
82	6.31075	3.201829
83	6.048125	3.231142
84	5.547775	3.248579
85	5.40915	3.267068
86	5.71085	3.320058
87	5.821	3.422306
88	6.095875	3.488281
89	5.88745	3.547491
90	5.743325	3.578636
91	6.237175	3.621943
92	5.8298	3.628169
93	4.498625	3.6117
94	3.859825	3.625234
95	4.5237	3.632873
96	4.1156	3.659969
97	5.899825	3.593693
98	3.507225	3.621556
99	4.2924	3.602226
100	7.3091	3.771317

[Jesse Wursten]

What error messages do you get?

[Zara Ruzim]

Error: command egenmore is unrecognized

[Ali Atia]

-egenmore- is a user-written command. To use it, you must install it by typing -ssc install egenmore-.

[Joro Kolev]

You need to install the user written collection egenmore

 ssc install egenmore

and then the command you have in mind is

egen corr1= corr(yield, DEM) , by(fid)

Error: command egenmore is unrecognized

[Rich Goldstein]

-egenmore- is not a command; it is a collection of -egen- functions (more later); second, -egenmore- is users-written and must be downloaded and installed; use -search- or -findit- to find and install; then use the functions within just as you would use any other -egen- function

[Zara Ruzim]

I already installed
. ssc install egenmore
checking egenmore consistency and verifying not already installed...
all files already exist and are up to date.

[Ali Atia]

The command to use is

 
 egen corr1= corr(yield, DEM) , by(fid)

[Nick Cox]

This is a comedy of small errors. The help for egenmore (once installed from SSC) does explain that you call up any of its functions in the usual way with egen, as Rich Goldstein also explains.


However, the function corr() won't accept a comma, and it does require prior sorting. Hence this will work

sysuse auto
bysort rep78 : egen corr = corr(mpg weight)

and so I imagine that this will work:

bysort fid : egen corr  = corr(yield DEM)

[Clyde Schechter]

I feel compelled to note, however, that if the example data shown in #1 is representative of the entire data set, the results will just be all missing values because fid identifies unique observations. And the correlation between two variables is undefined when there is only one pair observed.

[Nick Cox]

#10 Clyde Schechter Well, there you go. I failed to spot that too.

But here's the sample. At a wild guess, yield is some crop yield and DEM (Digital Elevation Model???) is surface height above some datum. Small variation in height doesn't rule out controls through moisture and nutrient status and even erosion and deposition.