Missing Values for two Variables: stata

[Bezi Nyem]

Hi Everyone,

I hope you are well.

I am running panel data for 12 countries over a 13 year time period. I have % control variables but within two of them (Edu and Health).
I have missing values. I don't want to drop these control variables, How do I fill the missing values?

I have provided the data for the two variables just in case its important to see it. It is currently in percentage

Thank you in advance for your time and help!

A	year	Health	Edu
Botswana	1991
Botswana	1992		31.11832
Botswana	1993		39.76681
Botswana	1994	6
Botswana	1995	14.2	62.46729
Botswana	1996	22.2	73.51929
Botswana	1997	25.8
Botswana	1998	25.7	78.52549
Botswana	1999	24.6	0
Botswana	2000	23.9	-78.5255
Botswana	2001	23.5
Botswana	2002	23.3	82.76215
Botswana	2003	23.1	17.64272
Cameroon	1991		20.20845
Cameroon	1992		23.81009
Cameroon	1993		25.45207
Cameroon	1994	2
Cameroon	1995	3.5
Cameroon	1996	4.7	25.94187
Cameroon	1997	5.3	27.88825
Cameroon	1998	5.4	26.64419
Cameroon	1999	5.2	35.94419
Cameroon	2000	4.9
Cameroon	2001	4.8	47.26565
Cameroon	2002	4.8	50.46743
Cameroon	2003	4.7	19.35992
Cote d'Ivoire	1991		18.72056
Cote d'Ivoire	1992		18.4261
Cote d'Ivoire	1993
Cote d'Ivoire	1994	3
Cote d'Ivoire	1995	4.3
Cote d'Ivoire	1996	5.4	24.28939
Cote d'Ivoire	1997	5.7
Cote d'Ivoire	1998	5.3
Cote d'Ivoire	1999	4.6
Cote d'Ivoire	2000	4.1
Cote d'Ivoire	2001	3.8
Cote d'Ivoire	2002	3.7
Cote d'Ivoire	2003	3.6	39.88876
Ghana	1991		38.99115
Ghana	1992		37.37887
Ghana	1993		35.94744
Ghana	1994	1.5
Ghana	1995	2.2
Ghana	1996	2.7	40.73941
Ghana	1997	2.9	41.02408
Ghana	1998	2.7	47.34028
Ghana	1999	2.4	55.77963
Ghana	2000	2.1
Ghana	2001	1.9	57.08299
Ghana	2002	1.9	58.13869
Ghana	2003	1.8	29.32197
Kenya	1991		31.97818
Kenya	1992		41.12302
Kenya	1993
Kenya	1994	6.4
Kenya	1995	10.9
Kenya	1996	12.8	38.52061
Kenya	1997	11.6	40.99545
Kenya	1998	9	47.8988
Kenya	1999	7	59.40621
Kenya	2000	6.2
Kenya	2001	6.1
Kenya	2002	6	67.64039
Kenya	2003	6
Madagascar	1991		34.07448
Madagascar	1992
Madagascar	1993		18.26771
Madagascar	1994	0.1
Madagascar	1995	0.1
Madagascar	1996	0.3
Madagascar	1997	0.5
Madagascar	1998	0.6	21.14828
Madagascar	1999	0.5	29.04219
Madagascar	2000	0.4
Madagascar	2001	0.4	36.59524
Madagascar	2002	0.3	38.02759
Madagascar	2003	0.3	15.99282
Malawi	1991		17.13872
Malawi	1992		16.71433
Malawi	1993		16.09533
Malawi	1994	8.8	18.06774
Malawi	1995	13.3	24.65532
Malawi	1996	16	36.01715
Malawi	1997	16.7	29.89747
Malawi	1998	15.8	27.28505
Malawi	1999	13.9	31.21402
Malawi	2000	12.1	33.05596
Malawi	2001	11.2	34.18793
Malawi	2002	10.8	34.44938
Malawi	2003	10.3	48.15252
Mauritius	1991		45.38366
Mauritius	1992		45.59499
Mauritius	1993		52.40914
Mauritius	1994	0.1
Mauritius	1995	0.1	66.75667
Mauritius	1996	0.4	74.98109
Mauritius	1997	0.9	81.02567
Mauritius	1998	1.1	88.65556
Mauritius	1999	1.2	88.13017
Mauritius	2000	1.2	89.24696
Mauritius	2001	1.1	90.41113
Mauritius	2002	1	91.84192
Mauritius	2003	1	4.48537
Niger	1991		4.87009
Niger	1992
Niger	1993		6.24449
Niger	1994	0.2	6.41819
Niger	1995	0.5	6.52356
Niger	1996	0.9	7.03912
Niger	1997	1.2	6.862
Niger	1998	1.2	9.90822
Niger	1999	1.1	11.27569
Niger	2000	0.8	13.51845
Niger	2001	0.7	14.35081
Niger	2002	0.6	15.56878
Niger	2003	0.6	17.00856
Nigeria	1991		28.68492
Nigeria	1992		27.07259
Nigeria	1993		24.59582
Nigeria	1994	1.4
Nigeria	1995	2.2
Nigeria	1996	3.2	23.41556
Nigeria	1997	3.8	29.42101
Nigeria	1998	4	34.69912
Nigeria	1999	3.8	35.09796
Nigeria	2000	3.6	43.83671
Nigeria	2001	3.5
Nigeria	2002	3.4
Nigeria	2003	3.3
South Africa	1991
South Africa	1992
South Africa	1993		66.07011
South Africa	1994	1
South Africa	1995	3.9
South Africa	1996	10.3	90.76656
South Africa	1997	17.1	87.15127
South Africa	1998	20.6	88.91318
South Africa	1999	20.8	89.49576
South Africa	2000	19.5	90.25352
South Africa	2001	18.9	91.05531
South Africa	2002	18.8	91.95773
South Africa	2003	18.8
Swaziland	1991
Swaziland	1992
Swaziland	1993
Swaziland	1994	2.2	46.67545
Swaziland	1995	9.4	47.33687
Swaziland	1996	18.5	44.48612
Swaziland	1997	24	42.11219
Swaziland	1998	25.6	46.62029
Swaziland	1999	25.7
Swaziland	2000	26.4	57.97358
Swaziland	2001	27.2	59.91126
Swaziland	2002	27.6	60.67129
Swaziland	2003	28

[Carlo Lazzaro]

Bezi:
- you can consider using -ipolate-;
-you may want to let things as they are and end up with an unbalanced panel dataset (that Stata can handle without any problem);
- you may wnat to consider -mi-.

All the usual remarks about missingness mechanisms (i.e.: is the missingness informative or not) apply.

[Nick Cox]

See also http://www.statalist.org/forums/foru...-interpolation

[Bezi Nyem]

See also http://www.statalist.org/forums/foru...-interpolation

Nick, thank you for the link.

I just have one more question. Do i go about using mipolate as specified in the link below:
http://www.statalist.org/forums/foru...tions-in-panel

using the following command
ipolate Edu year, gen (Edu1)?

I have used the above command and I do get new figures but stata also changes existing figure

mipolate Edu year, gen (Edu1) (Apologies I don't know how to copy the table directly from Stata yet) I had to move it to excel first.

Edu	Edu1
20.28768	23.57115
24.83205	26.48823
31.11832	30.15479
39.76681	31.64988
51.41528	30.64417
	36.31811
73.51929	43.61056
75.65077	46.20282
78.52549	47.05804
	48.3762
	54.64753
	53.85754
82.76215	59.15254
17.64272	23.57115
20.20845	26.48823
23.81009	30.15479
25.45207	31.64988
	30.64417
	36.31811
25.94187	43.61056
27.88825	46.20282
26.64419	47.05804
35.94419	48.3762
	54.64753
47.26565	53.85754
50.46743	59.15254
19.35992	23.57115
18.72056	26.48823
18.4261	30.15479
	31.64988
	30.64417
	36.31811
24.28939	43.61056
	46.20282
	47.05804
	48.3762
	54.64753
	53.85754
	59.15254
39.88876	23.57115
38.99115	26.48823
37.37887	30.15479
35.94744	31.64988
	30.64417
	36.31811
40.73941	43.61056
41.02408	46.20282
47.34028	47.05804
55.77963	48.3762
	54.64753
57.08299	53.85754
58.13869	59.15254
29.32197	23.57115
31.97818	26.48823
41.12302	30.15479
	31.64988
	30.64417
	36.31811
38.52061	43.61056
40.99545	46.20282
47.8988	47.05804
59.40621	48.3762
	54.64753
	53.85754
67.64039	59.15254
	23.57115
34.07448	26.48823
	30.15479
18.26771	31.64988
	30.64417
	36.31811
	43.61056
	46.20282
21.14828	47.05804
29.04219	48.3762
	54.64753
36.59524	53.85754
38.02759	59.15254
15.99282	23.57115
17.13872	26.48823
16.71433	30.15479
16.09533	31.64988
18.06774	30.64417
24.65532	36.31811
36.01715	43.61056
29.89747	46.20282
27.28505	47.05804
31.21402	48.3762
33.05596	54.64753
34.18793	53.85754
34.44938	59.15254
48.15252	23.57115
45.38366	26.48823
45.59499	30.15479
52.40914	31.64988
	30.64417
66.75667	36.31811
74.98109	43.61056
81.02567	46.20282
88.65556	47.05804
88.13017	48.3762
89.24696	54.64753
90.41113	53.85754
91.84192	59.15254
4.48537	23.57115
4.87009	26.48823
	30.15479
6.24449	31.64988
6.41819	30.64417
6.52356	36.31811
7.03912	43.61056
6.862	46.20282
9.90822	47.05804
11.27569	48.3762
13.51845	54.64753
14.35081	53.85754
15.56878	59.15254
17.00856	23.57115
28.68492	26.48823
27.07259	30.15479
24.59582	31.64988
	30.64417
	36.31811
23.41556	43.61056
29.42101	46.20282
34.69912	47.05804
35.09796	48.3762
43.83671	54.64753
	53.85754
	59.15254
	23.57115
	26.48823
	30.15479
66.07011	31.64988
	30.64417
	36.31811
90.76656	43.61056
87.15127	46.20282
88.91318	47.05804
89.49576	48.3762
90.25352	54.64753
91.05531	53.85754
91.95773	59.15254
	23.57115
	26.48823
	30.15479
	31.64988
46.67545	30.64417
47.33687	36.31811
44.48612	43.61056
42.11219	46.20282
46.62029	47.05804
	48.3762
57.97358	54.64753
59.91126	53.85754
60.67129	59.15254

[Bezi Nyem]

Bezi:
- you can consider using -ipolate-;
-you may want to let things as they are and end up with an unbalanced panel dataset (that Stata can handle without any problem);
- you may wnat to consider -mi-.

All the usual remarks about missingness mechanisms (i.e.: is the missingness informative or not) apply.

Dear Carlo,

Thanks for your response. My worry is that if I let things as they are, my sample size might reduce once I apply first differences no?
Also I was unable to perform a unit root test on the variables with missing values.

[Carlo Lazzaro]

Bezi:
you're right as far as the first point is concerned (you are probably right about the second one, too, but I'm not familiar with Sir Clyve Granger's machinery).
Can't say whether in your case sample shrinkage is a limited or unaffordable damage (it boils down to how many observations will be listwise deleted letting things as they are).

[Bezi Nyem]

Nick and Carlo, thank you. for your help, I figured out my error and corrected it. Thank you both!