Hello, I have a simple case of survey data and am puzzled why svy: mean, over gives me different results than a bivariate regression.
Code:
 svy, subpop(if !inlist(nutr_no,2,7,23) & age_sex_grp!=1): mean eardiff, over(nutr_no)
gives me this:

svy, subpop(if !inlist(nutr_no,2,7,23) & age_sex_grp!=1): mean eardiff, over(nutr_no)
(running mean on estimation sample)

Survey: Mean estimation

Number of strata = 1 Number of obs = 346,623
Number of PSUs = 102 Population size = 404,507,239
Subpop. no. obs = 253,020
Subpop. size = 343,310,430
Design df = 101

Carbohydrate: nutr_no = Carbohydrate
Protein: nutr_no = Protein
Lipids: nutr_no = Lipids
Vitamin_A: nutr_no = Vitamin_A
Vitamin_C: nutr_no = Vitamin_C
Vitamin_E: nutr_no = Vitamin_E
Thiamin: nutr_no = Thiamin
Riboflavin: nutr_no = Riboflavin
Niacin: nutr_no = Niacin
Vitamin_B6: nutr_no = Vitamin_B6
Folate: nutr_no = Folate
Vitamin_B12: nutr_no = Vitamin_B12
Calcium: nutr_no = Calcium
Copper: nutr_no = Copper
Iron: nutr_no = Iron
Magnesium: nutr_no = Magnesium
Phosphorus: nutr_no = Phosphorus
Selenium: nutr_no = Selenium
Zinc: nutr_no = Zinc

--------------------------------------------------------------
| Linearized
Over | Mean Std. Err. [95% Conf. Interval]
-------------+------------------------------------------------
eardiff |
Carbohydrate | 4.941668 .5116118 3.926768 5.956568
Protein | 12.67411 .6648757 11.35517 13.99304
Lipids | 24.04572 .7124884 22.63233 25.4591
Vitamin_A | 34.32649 .9846294 32.37325 36.27973
Vitamin_C | 67.61416 1.6064 64.42749 70.80082
Vitamin_E | 37.09219 .7086652 35.68639 38.49799
Thiamin | 29.00894 .658244 27.70316 30.31472
Riboflavin | 26.34825 .6733115 25.01258 27.68392
Niacin | 27.31367 .6336908 26.05659 28.57074
Vitamin_B6 | 49.28489 .9920848 47.31686 51.25292
Folate | 39.16276 .7442579 37.68635 40.63917
Vitamin_B12 | 38.55372 .7172662 37.13086 39.97659
Calcium | 48.0474 .9347514 46.1931 49.90169
Copper | 41.01134 .7896239 39.44494 42.57775
Iron | 154.65 1.524378 151.626 157.6739
Magnesium | 41.67131 .8602638 39.96478 43.37785
Phosphorus | 75.87354 1.374524 73.14686 78.60023
Selenium | 38.47203 .7368524 37.01031 39.93375
Zinc | 104.1346 1.137811 101.8775 106.3917
--------------------------------------------------------------

But then when I use regress instead, I get different results. What is the difference in the estimation method?
Code:
 svy, subpop(if !inlist(nutr_no,2,7,23) & age_sex_grp!=1): reg eardiff i.nutr_no
svy, subpop(if !inlist(nutr_no,2,7,23) & age_sex_grp!=1): reg eardiff i.nutr_no
(running regress on estimation sample)

Survey: Linear regression

Number of strata = 1 Number of obs = 346,623
Number of PSUs = 102 Population size = 404,507,239
Subpop. no. obs = 253,020
Subpop. size = 343,310,430
Design df = 101
F( 18, 84) = 2016.08
Prob > F = 0.0000
R-squared = 0.2240

------------------------------------------------------------------------------
| Linearized
eardiff | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
nutr_no |
Protein | 7.73244 .3295289 23.47 0.000 7.078743 8.386136
Lipids | 19.10405 .3606008 52.98 0.000 18.38871 19.81938
Vitamin_A | 29.38482 .7260651 40.47 0.000 27.9445 30.82514
Vitamin_C | 62.67249 1.407138 44.54 0.000 59.88111 65.46387
Vitamin_E | 32.15052 .3365246 95.54 0.000 31.48295 32.8181
Thiamin | 24.06727 .2778912 86.61 0.000 23.51601 24.61854
Riboflavin | 21.40658 .3287791 65.11 0.000 20.75437 22.05879
Niacin | 22.372 .2617531 85.47 0.000 21.85275 22.89125
Vitamin_B6 | 44.34322 .745232 59.50 0.000 42.86488 45.82156
Folate | 34.22109 .3820805 89.57 0.000 33.46315 34.97903
Vitamin_B12 | 33.61205 .3416303 98.39 0.000 32.93435 34.28976
Calcium | 43.10573 .6527606 66.04 0.000 41.81083 44.40063
Copper | 36.06968 .4424423 81.52 0.000 35.19199 36.94736
Iron | 149.7083 1.152961 129.85 0.000 147.4212 151.9955
Magnesium | 36.72965 .6259886 58.67 0.000 35.48785 37.97144
Phosphorus | 70.93188 1.177134 60.26 0.000 68.59676 73.26699
Selenium | 33.53036 .3579034 93.69 0.000 32.82038 34.24034
Zinc | 99.19295 .7791701 127.31 0.000 97.64729 100.7386
|
_cons | 4.941668 .5116118 9.66 0.000 3.926768 5.956568
------------------------------------------------------------------------------


Thanks for the help!!
-Kate