Easier way to get casewise statistics using tabstat?

[Amanda Ode]

Hi,

I need a command/code that help me to get casewise descriptive statistics using "tabstat" without needing to match variables with the same number of observations manually.

Example:
I have the following variables and number of observations in my data set:

Variable	Obs
Weight-a	100
Weight-b	90
BMI-a	120
BMI-b	110
Hight-a	70
Hight-b	72

Right now, I type "tabstat Weight-a Weight-b, casewise stat(mean sd min max)" for the Weight variables and a similar one for the BMI variables etc.

What I actually need is to compare Weight-a to Weight-b (90 pairwise obs), BMI-a to BMI-b (110 pairwise obs), and Hight-a to Hight-b (70 pairwise obs), respectively, using one code with the "tabstat" command.

I heard about "loop" and "return list" commands for instance but I don't know how to use them.

[Mike Lacy]

To clarify what you want: It sounds like you want a listing in which variables are the rows, and rows are sorted so that variables with the same number of non-missing observations appear consecutively.
If you want to learn about the return list, -help return list- will give some information.

[David Benson]

Hi Amanda,

1) It's not quite clear to me what you want, but here is some sample data and commands to help you get going
2) Do you want the avg change between weight_a and weight_b, the percent change, or the correlation?
3) It would be *really* helpful if you could share a sample of your data using dataex. If you need help with dataex, I have a video tutorial here

* Example generated by -dataex-. To install: ssc install dataex
clear
input byte id int(weight_a weight_b) float(BMI_a BMI_b) byte(height_a height_b)
 1 155 150 29.28 25.74 61 64
 2 174 195 29.86 30.54 64 67
 3 190 200 31.61 30.41 65 68
 4 158 170 24.74 24.39 67 70
 5 183 177 27.82 26.91 68 68
 6 184 194 27.17 28.65 69 69
 7 191 184  27.4     . 70  .
 8 205 199 27.04     . 73  .
 9 210 203 26.25 25.37 75 75
10 235   . 27.86     . 77 77
end

------------------ copy up to and including the previous line ------------------

gen weight_change = weight_b - weight_a
gen weight_pct = (weight_b - weight_a ) / weight_a
gen BMI_change = BMI_b - BMI_a


. list id weight_a weight_b weight_ch weight_p BMI_a BMI_b BMI_change height_a height_b , noobs N( weight_a wei
> ght_b weight_ch weight_p BMI_a BMI_b BMI_change height_a height_b ) mean( weight_a weight_b weight_ch weight_
> p BMI_a BMI_b BMI_change height_a height_b )

      +----------------------------------------------------------------------------------------------------+
      | id   weight_a   weight_b   weight~h   weight_~t    BMI_a   BMI_b   BMI_cha~e   height_a   height_b |
      |----------------------------------------------------------------------------------------------------|
      |  1        155        150         -5   -.0322581    29.28   25.74   -3.540001         61         64 |
      |  2        174        195         21    .1206897    29.86   30.54    .6800003         64         67 |
      |  3        190        200         10    .0526316    31.61   30.41   -1.200001         65         68 |
      |  4        158        170         12    .0759494    24.74   24.39   -.3500004         67         70 |
      |  5        183        177         -6   -.0327869    27.82   26.91   -.9099998         68         68 |
      |----------------------------------------------------------------------------------------------------|
      |  6        184        194         10    .0543478    27.17   28.65        1.48         69         69 |
      |  7        191        184         -7   -.0366492     27.4       .           .         70          . |
      |  8        205        199         -6   -.0292683    27.04       .           .         73          . |
      |  9        210        203         -7   -.0333333    26.25   25.37   -.8799992         75         75 |
      | 10        235          .          .           .    27.86       .           .         77         77 |
      |----------------------------------------------------------------------------------------------------|
 Mean |         188.5    185.778   2.444444    .0154803   27.903   27.43   -.6742859       68.9      69.75 |
    N |            10          9          9           9       10       7           7         10          8 |
      +----------------------------------------------------------------------------------------------------+


. tabstat weight_a weight_b weight_ch weight_pct BMI_a BMI_b BMI_change height_a height_b , stats(n mean median
>  min max) col(stat)

    variable |         N      mean       p50       min       max
-------------+--------------------------------------------------
    weight_a |        10     188.5       187       155       235
    weight_b |         9  185.7778       194       150       203
   weight_ch |         9  2.444444        -5        -7        21
  weight_pct |         9  .0154803 -.0292683 -.0366492  .1206897
       BMI_a |        10    27.903     27.61     24.74     31.61
       BMI_b |         7     27.43     26.91     24.39     30.54
  BMI_change |         7 -.6742859 -.8799992 -3.540001      1.48
    height_a |        10      68.9      68.5        61        77
    height_b |         8     69.75      68.5        64        77
----------------------------------------------------------------

. summ weight_a weight_b weight_ch weight_pct BMI_a BMI_b BMI_change height_a height_b

    Variable |        Obs        Mean    Std. Dev.       Min        Max
-------------+---------------------------------------------------------
    weight_a |         10       188.5    24.07973        155        235
    weight_b |          9    185.7778    17.44834        150        203
   weight_ch |          9    2.444444    10.76001         -7         21
  weight_pct |          9    .0154803    .0605473  -.0366492   .1206897
       BMI_a |         10      27.903    1.936452      24.74      31.61
-------------+---------------------------------------------------------
       BMI_b |          7       27.43    2.471592      24.39      30.54
  BMI_change |          7   -.6742859    1.589904  -3.540001       1.48
    height_a |         10        68.9    5.021067         61         77
    height_b |          8       69.75     4.26782         64         77

[Amanda Ode]

Thank you for your answer David even if it is not what I really asked about. I think Mike's question is closer to what I really meant. I will post the question again with the same title and hopefully with better explanation.