Dear Stata users,

I have the following panel data set covering all 50 states over 13 years.

Code:
. describe
  obs:       452,755                          
 vars:             4                          3 Mar 2020 23:39
 size:     7,244,080                          
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
              storage   display    value
variable name   type    format     label      variable label
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
year                   long    %12.0g                CCES year
state                  long    %12.0g     state      State (FIPS)
st                     long    %12.0g     st         State abbreviation (FIPS)
approval_gov           long    %12.0g     approval_gov  Governor approval
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Sorted by: 



. tab approval_gov,m

               Governor approval |      Freq.     Percent        Cum.
---------------------------------+-----------------------------------
                Strongly Approve |     67,404       14.89       14.89
      Approve / Somewhat Approve |    139,940       30.91       45.80
Disapprove / Somewhat Disapprove |     84,271       18.61       64.41
             Strongly Disapprove |    117,282       25.90       90.31
          Never Heard / Not Sure |     40,318        8.91       99.22
  Neither Approve Nor Disapprove |      1,414        0.31       99.53
                               . |      2,126        0.47      100.00
---------------------------------+-----------------------------------
                           Total |    452,755      100.00

. tab approval_gov,m nolab

   Governor |
   approval |      Freq.     Percent        Cum.
------------+-----------------------------------
          1 |     67,404       14.89       14.89
          2 |    139,940       30.91       45.80
          3 |     84,271       18.61       64.41
          4 |    117,282       25.90       90.31
          5 |     40,318        8.91       99.22
          6 |      1,414        0.31       99.53
          . |      2,126        0.47      100.00
------------+-----------------------------------
      Total |    452,755      100.00



.    tab year

    CCES year    Freq.    Percent    Cum.
                
    2006    36,421    8.04    8.04
    2007    9,999    2.21    10.25
    2008    32,800    7.24    17.50
    2009    13,800    3.05    20.55
    2010    55,400    12.24    32.78
    2011    20,150    4.45    37.23
    2012    54,535    12.05    49.28
    2013    16,400    3.62    52.90
    2014    56,200    12.41    65.31
    2015    14,250    3.15    68.46
    2016    64,600    14.27    82.73
    2017    18,200    4.02    86.75
    2018    60,000    13.25    100.00
                
    Total    452,755    100.00
The dataset is 4 variable wide and 452,755 observations long because each state year contains individual responses to a survey question about gubernatorial approval. Thus, there are 495 observations for Alabama in 2006, 131 obs for Alaska in 2006 etc. I would like to reshape the dataset and create 6 variables " Strongly_Approve", "Approve", "Disapprove", "Strongly_Disapprove", "Never_Heard", "Neither_Approve_Nor_Disapprove" instead of "approval_gov". Each of these 6 new variables would count the number of occurrences of each type of response in a given state year so that my dataset is 650 observations long (50 states by 13 years). There are about 2,000 missing values in approval_gov, which can be ignored/dropped. I tried to use `reshape`, but it does not seem to be able to give me what I want. I would be grateful for any help.