[Brendan Churchill]

However, we also want to assign a value of 1 to observations that fall below the tagged observation where status does not change. This can be useful because you may want to investigate whether the length of time spent in a state (for unemployment, also called "spell of unemployment") affects well-being. It may be that there is no significant change in well-being for individuals who transition out of and into employment in the space of a year but not for those who spend a longer period out of employment. This allows you to check this.

Also, you mentioned in #9 about investigating length of time in a state or between transitions (e.g. time out of the labour force) - how would I, for example calculate the spells between one state and another into a kind of continuous variable?

Best
Brendan

[Andrew Musau]

Thanks Andrew - most helpful
and if I wanted to do it separately for each outcome, e.g. FT and PT > OUT or PT and OUT > FT in the next year, how would I do that?

If you want to consider a transition into one state only, e.g., into full time employment, then change

inlist(employment, 1, 2, 3)

to

inlist(employment, 1)

in #14. This will give you transition from any state to full-time employment. The same for any state to part-time

inlist(employment, 2)

and so on.

Also, you mentioned in #9 about investigating length of time in a state or between transitions (e.g. time out of the labour force) - how would I, for example calculate the spells between one state and another into a kind of continuous variable?

Taking your initial data example, after creating the transition variable, you can create a variable that measures the duration in a particular state in the following way

* Example generated by -dataex-. To install: ssc install dataex
clear
input byte(losatyh esdtl id) float(year transition)
 5 6  1 2008 0
 6 2  1 2009 1
 5 2  1 2010 0
 6 2  1 2011 0
 4 2  1 2012 0
 4 2  1 2013 0
 5 2  1 2014 0
 8 2  1 2015 0
 8 2  1 2016 0
 8 2  1 2017 0
 7 2  2 2008 0
 8 1  2 2009 1
 6 1  2 2010 0
 7 3  2 2011 1
 6 1  2 2012 1
 8 5  2 2013 1
 7 2  2 2014 1
 8 2  2 2015 0
 4 6  2 2016 1
 5 5  2 2017 1
 2 6  3 2008 0
 5 6  3 2009 0
 5 6  3 2010 0
10 2  3 2011 1
 8 2  3 2012 0
 9 2  3 2013 0
 8 2  3 2014 0
10 2  3 2015 0
 8 2  3 2016 0
 8 2  3 2017 0
 8 2  4 2008 0
 8 2  4 2009 0
 8 6  4 2010 1
 8 5  4 2011 1
 7 6  4 2012 1
 7 5  4 2013 1
 7 3  4 2014 1
 6 4  4 2015 1
 8 3  4 2016 1
 7 5  4 2017 1
 7 2  5 2008 0
 8 4  5 2009 1
 7 4  5 2010 0
 7 2  5 2011 1
 6 2  5 2012 0
 8 1  5 2013 1
 7 2  5 2014 1
 5 2  5 2015 0
 7 1  5 2016 1
 7 6  5 2017 1
 4 2  6 2008 0
 5 2  6 2009 0
 7 2  6 2010 0
10 4  6 2011 1
10 2  6 2012 1
 9 3  6 2013 1
10 6  6 2014 1
10 5  6 2015 1
 6 6  6 2016 1
 5 6  6 2017 0
 5 6  7 2008 0
 6 6  7 2009 0
 8 6  7 2010 0
 7 6  7 2011 0
 5 6  7 2012 0
 2 6  7 2013 0
 2 6  7 2014 0
 1 6  7 2015 0
 3 6  7 2016 0
 1 6  7 2017 0
 5 6  8 2008 0
10 1  8 2009 1
10 2  8 2010 1
 9 2  8 2011 0
 8 2  8 2012 0
 8 2  8 2013 0
 5 5  8 2014 1
 8 3  8 2015 1
 5 5  8 2016 1
 1 6  8 2017 1
 5 2  9 2008 0
 5 2  9 2009 0
 7 1  9 2010 1
 5 2  9 2011 1
 6 2  9 2012 0
 6 2  9 2013 0
 8 5  9 2014 1
 8 5  9 2015 0
 8 5  9 2016 0
10 5  9 2017 0
 6 6 10 2008 0
 7 6 10 2009 0
 7 5 10 2010 1
 4 2 10 2011 1
 6 5 10 2012 1
 5 1 10 2013 1
 6 1 10 2014 0
 7 1 10 2015 0
 7 1 10 2016 0
 7 1 10 2017 0
end
label values losatyh OLOSATYH
label def OLOSATYH 5 "[5] Neither satisfied nor dissatisfied", modify
label def OLOSATYH 10 "[10] Totally satisfied", modify
label values esdtl OESDTL
label def OESDTL 1 "[1] Employed FT", modify
label def OESDTL 2 "[2] Employed PT", modify
label def OESDTL 3 "[3] Unemployed, looking for FT work", modify
label def OESDTL 4 "[4] Unemployed, looking for PT work", modify
label def OESDTL 5 "[5] Not in the labour force, marginally attached", modify
label def OESDTL 6 "[6] Not in the labour force, not marginally attached", modify

gen group= sum(transition) if transition==1
replace group=group[_n-1] if missing(group) & _n>1
replace group=0 if missing(group)
bys id group (year): egen duration= count(group)

It is not clear what you should do if you have missing years as you do not know whether there was any transition during these years. Therefore, this assumes a balanced panel.

Res.:

. l id year esdtl duration, sepby(id)

     +-----------------------------------------------------------------------------+
     | id   year                                                  esdtl   duration |
     |-----------------------------------------------------------------------------|
  1. |  1   2008   [6] Not in the labour force, not marginally attached          1 |
  2. |  1   2009                                        [2] Employed PT          9 |
  3. |  1   2010                                        [2] Employed PT          9 |
  4. |  1   2011                                        [2] Employed PT          9 |
  5. |  1   2012                                        [2] Employed PT          9 |
  6. |  1   2013                                        [2] Employed PT          9 |
  7. |  1   2014                                        [2] Employed PT          9 |
  8. |  1   2015                                        [2] Employed PT          9 |
  9. |  1   2016                                        [2] Employed PT          9 |
 10. |  1   2017                                        [2] Employed PT          9 |
     |-----------------------------------------------------------------------------|
 11. |  2   2008                                        [2] Employed PT          1 |
 12. |  2   2009                                        [1] Employed FT          2 |
 13. |  2   2010                                        [1] Employed FT          2 |
 14. |  2   2011                    [3] Unemployed, looking for FT work          1 |
 15. |  2   2012                                        [1] Employed FT          1 |
 16. |  2   2013       [5] Not in the labour force, marginally attached          1 |
 17. |  2   2014                                        [2] Employed PT          2 |
 18. |  2   2015                                        [2] Employed PT          2 |
 19. |  2   2016   [6] Not in the labour force, not marginally attached          1 |
 20. |  2   2017       [5] Not in the labour force, marginally attached          1 |
     |-----------------------------------------------------------------------------|
 21. |  3   2008   [6] Not in the labour force, not marginally attached          3 |
 22. |  3   2009   [6] Not in the labour force, not marginally attached          3 |
 23. |  3   2010   [6] Not in the labour force, not marginally attached          3 |
 24. |  3   2011                                        [2] Employed PT          7 |
 25. |  3   2012                                        [2] Employed PT          7 |
 26. |  3   2013                                        [2] Employed PT          7 |
 27. |  3   2014                                        [2] Employed PT          7 |
 28. |  3   2015                                        [2] Employed PT          7 |
 29. |  3   2016                                        [2] Employed PT          7 |
 30. |  3   2017                                        [2] Employed PT          7 |
     |-----------------------------------------------------------------------------|
 31. |  4   2008                                        [2] Employed PT          2 |
 32. |  4   2009                                        [2] Employed PT          2 |
 33. |  4   2010   [6] Not in the labour force, not marginally attached          1 |
 34. |  4   2011       [5] Not in the labour force, marginally attached          1 |
 35. |  4   2012   [6] Not in the labour force, not marginally attached          1 |
 36. |  4   2013       [5] Not in the labour force, marginally attached          1 |
 37. |  4   2014                    [3] Unemployed, looking for FT work          1 |
 38. |  4   2015                    [4] Unemployed, looking for PT work          1 |
 39. |  4   2016                    [3] Unemployed, looking for FT work          1 |
 40. |  4   2017       [5] Not in the labour force, marginally attached          1 |
     |-----------------------------------------------------------------------------|
 41. |  5   2008                                        [2] Employed PT          1 |
 42. |  5   2009                    [4] Unemployed, looking for PT work          2 |
 43. |  5   2010                    [4] Unemployed, looking for PT work          2 |
 44. |  5   2011                                        [2] Employed PT          2 |
 45. |  5   2012                                        [2] Employed PT          2 |
 46. |  5   2013                                        [1] Employed FT          1 |
 47. |  5   2014                                        [2] Employed PT          2 |
 48. |  5   2015                                        [2] Employed PT          2 |
 49. |  5   2016                                        [1] Employed FT          1 |
 50. |  5   2017   [6] Not in the labour force, not marginally attached          1 |
     |-----------------------------------------------------------------------------|
 51. |  6   2008                                        [2] Employed PT          3 |
 52. |  6   2009                                        [2] Employed PT          3 |
 53. |  6   2010                                        [2] Employed PT          3 |
 54. |  6   2011                    [4] Unemployed, looking for PT work          1 |
 55. |  6   2012                                        [2] Employed PT          1 |
 56. |  6   2013                    [3] Unemployed, looking for FT work          1 |
 57. |  6   2014   [6] Not in the labour force, not marginally attached          1 |
 58. |  6   2015       [5] Not in the labour force, marginally attached          1 |
 59. |  6   2016   [6] Not in the labour force, not marginally attached          2 |
 60. |  6   2017   [6] Not in the labour force, not marginally attached          2 |
     |-----------------------------------------------------------------------------|
 61. |  7   2008   [6] Not in the labour force, not marginally attached         10 |
 62. |  7   2009   [6] Not in the labour force, not marginally attached         10 |
 63. |  7   2010   [6] Not in the labour force, not marginally attached         10 |
 64. |  7   2011   [6] Not in the labour force, not marginally attached         10 |
 65. |  7   2012   [6] Not in the labour force, not marginally attached         10 |
 66. |  7   2013   [6] Not in the labour force, not marginally attached         10 |
 67. |  7   2014   [6] Not in the labour force, not marginally attached         10 |
 68. |  7   2015   [6] Not in the labour force, not marginally attached         10 |
 69. |  7   2016   [6] Not in the labour force, not marginally attached         10 |
 70. |  7   2017   [6] Not in the labour force, not marginally attached         10 |
     |-----------------------------------------------------------------------------|
 71. |  8   2008   [6] Not in the labour force, not marginally attached          1 |
 72. |  8   2009                                        [1] Employed FT          1 |
 73. |  8   2010                                        [2] Employed PT          4 |
 74. |  8   2011                                        [2] Employed PT          4 |
 75. |  8   2012                                        [2] Employed PT          4 |
 76. |  8   2013                                        [2] Employed PT          4 |
 77. |  8   2014       [5] Not in the labour force, marginally attached          1 |
 78. |  8   2015                    [3] Unemployed, looking for FT work          1 |
 79. |  8   2016       [5] Not in the labour force, marginally attached          1 |
 80. |  8   2017   [6] Not in the labour force, not marginally attached          1 |
     |-----------------------------------------------------------------------------|
 81. |  9   2008                                        [2] Employed PT          2 |
 82. |  9   2009                                        [2] Employed PT          2 |
 83. |  9   2010                                        [1] Employed FT          1 |
 84. |  9   2011                                        [2] Employed PT          3 |
 85. |  9   2012                                        [2] Employed PT          3 |
 86. |  9   2013                                        [2] Employed PT          3 |
 87. |  9   2014       [5] Not in the labour force, marginally attached          4 |
 88. |  9   2015       [5] Not in the labour force, marginally attached          4 |
 89. |  9   2016       [5] Not in the labour force, marginally attached          4 |
 90. |  9   2017       [5] Not in the labour force, marginally attached          4 |
     |-----------------------------------------------------------------------------|
 91. | 10   2008   [6] Not in the labour force, not marginally attached          2 |
 92. | 10   2009   [6] Not in the labour force, not marginally attached          2 |
 93. | 10   2010       [5] Not in the labour force, marginally attached          1 |
 94. | 10   2011                                        [2] Employed PT          1 |
 95. | 10   2012       [5] Not in the labour force, marginally attached          1 |
 96. | 10   2013                                        [1] Employed FT          5 |
 97. | 10   2014                                        [1] Employed FT          5 |
 98. | 10   2015                                        [1] Employed FT          5 |
 99. | 10   2016                                        [1] Employed FT          5 |
100. | 10   2017                                        [1] Employed FT          5 |
     +-----------------------------------------------------------------------------+

[Brendan Churchill]

Thanks Andrew Musau, this code has been invaluable to my work. I do have a question about it.

How would I just make a variable that captures years since transition, not just a count of years in the transition status as below:

 
 gen group= sum(transition) if transition==1 replace group=group[_n-1] if missing(group) & _n>1 replace group=0 if missing(group) bys id group (year): egen duration= count(group)

But one that counts every year/wave since transition. So if an ID had was in the sample for 9 years after initial transition, each year would be counted 1, 2, 3, 4, 5, 6.... at each wave

Any help would be most appreciated

best
Brendan

[Andrew Musau]

* Example generated by -dataex-. For more info, type help dataex
clear
input byte(losatyh esdtl id) float(year transition group duration)
 5 6  1 2008 0  0  1
 6 2  1 2009 1  1  9
 5 2  1 2010 0  1  9
 6 2  1 2011 0  1  9
 4 2  1 2012 0  1  9
 4 2  1 2013 0  1  9
 5 2  1 2014 0  1  9
 8 2  1 2015 0  1  9
 8 2  1 2016 0  1  9
 8 2  1 2017 0  1  9
 7 2  2 2008 0  1  1
 8 1  2 2009 1  2  2
 6 1  2 2010 0  2  2
 7 3  2 2011 1  3  1
 6 1  2 2012 1  4  1
 8 5  2 2013 1  5  1
 7 2  2 2014 1  6  2
 8 2  2 2015 0  6  2
 4 6  2 2016 1  7  1
 5 5  2 2017 1  8  1
 2 6  3 2008 0  8  3
 5 6  3 2009 0  8  3
 5 6  3 2010 0  8  3
10 2  3 2011 1  9  7
 8 2  3 2012 0  9  7
 9 2  3 2013 0  9  7
 8 2  3 2014 0  9  7
10 2  3 2015 0  9  7
 8 2  3 2016 0  9  7
 8 2  3 2017 0  9  7
 8 2  4 2008 0  9  2
 8 2  4 2009 0  9  2
 8 6  4 2010 1 10  1
 8 5  4 2011 1 11  1
 7 6  4 2012 1 12  1
 7 5  4 2013 1 13  1
 7 3  4 2014 1 14  1
 6 4  4 2015 1 15  1
 8 3  4 2016 1 16  1
 7 5  4 2017 1 17  1
 7 2  5 2008 0 17  1
 8 4  5 2009 1 18  2
 7 4  5 2010 0 18  2
 7 2  5 2011 1 19  2
 6 2  5 2012 0 19  2
 8 1  5 2013 1 20  1
 7 2  5 2014 1 21  2
 5 2  5 2015 0 21  2
 7 1  5 2016 1 22  1
 7 6  5 2017 1 23  1
 4 2  6 2008 0 23  3
 5 2  6 2009 0 23  3
 7 2  6 2010 0 23  3
10 4  6 2011 1 24  1
10 2  6 2012 1 25  1
 9 3  6 2013 1 26  1
10 6  6 2014 1 27  1
10 5  6 2015 1 28  1
 6 6  6 2016 1 29  2
 5 6  6 2017 0 29  2
 5 6  7 2008 0 29 10
 6 6  7 2009 0 29 10
 8 6  7 2010 0 29 10
 7 6  7 2011 0 29 10
 5 6  7 2012 0 29 10
 2 6  7 2013 0 29 10
 2 6  7 2014 0 29 10
 1 6  7 2015 0 29 10
 3 6  7 2016 0 29 10
 1 6  7 2017 0 29 10
 5 6  8 2008 0 29  1
10 1  8 2009 1 30  1
10 2  8 2010 1 31  4
 9 2  8 2011 0 31  4
 8 2  8 2012 0 31  4
 8 2  8 2013 0 31  4
 5 5  8 2014 1 32  1
 8 3  8 2015 1 33  1
 5 5  8 2016 1 34  1
 1 6  8 2017 1 35  1
 5 2  9 2008 0 35  2
 5 2  9 2009 0 35  2
 7 1  9 2010 1 36  1
 5 2  9 2011 1 37  3
 6 2  9 2012 0 37  3
 6 2  9 2013 0 37  3
 8 5  9 2014 1 38  4
 8 5  9 2015 0 38  4
 8 5  9 2016 0 38  4
10 5  9 2017 0 38  4
 6 6 10 2008 0 38  2
 7 6 10 2009 0 38  2
 7 5 10 2010 1 39  1
 4 2 10 2011 1 40  1
 6 5 10 2012 1 41  1
 5 1 10 2013 1 42  5
 6 1 10 2014 0 42  5
 7 1 10 2015 0 42  5
 7 1 10 2016 0 42  5
 7 1 10 2017 0 42  5
end
label values losatyh OLOSATYH
label def OLOSATYH 5 "[5] Neither satisfied nor dissatisfied", modify
label def OLOSATYH 10 "[10] Totally satisfied", modify
label values esdtl OESDTL
label def OESDTL 1 "[1] Employed FT", modify
label def OESDTL 2 "[2] Employed PT", modify
label def OESDTL 3 "[3] Unemployed, looking for FT work", modify
label def OESDTL 4 "[4] Unemployed, looking for PT work", modify
label def OESDTL 5 "[5] Not in the labour force, marginally attached", modify
label def OESDTL 6 "[6] Not in the labour force, not marginally attached", modify

bys id group (year): gen yearsince= year-year[1] if transition[1] & _N>1

Res.:

. l id- yearsince , sepby(id)

     +----------------------------------------------------+
     | id   year   transi~n   group   duration   yearsi~e |
     |----------------------------------------------------|
  1. |  1   2008          0       0          1          . |
  2. |  1   2009          1       1          9          0 |
  3. |  1   2010          0       1          9          1 |
  4. |  1   2011          0       1          9          2 |
  5. |  1   2012          0       1          9          3 |
  6. |  1   2013          0       1          9          4 |
  7. |  1   2014          0       1          9          5 |
  8. |  1   2015          0       1          9          6 |
  9. |  1   2016          0       1          9          7 |
 10. |  1   2017          0       1          9          8 |
     |----------------------------------------------------|
 11. |  2   2008          0       1          1          . |
 12. |  2   2009          1       2          2          0 |
 13. |  2   2010          0       2          2          1 |
 14. |  2   2011          1       3          1          . |
 15. |  2   2012          1       4          1          . |
 16. |  2   2013          1       5          1          . |
 17. |  2   2014          1       6          2          0 |
 18. |  2   2015          0       6          2          1 |
 19. |  2   2016          1       7          1          . |
 20. |  2   2017          1       8          1          . |
     |----------------------------------------------------|
 21. |  3   2008          0       8          3          . |
 22. |  3   2009          0       8          3          . |
 23. |  3   2010          0       8          3          . |
 24. |  3   2011          1       9          7          0 |
 25. |  3   2012          0       9          7          1 |
 26. |  3   2013          0       9          7          2 |
 27. |  3   2014          0       9          7          3 |
 28. |  3   2015          0       9          7          4 |
 29. |  3   2016          0       9          7          5 |
 30. |  3   2017          0       9          7          6 |
     |----------------------------------------------------|
 31. |  4   2008          0       9          2          . |
 32. |  4   2009          0       9          2          . |
 33. |  4   2010          1      10          1          . |
 34. |  4   2011          1      11          1          . |
 35. |  4   2012          1      12          1          . |
 36. |  4   2013          1      13          1          . |
 37. |  4   2014          1      14          1          . |
 38. |  4   2015          1      15          1          . |
 39. |  4   2016          1      16          1          . |
 40. |  4   2017          1      17          1          . |
     |----------------------------------------------------|
 41. |  5   2008          0      17          1          . |
 42. |  5   2009          1      18          2          0 |
 43. |  5   2010          0      18          2          1 |
 44. |  5   2011          1      19          2          0 |
 45. |  5   2012          0      19          2          1 |
 46. |  5   2013          1      20          1          . |
 47. |  5   2014          1      21          2          0 |
 48. |  5   2015          0      21          2          1 |
 49. |  5   2016          1      22          1          . |
 50. |  5   2017          1      23          1          . |
     |----------------------------------------------------|
 51. |  6   2008          0      23          3          . |
 52. |  6   2009          0      23          3          . |
 53. |  6   2010          0      23          3          . |
 54. |  6   2011          1      24          1          . |
 55. |  6   2012          1      25          1          . |
 56. |  6   2013          1      26          1          . |
 57. |  6   2014          1      27          1          . |
 58. |  6   2015          1      28          1          . |
 59. |  6   2016          1      29          2          0 |
 60. |  6   2017          0      29          2          1 |
     |----------------------------------------------------|
 61. |  7   2008          0      29         10          . |
 62. |  7   2009          0      29         10          . |
 63. |  7   2010          0      29         10          . |
 64. |  7   2011          0      29         10          . |
 65. |  7   2012          0      29         10          . |
 66. |  7   2013          0      29         10          . |
 67. |  7   2014          0      29         10          . |
 68. |  7   2015          0      29         10          . |
 69. |  7   2016          0      29         10          . |
 70. |  7   2017          0      29         10          . |
     |----------------------------------------------------|
 71. |  8   2008          0      29          1          . |
 72. |  8   2009          1      30          1          . |
 73. |  8   2010          1      31          4          0 |
 74. |  8   2011          0      31          4          1 |
 75. |  8   2012          0      31          4          2 |
 76. |  8   2013          0      31          4          3 |
 77. |  8   2014          1      32          1          . |
 78. |  8   2015          1      33          1          . |
 79. |  8   2016          1      34          1          . |
 80. |  8   2017          1      35          1          . |
     |----------------------------------------------------|
 81. |  9   2008          0      35          2          . |
 82. |  9   2009          0      35          2          . |
 83. |  9   2010          1      36          1          . |
 84. |  9   2011          1      37          3          0 |
 85. |  9   2012          0      37          3          1 |
 86. |  9   2013          0      37          3          2 |
 87. |  9   2014          1      38          4          0 |
 88. |  9   2015          0      38          4          1 |
 89. |  9   2016          0      38          4          2 |
 90. |  9   2017          0      38          4          3 |
     |----------------------------------------------------|
 91. | 10   2008          0      38          2          . |
 92. | 10   2009          0      38          2          . |
 93. | 10   2010          1      39          1          . |
 94. | 10   2011          1      40          1          . |
 95. | 10   2012          1      41          1          . |
 96. | 10   2013          1      42          5          0 |
 97. | 10   2014          0      42          5          1 |
 98. | 10   2015          0      42          5          2 |
 99. | 10   2016          0      42          5          3 |
100. | 10   2017          0      42          5          4 |
     +----------------------------------------------------+

.

[Brendan Churchill]

Thanks Andrew Musau, most helpful. One more can you augment the variable to include years BEFORE the transition too?
best
Brendan

[Andrew Musau]

You have several transitions. How do you propose to do that? Show your intended results using a data example, e.g., #19.