Application 1 (cont.)
How long do PhD students study for? (cont.)
The following year the number of students admitted to the same institution rose rapidly.
Exercise
Calculate the expected duration of study in the 2000–2001 academic year and compare it to the expected duration of 4.1 years in 1999–2000.
Full-time research students registered at a UK school of public health.
| Year of study (x) |
Px 1/12/2000 | Px 01/12/2001 | Growth rate rx |
Px(2000-01) | Lx |
|---|---|---|---|---|---|
| 1 | 37 | 63 | 0.5308 |
48.3 |
63.0 |
| 2 | 50 | 36 | -0.3276 |
42.4 |
61.2 |
| 3 | 52 | 50 | -0.0391 |
51.0 |
61.3 |
| 4 | 29 | 40 | 0.3207 |
34.1 |
47.1 |
| 5 | 15 | 7 | -0.7601 |
10.2 |
11.4 |
| 6+ | 2 | 1 | -0.6913 |
1.4 |
0.5 |
| Total | 183 | 196 | 187.4 |
244.5 |
|
|
Expected duration of study =∑Lx/l0 = 244.5/63 = | 3.9 years
|
|||
The correct answer is 3.9
Please attempt the answer.
Yes, that’s correct! The answer is 3.9.
Clearly, the population of research students being analysed was neither a stationary nor stable one as the number of new students registering varied greatly from year to year. Taking out these erratic fluctuations in the rate of growth of the student population, however, reveals that the expected length of registration of a student was far less volatile at about 4 years in both the 1999–2000 and 2000–2001 synthetic cohorts.
No, that’s not right.
This method can be used to calculate the expected duration of a stay in any institutional population, such as a school, prison, or a nursing care home, or more generally, life expectancy in any population. In particular, Preston and Bennett (1983) describe how it can be used to estimate adult mortality at the national level using data from two successive censuses of a country.
Whatever the application, the only data required are two enumerations of the population tabulated by duration since entering the population and the number of people entering the population in the period in between the two enumerations.
The key assumption made by the method is that there are no unrecorded exits from the population via other means. For example, the estimation of mortality from two national censuses requires the assumption that the country is closed to migration or, at least, that the age-specific net migration rates are trivial in size compared to the age-specific death rates.