Вы находитесь на странице: 1из 1

Demographie Survey in Panama in 1976.

They illustrate some of the of this type are often due to changing marriage pattems, but the fact
probiems encountered when dealing with real dala. that the proportion single declines at ages 50-54 suggests that the rela-
A quiclc examination of table 187 reveals that, especially at older tively high value observed at ages 45-49 may be due, at least in part, to
ages, the marital status of a substanţial proportion of the persons misreporting of status rather than to true changes in the marriage pat­
interviewed is unknown. It is difficult to imagine a reason for this tems of the past.
omission since a priori it would seem easy for a person to remember Thetefore, to calculate the singulate mean age of marriage in this
whether he or she had been married. case, the reported proportions single were adjusted by replacing the
A possible source of misunderstanding may be that the term “ever value of 0.0470 for age group 45-49 by the average of the two values
married” was used in a broad sense to mean “having belonged to a adjacent to it, which happen to be also the two smallest proportions
stable union". Because in many countries, unions not legalized by single observed. Thus, it was assumed that the proportion single at
marriage are socially unacceptable, a respondent in such a situation ages 45-49 was 0.0324. Under this assumption, the singulate mean age
might tend to avoid acknowledging his or her “marital status”. It then at marriage is computed in the usual way. The results of the main
seems plausible to assume that most of the “unknowns” were, in fact, steps are shown below:
“ever married”. The validity of this assumption is, however, impossi-
ble to establish without recourse to further Information. But if the RSi = 5.0(1.584)= 7.92
data available are to be used in computing singulate mean ages at
marriage for each sex, it is necessaiy to make some assumption about AS2= 7.92 +15.0= 22.92
their true meaning. As suggested, one extreme assumption is to sup-
pose that all the “unknowns" are ever married. Another, less extreme RN = (0.0324 +0.0323) /2.0 = 0.0324
possibility is to assume that both the single and the ever-married per­
sons within each age group have an equal probability of being RM = 0.9677
classified as “unknown” (this assumption would be plausible if most of
the unknowns were the result of random errors produced while Pro­ RS }= 50.0(0.0324)= 1.62
cessing the data). Under this assumption, the proportion single in
each age group would remain the same if the unknowns were ignored SMAM = (RS2 -RS3 )/RM = 22.01.
as if they had declared their marital status properly.
Each of these assumptions leads to a different way of computing the It is interesting to note that the value of SMAM would have been
proportion single in each age group and to somewhat different esti- 21.88 if the observed data, without adjustment, had been used, a value
mates of the desired singulate mean ages at marriage. The computa­ that is very similar to that obtained above.
ţional procedures and results are shown below.
(b) Calculation of singulate mean age at marriage assuming that those of
(a) Calculation of singulate mean age at marriage ignoring those of un­ unknown marital status are ever married
known marital status The values of SMAM are computed next by assuming that the
In the first instance, the proportions single are calculated by ignor­ “unknowns” belong, in fact, to the ever-married categoty. The pro­
ing the unknowns. The resulting proportions are shown in table 188 portions single obtained under this assumption are shown in table 189.
Then, using these proportions single, the singulate mean ages at mar­ The computations needed to arrive at the final value of SMAM are
riage are computed as usual. summarized in cases 3 and 4.
(i) Case 3: males (unknowns considered to be ever married)
The results for case 3 are shown below:
T able 188. P roportions single ignorino those of unknown
MARITAL STATUS, PANAMA, 1976 RS 2 = 5.0(2.416)+15.0= 27.08

Agt RS 2 = 50.0( (0.0783 +0.0542)/2.0) = 50.0(0.0663)= 3.3125


Omm Mak$ Ftmaies
iîy (2) m
SMAM = (*Sj-/tSj)/0.9337 = 25.45.
15-19............... 0.9749 0.8101
20-24................ 0.6947 0.3918
(ii) Case 4: females (unknowns considered tobe ever married)
25-29................ 0.3257 0.1608
30-34................ 0.1722 0.0936 Once moţe, the inconsistency apparent in age group 45-49 was
35-39................ 0.1248 0.0628 corrected by replacing the reported proportion single, 1/(7) (0.0441),
40-44................ 0.0946 0.0325 by the average of those adjacent to it (0.0305). An outline of the cal-
45-49................ 0.0893 0.0470 (0.0324) culations follows:
50-54................ 0.0626 0.0323
RS 2 = 5.0(1.5605)+15.0= 22.80

RS 2 = 50.0(0.0302)= 1.51
(i) Case 1: males (unknowns ignored)
The results of the main steps for case 1 are given below: SMAM = (KS2—/tSj)/0.9698= 21.95.
/tS,= 5.0(2.4762)= 12.381

*S2=/IS,+15.0= 27.381 T able 189. P roportions single assuming that those of unknown
MARITAL STATUS ARE EVER MARRIED, PANAMA, 1976
RN = (0.0893 +0.0626)/2.0 = 0.076 Proportions Angli
Agr
Males Ftmatn
RM = 0.924 *zr m tl)
RS 2 =S0.0RN = 3.798 15-19................ 0.9731 0.8056
20-24................ 0.6861 0.3847
SMAM = (RS2 -RS2 )/RM = 25.52. 25-29................ 0.3143 0.1575
30-34................ 0.1624 0.0910
35-39................ 0.I16S 0.0602
(ii) Case 2: females (unknowns ignored) 40-44................ 0.0853 0.0310
Note that the proportions single among females do not, as expected. 45-49................ 0.0783 0.0441 (0.0305)
deciease consistently as age increases. The proportion single at ages 50-54................ 0.0542 0.0300
45-49, for example, is greater than that at ages 40-44. Inconsistencies

226

Вам также может понравиться