4. Secondary Analyses

As part of its work for the Arroyo Center, RAND undertook a study of family programs and readiness in the Army (Burnam et al., 1992). A Survey of Army Families was fielded in 1987 with the following goals: (1) to determine the effect of personal and family problems on Army duties (primarily readiness), (2) to evaluate the effects of Army practices (e.g., deployments and rotations) on individual and family well-being, (3) to systematically measure the need for, use of, and effectiveness of important services (including child care), and (4) to determine the relative magnitude of specific concerns and identify particularly affected subpopulations or locations.

Although the RAND Arroyo Center Survey of Army Families included previous reports from the child care programs (see Burnam et al., 1992), our analyses provide greater detail than was previously reported about issues that are particularly relevant to the present study. In this section we analyze those data to address the following questions:

  1. To what extent is time lost from duty a problem for families with children? Does this vary by type of child care arrangement or by the employment status of the spouse?
  2. Which subgroups lose the most time to duty?
  3. Do CDCs and FDC appear to meet all the child care needs of the families that they serve?

Survey Measures

Child Care Arrangements. We analyze the child care arrangements of the respondents' youngest (or only) preschool-aged child[1] by defining the "modal child care arrangement" as the type of child care relied upon the most hours during the week by each respondent. We created six child care categories: (1) Army Child Development Center (CDC); (2) civilian day care center, preschool, or kindergarten; (3) Army Family Child Care (FDC);[2] (4) care by a nonrelative (including babysitters, friends, and neighbors); (5) care by a relative (including spouse, siblings, or self care); and (6) no care. We classify someone as having child care if the respondent reported using at least six hours of care per week.[3] As a result, the group referred to as having no care includes those respondents who reported five or fewer hours of care per week. We assumed that parents using fewer than six hours per week are not regular users of child care (e.g., they use child care only periodically).

If parents reported using more than one type of care for equal amounts of time each week, they were assigned to the category representing the most formal type of care. For example, if a parent reported using Army CDC and a nonrelative sitter for 20 hours per week each, the modal type of child care arrangement for that family was classified as Army CDC.

Spouse Employment Status. We analyze spouse's employment status by distinguishing families in which the military parent was: (1) single, (2) married to a nonworking spouse, or (3) married to a spouse who was currently in the labor force.[4]

Problems with Child Care and Children. We assess the extent of time lost from duty by determining the number of days of absence for child care reasons in the past year, excluding leave (vacation) time. The total amount of time lost to duty was derived by summing responses to two questions concerning the amount of time lost due to: (1) caring for child(ren) on a daily basis (e.g., supervision, discipline), and (2) other child care concerns (e.g., the child is sick, a visit to the school).[5] The period of recall in the survey for these two questions was a month. To obtain annual measures, we multiplied each value by 12.[6]

We examine the perceived adequacy of child care during military exercises by creating an indicator variable for soldiers according to whether they rated their child care arrangements during the most recent planned deployment or exercise of two or more weeks (if such a deployment had occurred in the previous year) as fair or poor, or would have been fair, poor, or impossible if they had been deployed for six months or more.

We created an indicator variable for classifying parents' overall rating of their child care arrangements. This variable takes on the value "1" if the parents reported their overall satisfaction with their child care arrangement as fair or poor, and "0" otherwise.

We also created an indicator of parental uncertainty concerning the Army's helpfulness with child care during prolonged separation. Based on the question "If a military conflict separated you from your family for six months or more, how sure are you that the Army would help your family with child care?" a 1 indicates that a soldier responded "somewhat unsure" or "completely unsure" to the question. This variable takes on the value of 0 for all other responses.


The results of our analyses are reported in Tables 4.1-4.5. These results are based on our sample of 1031 respondents with accompanying children aged 0-5 years at the time of the survey. Because of nonresponses to individual items, the sample size obtained for individual measures varies. The data are weighted to adjust for differences in response rates, oversampling of unique groups, and for underrepresentation of the universe of Army members and spouses. As a result of this weighting, the responses represent about 80 percent of the Army's 777,000 soldiers in 1987 with accompanying children aged 0-5 years. To the extent that this population has changed since the fielding of this survey (in 1987), the results may not be completely generalizable to this population today.

Table 4.1 contains information about the use of various types of child care. Several interesting results emerge. First, nearly 83 percent of all families with preschool-aged children rely on some kind of child care for at least six hours per week.

Table 4.1
Distribution of Child Care Combinations

Type of Child Care Percent
< 6 hours weekly (N = 138) 17.4
Only one type of care (N = 294) 37.2
Two or more types of care (N = 358) 45.4
Army CDC, 1+ hours (N = 115) 17.6
1-5 hours 7.5
Only, 6+ hours 3.5
Plus other, 6+ hours 6.6
Modal care 6+ hours 8.0
Army FDC, 1+ hours (N = 45) 6.9
1-5 hours 2.0
Only, 6+ hours 1.1
Plus other, 6+ hours 3.8
Modal care 6+ hours 2.7
Civilian, 1+ hours (N = 74) 11.3
1-5 hours 0.9
Only, 6+ hours 3.4
Plus other, 6+ hours 7.0
Modal care 6+ hours 13.9
Nonrelative, 1+ hours (N = 452) 69.3
1-5 hours 30.7
Only, 6+ hours 16.7
Plus other, 6+ hours 21.9
Modal care 6+ hours 30.5
Relative, 1+ hours (N = 337) 51.8
1-5 hours 25.2
Only, 6+ hours 5.4
Plus other, 6+ hours 21.2
Modal care 6+ hours 14.8

NOTE: Sample is restricted to those reporting use of one or more hours of care weekly (N = 652).

Second, a large percentage (45 percent) of families rely on more than one type of child care. Indeed, among families using child care for at least six hours per week, 55 percent (45.4/(45.4 + 37.2)) rely on more than one type of care, suggesting that the child care needs of most families cannot be filled by one type of care.

Third, whereas 17.6 percent of all Army families report some use of Army CDCs, only 3.5 percent rely on them exclusively. Furthermore, only 8.0 percent use the CDC as their modal (i.e., most frequently used) type of care arrangement.

Fourth, whereas 6.9 percent of Army families report some use of FDC, only 1.1 percent rely on FDC exclusively. And only 2.7 percent use FDC as their modal care arrangement.

Fifth, a much lower percentage (less than 1 percent) of families use civilian day care centers for five or fewer hours of care per week than use Army CDC (7.5 percent). This is probably a reflection of the fact that most civilian child care centers do not offer hourly care. To the extent that some Army families need such care, the Army seems to be meeting at best some of that need.

Table 4.2 shows the distribution of modal type of care by spouse employment status. A greater proportion of single-parent families report relying on Army CDC as their modal child care arrangement, which is not surprising given the priority this group receives for CDC slots.[7] Also noteworthy is the fact that civilian day care centers predominantly are used by families with working spouses. Finally, 33.5 percent of single-parent families report using less than six hours of child care per week. One explanation for these data may be that some children of single parents are not living with them.[8]

Table 4.2
Modal Type of Child Care by Spouse Employment Status

Spouse Employment Status
Modal Type of Child Care All Single Nonworking Working
Army CDC 7.9 18.0 6.0 9.8
Civilian 7.3 3.8 4.3 13.0
Army FDC 2.7 7.4 0.5 5.9
Nonrelative 30.6 35.2 18.6 51.1
Relative 15.0 2.2 18.4 10.9
1-5 hours/no use 36.5 33.5 52.2 9.3
All 100.0 100.0 100.0 100.0

NOTE: The percentage of parents using each type of child care includes all parents, even those using no care and those using less than six hours weekly. The distribution is based on weighted estimates of a sample (N = 1031).

Table 4.3 displays findings from the various measures of problems resulting from child care or children as a function of the employment status of the spouse. Parents on average lose about three days per year from duty for child care reasons, not an insignificant amount of time when viewed from the perspective of an employer. Indeed, given the large number of families with young children in the military, a substantial number of work days are lost per year.[9] Not surprisingly, single parents report a significantly higher number of days (eight) per year lost from duty for these reasons. This is more than 2-1/2 times the overall average. Even in families in which the spouse does not work, the military member still reports losing duty time because of children.

Table 4.3
Problems with Child Care or Children by Spouse Employment Status

Spouse Employment Status
Type of Problem All Single Nonworking Working
Days lost from duty for child care reasons (N = 914) 3.1 8.0[a] 1.9[b] 4.5[b]
% reporting fair or poor child care during deployment (N = 289) 32.6 32.2 29.8 35.8
% rating child care arrangements fair or poor (N = 937) 12.2 13.6 12.2 12.0
% unsure Army would help (N = 988) 48.8 72.5[a] 43.1[b] 55.4[c]

NOTE: Row 1 entries are means. Cell entries for rows 2-4 are percentages. Entries significantly different from one another at the 5 percent level are denoted by unique letters [ ] in the same row.

Nearly one-third of all families report fair or poor child care during deployments of two or more weeks duration. There are no statistically significant differences across the groups, indicating that for a large proportion of all families, regardless of spouse employment status, child care during deployments presents a problem.

Approximately 12 percent of families rate their child care arrangement(s) as fair or poor. Although the proportion of single parents reporting fair or poor child care arrangements is higher than among families with either working or nonworking spouses, the difference is not statistically significant.[10]

Single parents are more likely than married parents to report being unsure about whether the Army would help with child care during a military conflict lasting six months or more. Overall, nearly 50 percent of all Army families were somewhat or completely unsure about whether the Army would help out in such a situation.

Table 4.4 shows child care ratings and problems as a function of the modal type of child care arrangement. It is interesting to note that the highest number of days lost to duty because of children and child care can be found among CDC users. The lowest number of days lost to duty are found among FDC users. This suggests that CDC care may be a less reliable source of care than FDC, even though the perception of many parents we interviewed in our site visits was to the contrary, as discussed in Section 3. The association between CDC care and more days lost is not surprising, given the more flexible nature of FDC, as discussed in Section 3. Moreover, these results are consistent with civilian data, as discussed below.

Table 4.4
Problems with Child Care by Child Care Arrangements

Child Care Arrangement
Type of Problem with Child Care All CDC Civilian FDC Non-relative Relative 1-5
Days lost from duty for child care reasons (N = 922) 3.1 7.0[a] 3.1 1.4[b] 2.2[b] 4.3 2.7
% reporting fair or poor child care during deployment (N = 293) 32.8 45.4 24.7 23.2 34.8 44.1 25.6
% rating child care arrangements fair or poor (N = 942) 12.4 16.9[b] 10.4 3.1[a] 10.8[b] 16.1[b] 12.3[b]
% unsure Army would help (N = 998) 49.0 41.4[b] 47.7 44.5 57.0[a] 49.3 44.4[b]

NOTE: Row 1 entries are means. Cell entries for rows 2-4 are percentages. Entries significantly different from one another at the 5 percent level are denoted by unique letters.

Similar patterns of responses are reported for the rating of both daily child care and child care during deployments. A significantly higher proportion of CDC users report fair or poor ratings than do FDC users. It is interesting that our site visits revealed that most parents strongly prefer the CDC, yet these data show that parents using FDC were less dissatisfied with their child care arrangements than were CDC users. This is, however, consistent with information from several FDC parents whom we interviewed who indicated that they had been opposed to FDC only until they actually tried it. Once they used FDC, they were very satisfied with it.

A few other results from Table 4.4 deserve mention. Parents using both CDC and FDC are less likely than the overall average to be unsure that the Army will help with child care during an extended military conflict. Each of these differences was significant.

Table 4.5 shows the reasons respondents offered for using civilian child care. Less than 25 percent of the families using civilian day care centers reported using such care because Army child care was unavailable to them either because it was full (and they were on a waiting list) or because it was not available for other reasons. A much larger proportion report using civilian care because of its characteristics (e.g., convenience of location, hours of care, or quality of care). It is interesting that a higher proportion identified location convenience as the reason for being in civilian day care than the proportion who chose it for its quality. This may mean either that parents care more about convenience of care than quality of care in general (Johansen, 1990), or that quality of care in military settings is acceptable, whereas location presents a problem.[11]

Table 4.5
Reasons for Using Civilian Instead of Army Child Care

Reason N %
Waiting to get into Army program 11 14.3
Army program needed is not available 7 9.1
Convenience of civilian service location 44 57.1
Convenience of civilian service hours 32 41.6
Quality of civilian service care 37 48.0
Cost of civilian service 22 28.6

NOTE: The total number of parents who used civilian care is 77. Respondents may have provided more than one reason.


Results of our secondary data analyses using the Survey of Army Families fielded in 1987 by RAND raise several interesting points.

First, most families use some type of child care, and most families using child care rely on more than one type of child care to meet their child care needs. Furthermore, whereas a large number of families report some use of Army CDCs, a much smaller number use it as their only or even modal type of care. Clearly, CDCs were not meeting most families' child care needs. Second, a significant amount of time is lost from duty because of problems with children and child care by all families, but the most time is lost by parents with children in CDC, and the least by families using FDC. Third, although most parents seem to express a strong preference for CDCs, a higher proportion of CDC users than FDC users report fair or poor ratings of their child care arrangements. Finally, among parents using civilian day care centers, the proportion reporting use of such care because of the characteristics of that care is much greater than the proportion reporting that they use civilian care because Army CDC was not available to them.

[1]The Survey of Army Families obtained child care data only for the youngest or the sole child in each study age group. We selected data for only the youngest child in families with more than one child.

[2]As noted above, the Army calls its quarters-based program family child care (FCC), but for consistency with the rest of the report, we will continue to use FDC to describe this care.

[3]This criterion was selected so that children who attend part-day preschool programs are included. Children often enroll in these programs for two or three mornings or afternoons a week, for two- to three-hour sessions.

[4]Since only about 300 of the 3100 spouses surveyed are also in the military, and of those fewer than 250 have accompanying children, the numbers were too small to conduct separate analyses for dual military families.

[5]The two questions were:

In the past month, how much time did you take off from duty ("your job" for spouses) for the following FAMILY reasons? (Please count time when you arrived late or left early, but do NOT include leave time.)

a. Caring for child(ren) on a daily basis (for example, supervision or discipline).
b. Other care of child(ren) (for example, sick child or visit to school).

[6]There may be seasonal differences in time lost that make our transformation somewhat questionable.

[7]Although the Army gives single-parent families priority for family-based care as well, parental preference for CDC care leads to greater exercise of the CDC priority.

[8]Numerous respondents told us that parents are disinclined to report this, as accompanying minors entitle parents to more generous living allowances. We were not able to estimate possible numbers.

[9]Comparable data from the civilian sector are not available. However, 7 percent of employed mothers in a 1987 Current Population Survey sample reported that they had lost time from work in the past month because of failed child care arrangements (U.S. Census Bureau, 1990). Over a year's time, this percentage would be substantially higher (Hayes et al., 1990).

[10]The quality of care in both FDC homes and CDCs since this survey was fielded has been improved, so these results should be interpreted with caution.

[11]Again, these results should be interpreted with care because of the changes that may have occurred in the military (and civilian) setting(s) since the Army family survey was fielded in 1987. Furthermore, these results are generalizable to the military at large only to the extent that the other military services are similar to the Army.

Previous section
Next section