A Glimpse of Child Hunger in Canada

4. Results

4.1 The Scope of the Problem

A total of 206 families reported ever experiencing hunger among 16639 total families. Although they are only 1.2% of the NLSCY sample, these families represent about 57,000 Canadian families.

4.2 Comparing Hungry Families with Other Families in Canada

The average age of children from hungry households was significantly older than other children 7.0 (s.d.=2.9) compared with 6.4 years (s.d.=2.9 years), p=.002. There was a significant difference in the number of siblings in the households reporting hunger although the difference was small [1.7 (s.d.=1.1) in hungry households versus 1.4 (s.d.=1.0) in non-hungry households, p<.000]. The ages of the Persons Most Knowledgeable (PMK) about the child did not differ significantly between those ever experiencing hunger (33.2 [s.d.=6.7] years) and those who did not.

The household composition and relationship of the PMK to the child was significantly different among families reporting ever experiencing hunger compared with those who did not. In particular, these children were significantly more likely to be living in a lone parent household with 58.3% of respondents from hungry families living in such an arrangement compared with 15.0% of other respondents (p<.000). Relatively few hungry families lived with two biological parents (21.7%) compared with those who lived in non-hungry families (65.0%), p<.000. The PMK was less likely to be the birth mother of the hungry child compared to other children (93.2% versus 96.8%, p=.003) and children were more likely to be living with a birth father than in other families, p<.000.

Families reporting ever experiencing hunger were significantly more likely to live in large urban areas with populations 500,000 or greater (28.2% versus 17.5%) compared with smaller urban areas or rural communities (p<.000) [Table 1]. These families resided throughout Canada without significant differences among the regions.

Table 1: Reported Occurrence of Hunger by Community Size*
Community Size	Occurrence of Hunger
	Ever n(%)	Never n(%)
Population 500,000+	58(28.2)	2877(17.5)
Population 100,000-499,000	37(18.0M)	3574(21.7)
Population 15,000-99,999	33(16.0M)	2988(18.2)
Population <15,000/Rural	78(37.5)	6994(42.6)
TOTAL	206(100%)	16433(100%)
* p<.000 M=marginal result Source: NLSCY

Anglo-Celtic (Canadian, English, Scottish, Irish) ancestry was most prevalent among those reporting both hunger and non-hunger states (43.2% of hungry families versus 48.4% of non-hungry families) and the difference was not significant (NS). The only ethnic group that was significantly associated with hunger were persons of aboriginal descent (North American Indian, Inuit, Métis). Compared with their prevalence of 3.9% in the total survey population, Aboriginal people were four times (16.0%M) more likely to report ever experiencing hunger than did households representing other ethnic groups, p<.000. Immigration status and being a member of other racially visible groups were no different among those reporting hunger than among those not reporting hunger.

Given that hunger is an indicator of extreme disadvantage, it is not surprising that income variables represented striking differences between hungry and non-hungry families. Household income falling below the low income cut-off (a standard Statistics Canada derived variable, [Human Resources Development Canada & Statistics Canada, 1996]), occurred significantly more often in hungry versus non-hungry households (p<.000). Income adequacy, a derived variable that ranged from lowest to highest, was also significantly different with 66.0% of hungry household incomes falling in the lowest or lower middle levels versus 19.9% for other families (p<.000).

Income from all sources varied significantly among hungry and non-hungry families. Income from wages and salaries was reported by 56.8% of families reporting ever experiencing hunger compared with 87.0% of other families reporting this source of income, p<.000. Hungry families were also significantly more likely to report income from social assistance/welfare (68.9% versus 14.7%, p<.000), and significantly less likely to report income from self-employment than other families (12.1%M versus 28.2%, p<.000). There was no difference in the reporting of income from employment insurance (18.4%M for hungry versus 20.7% for others, NS). As well, income from non-listed sources was reported significantly more often from hungry households than non-hungry households (14.6%M versus 8.5%, p=.002). This might be a measure of participation in the informal economy.

Wages and salaries were reported as the main source of income in 38.8% of hungry families while 56.8% reported social assistance or welfare as the main source. This indicates that the 'working poor' represent more than one-third of those who report ever experiencing hunger in the family.

The NLSCY also asked about earned income. The mean personal income of the PMK was significantly lower for those reporting hunger than those never reporting hunger ($12,816 versus $17,153, p<.000) although the earnings gap was most apparent in mean household income which was $21,255 for hungry families compared with $47,958 for other families, p<.000.

The PMK's main reported activity reflected income sources: 60.7% of respondents reported caring for the family while 23.8%M reported caring for the family and working. These activities were significantly different from other families where the percentages were 48.3% and 39.8% respectively, p<.000.

4.3 Health Status

While income differences were striking, so too were differences in health status among those reporting or not reporting hunger. PMKs were asked to rate their health from excellent to poor. PMKs of hungry families reported significantly poorer health than other PMKs (p<.000). Whereas 74.3% of other PMKs were reported to have very good or excellent health, 50.0% of PMKs from hungry families were reported to have this health status (p<.000) [Table 2]. Presence of a chronic health condition in the PMK (57.8%) was significantly different among families that reported ever experiencing hunger compared with all NLSCY families (42.7%, p<.000). Chronic health conditions that were reported significantly more often by the PMK in a hungry household included back problems [18.9% versus 10.5% (p<.000)] and migraine headaches [21.4% versus 10.3% (p<.000)]. PMKs from hungry households were significantly more likely to report activity limitations at home (18.0%M versus 7.2%, p<.000).

Table 2: Reported Health Status of PMK and Child by Occurrence of Hunger
	PMK		Child
Health	Hunger Ever* %	Never %	Hunger Ever* %	Never %
Rating	n=206	n=16328	n=206	n=16433

Excellent	14.6	33.7	44.7	59
Very Good	35.4	40.6	26.2	28.2
Good	35.4	21.2	19.4	10.7
Fair/Poor	14.5M	4.5	9.7U	2.1
* p,.000 M=marginal estimate U=unreliable estimate Source: NLSCY

The health situation of children in hungry households was equally troubling as measured by the PMK's reported health of the child (Table 2). Children of hungry families were reported to have significantly poorer health than other children (p<.000). Whereas 87.2% of other children were reported to have very good or excellent health, 70.9% of children from hungry families were reported to enjoy this health status. Ever having had asthma was the only health condition that differed between children of hungry and other families. While 12.6%M of non-hungry children were ever diagnosed with asthma; the risk was 1.8 times higher at 22.8% for hungry children (p<.000). Inhalant use by the child was 1.6 times (10.7%M versus 6.7%) higher in hungry children than in other children, confirming the on-going nature of the condition in both groups, and implying a relatively equitable access to asthma medications for children in either type of household.

Cigarette use in the households of hungry families was 1.7 times higher in PMKs reporting ever experiencing hunger and is likely related to asthmatic problems in children (Chen, Rennie, & Dosman, 1996): 49.0% of PMKs from hungry households reported daily cigarette use compared with a daily smoking rate of 29.7% by other PMKs (p<.000). Alcohol use in the past year was significantly more common among other PMKs (79.9% vs 69.9%, p<000) than among PMKs reporting family hunger. Alcohol abuse was not determined.

Stepwise multiple logistic regression was conducted to predict the risk of hunger in NLSCY families. The independent predictors of hunger were low household income; single parent status; main source of income is social services/welfare; child's health is fair/poor; PMK's poorer self-reported rating of health; ethnic group is aboriginal; and parent looking for work (Model 1). The odds of households with low income ever experiencing hunger was very high (96%). Single parenthood had almost a 50% chance of being associated with ever experiencing hunger.

Model 1: Logistic Regression Analysis of Risk Factors for Hunger Ever
Variable	Odds Ratio	95% C.I.	p-value
Main Income Source (Social Assistance)	2.1	1.4-3.1	0.0001
Household Income	0.96	.95-.97	0.0000
PMK's Health is Poor	7.6	3.5-16.5	0.0000
Number of Parents	0.49	.35-.70	0.0001
Child's Health Fair/Poor	3.2	1.3-3.0	0.0000
Ethnic Origin (Aboriginal)	2.0	1.3-3.0	0.0009
Main Activity - Looking for work	4.6	2.4-9.0	0.0000
Source: NLSCY

4.4 Profile of Hungry Families

The 206 families who reported that their child had ever experienced hunger because the family had run out of food or money to buy food were further analyzed. Male children represented 48.5% of these children. The educational attainment of mothers from hungry households was surprisingly diverse although it was significantly less than for other mothers (p<.000): 32.0% had less than high school education; 15.0% had completed high school; 28.2% had some post-secondary education; and 24.7% had a university degree or more education.

Among the 46.1% of homes that had a male spouse or male PMK, the father's educational attainment demonstrated lower achievement than achieved by the mothers: 51.1% had less than high school; 14.4% had completed high school; 20.0% had some post-secondary; and 14.4% had achieved a university degree.

4.5 Frequent Versus Occasional Hunger

Respondents were asked to indicate the frequency of their family's reported hunger (Figure 1). The choices ranged from more often than end of each month, to occasionally. Families were therefore divided into two groups—those reporting frequent hunger (35%) where hunger was reported at least every few months, and those reporting hunger occasionally.

A few significant differences were found between the frequently hungry and occasionally hungry groups. Children who were frequently hungry were significantly more likely to be living with their birth mother than were those who were occasionally hungry (98.6% versus 90.3%, p<.02). PMKs of frequently hungry households were also significantly more likely to report the presence of a chronic condition (68.1% versus 52.2%, p<.03). Among the small number of PMKs who were birth fathers, all reported occasional hunger.

Source of income differed significantly between those who reported frequent hunger and those who reported occasional hunger. Frequently hungry families were significantly more likely to depend upon social assistance/welfare than were occasionally hungry families (68.1% versus 37.3%, p<.000). While differences in personal income of the PMK were not significant, frequently hungry households' mean income of $18,110 was significantly less by $4,835 than the $22,945 mean income of occasionally hungry households (p<.02).

On stepwise multiple logistic regression, the independent predictor of frequent hunger was main source of income is social assistance/welfare (Model 2).

Model 2: Logistic Regression Analysis of Risk Factors for Frequent Hunger
Variable	Odds Ratio	95% C.I.	p-value
Main Income Source (Social Assistance)	5.1	2.5-10.4	0.0000
Source: NLSCY

4.6 Responses to Lack of Food

Of the 206 families reporting hunger, 34% reported that the parent skipped meals or ate less when the family had run out of food or money to buy food; 4.9%¹ reported that the child skipped meals or ate less; and 26.7% reported that they cut down on the variety of food that the family usually eats. There were no differences in responses to meal skipping by frequency of hunger.

The independent predictors of a parent skipping meals or eating less were: PMK reports a chronic condition, and main source of income is salaries or wages (Model 3).

Model 3: Logistic Regression Analysis of Risk Factors for Parent Eating Less as Response to Hunger
Variable	Odds Ratio	95% C.I.	p-value
Presence of Chronic Condition (PMK)	4.9	2.4-10.2	0.0000
Main Income Source (Wages/Salaries=1)	0.29	.11-.77	0.01
Source: NLSCY

4.7 Coping Strategies

Use of any of a number of coping strategies for the securing of additional food when the family had run out of food or money to buy food was also solicited. Seeking help from relatives or from the food bank was reported equally by 31.1% of respondents; seeking help from friends was utilized by 15.5%M; and seeking help from a social worker/government office was reported by very few people. Most respondents (64.1%) used one coping strategy with 18.4% reported using two coping strategies, and 17.4% using three or more strategies.

The only difference found among the coping strategies of one-parent versus two-parent households was that 38.3% of one parent households used a food bank versus 20.9% of two parent households (p=.008).

Coping strategies of frequently hungry and occasionally hungry families were similar as were child health status ratings and PMK self-reported health. Smoking did vary with 72.2% of PMKs reporting frequent hunger smoking daily or occasionally compared with a smoking rate of 50.7% in occasionally hungry respondents, p<01.

Households indicating the most commonly reported strategies—food bank use, and seeking help from relatives—were very different (Models 4 and 5). The independent predictors of food bank use were: residence in Ontario region; no reported activity limitation by PMK; and single parent status. The independent predictors of seeking help from relatives were: fewer numbers of children in the household aged 0-17 years; PMK does not report a chronic condition; there are two biological parents; and household income is not very low.

Model 4: Logistic Regression Analysis of Risk Factors for Food Bank Use
Variable	Odds Ratio	95% C.I.	p-value
Region (Ontario)	3.5	1.7-7.2	0.0008
PMK Activity Limitation - Home	0.16	.04-.55	0.0004
Two Biological Parents	0.38	.18-.83	0.015
Source: NLSCY

Model 5: Logistic Regression Analysis of Risk Factors for Seeking Help from Relatives as a Response to Hunger
Variable	Odds Ratio	95% C.I.	p-value
Number of Children Aged 0-17 in Household	0.42	.28-.63	0.0000
Presence of Chronic Condition (PMK)	0.41	.19-.87	0.02
Middle Level Income (vs Higher)	0.05	.015-.52	0.01
Two Biological Parents	3.2	1.4-7.1	0.004
Source: NLSCY

The independent predictors of seeking help from friends were: low, rather than very low, income; and mother is unemployed (Model 6).

Model 6: Logistic Regression Analysis of Risk Factors for Seeking Help from Friends as a Response to Hunger
Variable	Odds Ratio	95% C.I.	p-value
Income Level (low vs very low)	0.28	.10-.77	0.01
Mother Unemployed	8.7	2.6-28.7	0.004
Source: NLSCY

¹Unreliable estimate

Contents

A Glimpse of Child Hunger in Canada - October 1998