Asia Pacific J Clin Nutr (1994) 3, 69-82

Adiposity, dietary and physical activity patterns in ethnic Chinese youths: a cross-country comparison of Singaporean Chinese and Chinese Americans

M.C. Wang¹, T.F. Ho², G. Block¹, M. Lee³, J. Anderson⁴ and Z.I. Sabry¹

^1. School of Public Health, University of California, USA;
^2. Dept of Physiology, National University of Singapore, Singapore;
^3. Dept of Epidemiology & Biostatistics, University of California, San Francisco, USA;
^4. Dept of Anthropology, University of California, Berkeley, USA.

During the last decade, childhood obesity has been on the increase in Singapore and many newly industrialized Asian countries. We compared the mean body mass index (BMI) and triceps skinfold (TSF) values, as well as the dietary and physical activity patterns of Singaporean Chinese and Chinese American youths. Chinese Americans had a higher mean BMI but a lower mean TSF than Singaporean Chinese. Dietary comparisons suggest that Singaporean Chinese ate fish and grain products more often than Chinese American youths, while Chinese American youths consumed processed meats, dairy products and snack foods more frequently. Mean frequency of consumption of low fat, traditional Chinese foods such as rich porridge was higher among the Singaporean Chinese, while typical 'American' foods including cheese were consumed more often among the Chinese Americans. Certain food items that were more 'neutral' in terms of their cultural identity, such as carbonated drinks, cookies and bread were consumed with the same mean frequencies in both cohorts. In terms of physical activity, Singaporean Chinese youths, on average, spent more time in sedentary activities, less time sitting, and more time in light or moderate activities. The mean time spent on vigorous activities per day was only one hour in both cohorts. Our study suggests differences in body fat distribution and composition, as well as in dietary and activity patterns, between Chinese American and Singaporean Chinese youths. There is a need to develop obesity indicators that are appropriate for the specific populations involved, and to carefully investigate environmental influences on childhood obesity.

---

Introduction

Childhood obesity has recently been recognized as an emerging problem in many newly industrialized Chinese societies^1-4. For example, in Singapore, in 1976, only 2% of Primary Six schoolchildren (mostly 12-year-olds) were identified as obese, based on a relative weight of (120% of standard weight-for-height from the Harvard growth standards. In 1983, the corresponding rate was 12%¹. Recent data from the Ministry of Health in Singapore shows that the prevalence of obesity among Primary Six children was 19% for boys and 12% for girls in 19905. In Hong Kong, the prevalence of overweight among a selected g 1000 roup of adolescents was found to be about 3-4%². Table 1 summarizes selected published studies on childhood obesity in ethnic Chinese populations. Comparison of data across these studies is difficult since each study uses a different criteria for defining overweight.

Table 1. Studies of childhood obesity in ethnic Chinese populations.

Country	Population	Variables measured	Analysis	Results
Singapore1	705 511 school children measured from 1976-83 (51% boys, 49% girls): 346,208 6-7 yr olds 359,303 11-12 yr olds	Weight, standing height, relative weight (using the Harvard growth standards for ideal weight-for height).	Obesity defined as ( 120% of ideal weight-for-height Computed prevalence by year, gender, and age	Annual prevalence: Boys: 1.4% (1974); 8.2% (1983) Girls: 1.0%(1974); 8.5% (1983) 6 7 years:1.4%(1976); 3.1% (1983) 11-12 years: 2.2%(1976); 12.1%(1983)
Hong Kong2	1535 adolescents aged 14-27 years (82% aged 16-21 years) from two secondary schools and one commercial school	Height, weight BMI (weight/height2): physical activity assessed by questionnaire; parents' height and weight as reported by the subjects.	A BMI of 25 is the cut-off for defining overweight Computed percent overweight by age and sex Computed energy expenditure by converting hours spent in various activities to metabolic equivalence Used univariate analysis to assess associations of activity and parents' BMI with subject' s overweight	1) Per cent overweight: 14-15 years 4.1% 16-17 years 2.6% 18-19 years 2.5% 20+ years 5.3% All age groups: 3.5% Males: 4.0% Females: 3.0% 2) Time spent in exercise and activity score were not related to overweight 3) Correlation between parents' BMI and subject' s BMI was not different from zero.
Taiwan3	20 677 individuals aged 3-70 years (49% males, more than 70% (20 years) were measured in 12 districts selected by stratified random sampling	Mid-arm circumferences (TSF)	Computed prevalence based on triceps skinfold (TSF) and > 120% of average body weight for each age gr 1000 oup, for 10-15 year old boys and girls (1986-88) Compared prevalences with a previous survey ( 1980-82)	1) Prevalence (only results for 10 and 15 year olds are given here): Based on 120% of average weight:   Boys Girls   10yr 15yr 10yr 15y 1980-82 8.4 10.2 4.6 9.2% 1986-88 14.4 10.5 14.1 9.7% Based on TSF: 1986-88 21.5 18.6 10.0 14.8% 2)Mean BMI (1986-88) Boys (8-19yr): 19.0-27.2 kg/m2 Girls(8-19yr): 18.6-24.5kg/m2 Mean TSF (1986-88) Boys(8-19yr): 10.3-13.2mm Girls(8-19yr): 12.6 19.9mm
China4	4314 boys and girls aged 7-18 years from the Northern Chinese	BMI, subscapular and TSF	Cut-offs for overweight defined by standard deviation from predicted values of BMI and the sum of measured skinfolds (SF) 1000	1) Prevalence of overweight (1983): Criteria: BMI Sum SF Both Boys: 7.1 5.6 2.9% Girls: 9.0 6.0 4.0 2)Range of mean BMI (1983) Boys: (7-18yr): 15.0-20.0 kg/m2 Girls :(7-18 yr): 14.7-20.1 Range of mean sum SF (1983) Boys: (7-18yr): 12.9-16.2mm Girls: (7-18yr): 15.1-29.5

The causes of obesity are multi-factorial. Both genetic and environmental factors have been implicated⁶. Among the environmental causes, diet and physical activity have been most widely studied. This paper will focus on these environmental factors.

Dietary patterns among the Chinese

Although the Chinese diet is generally considered low in fat, rapid economic development in many Chinese-dominated newly industrialized nations have resulted in increasing meat and animal fat consumption. In Singapore for example, where 78% of the population is of Chinese ethnicity, food availability data from the Food & Agricultural Organisation (FAO)⁷ indicates that fat levels doubled between 1961-63 (41 grams) and 1986-88 (81 grams). A survey of 40 Chinese households in Singapore carried out in 1984-85 revealed that dietary patterns were influenced by affluence and the primary language spoken⁸. In particular, more affluent households tended to purchase more red meat/offal, poultry, fruit, eggs and vegetables other than green leafy, while lower-income households were consuming more eggs and milk in 1984 than in 1970. The affluent households also tended to use corn or soybean oil as opposed to lard or other vegetable oils, and to eat high fibre biscuits more than the less affluent households. There were also some indications that more English-speaking than Chinese-speaking households had made changes in their diet during the last five years.

Similarly, in Taiwan, fat availability per capita increased from 86 grams (28% of calories) in 1980 to 118 grams (36% of calories) in 1986⁹. Finally, in China, fat availability per person in both rural and urban regions increased by 180% from 16 grams or 8% of calories (1961-63) to 44 grams or 15% of calories (1986-88)⁷. Increases in the availability of processed foods that are high in fat content, and increased consumption of meat^7,9 ar 1000 e likely explanations for the increases in fat intake seen in these societies.

In the United States, the process of acculturation can be expected to lead to major changes in the dietary patterns of immigrants from China, Hong Kong, Taiwan, Singapore and elsewhere. Indeed, several studies have documented changes in the food habits of various Chinese immigrant groups in the United States and Canada^10-13, Yang & Fox, in 1979, administered a questionnaire designed to assess changes in food habits in a group of first generation immigrant Chinese adults living in Nebraska¹⁰. They reported more people following 'American style' breakfast and lunch patterns. Dinner, for most subjects, however, generally remained Chinese-style. They also found that immigrants from Taiwan made fewer changes in food habits than those from China or Hong Kong.

Physical activity

There is a dearth of information on the physical activity patterns of ethnic Chinese children and youths. In a survey of 887 youths, all over 14-years-old, from two schools in Hong Kong, it was found that 25% and 40% of the respondents did not report having recently taken any moderate or vigorous exercise, respectively. The relationship between activity and overweight was weak and statistically insignificant². In a population-based cross-country study of colorectal cancer of Chinese Americans and of Chinese living in China, both higher energy expenditure and intakes were observed among the Chinese in China. The risk of colorectal cancer was observed to increase the duration of exposure to a sedentary lifestyle and high saturated fat consumption. Differences in the rates of colorectal cancer between the Chinese Americans and the population in China could be explained by differences in these lifestyle factors¹⁴.

The purpose of this paper is to compare adiposity, using body mass index (BMI) and triceps skinfolds (TSF), as well as dietary and physical activity patterns, between ethnic Chinese youths living in Singapore, and their age and sex counterparts living in California.

Methods

Recruitment

Recruitment efforts were directed at high school and college-aged individuals. In Singapore, a total of 280 subjects, aged 17-22 years, were voluntarily recruited from among first year female students of the National University of Singapore (103 females), Army recruits (117 males) and students at a polytechnic (40 females and 20 males). Males were not recruited from the university because in Singapore, young men usually attend two years of military service before entering university. The Army recruits in the cohort were just about to begin their two year stint of military service and had not yet undergone rigorous physical training. Only recruits with a t least a certain level of secondary school education were selected. At the polytechnic, where there was an obesity screening program, special efforts were made to recruit obese individuals. In California, 1 13 subjects aged 16-22, were initially recruited from the Bay Area of California. Recruitment sources were the University of California at Berkeley, several Chinese social clubs at high schools and the Chinese School in San Jose.

The eligibility criteria required the individuals to:

• have lived in the country of study for more than 8 years
• have no medical problems that would predispose them to obesity or growth disorders, such as congenital hip dislocations, thyroid problems, or congenital heart disease.
< 1000 /ul>

In the analysis of the growth data, two additional criteria were imposed: subjects had to be of Southern Chinese origin, and be at least 17-years-old. These criteria were instituted for the following reasons. Most Singaporean Chinese originated from Southern China, and growth differences are known to exist between Northern and Southern Chinese¹⁵. None of the Singaporean participants was younger than 17 years.

Data collection and analysis

The protocol for the conduct of this study was approved by the Committee for the Protection of Human Subjects at the University of California. In accordance with its policies, informed consent to participate was obtained from every subject.

Standard protocol for measuring height, weight and skinfolds was followed¹⁶. A detailed description of the protocol is given in another paper¹⁷. Mean and median values were computed, and statistical differences between the two cohorts were assessed using the two-tailed student's t-test. The means and medians were also compared with the reference population in the United States¹⁸.

Dietary patterns of the participants were determined using a food frequency questionnaire developed specifically for Chinese American and Singaporean Chinese youths. Nutrient values of each of the 120 listed food items were derived from one of the following food composition tables: 1. Handbook No. 8 (United States Department of Agriculture); 2. Journal of Food Composition and Analysis (Special issue: Chinese Food Composition Tables; 3(3,4)), 1990; 3. Nutrient composition of Malaysia Foods (ASEAN Sub-committee on protein: Food habits research and development, 1988; 4. Guo-ming ying-yang chindao so chih. National Nutrition Guidelines (Ministry of Health, Republic of China), 1991.

The validity of this dietary questionnaire, and the methodology for analyzing the dietary data have been described elsewhere¹⁹. Mean macronutrient intakes, and mean frequency of consumption of selected food groups were estimated, and compared between the two cohorts.

Physical activity patterns were assessed from a brief questionnaire that asked for estimations of time spent on vigorous, light to moderate, sitting down and sedentary activities during typical week days and weekends, in the past year. Somewhat similar questions have been applied to the Chinese American adult population in an epidemiologic study of colorectal cancer¹⁴. The subjects were also asked if, when compared to others of their age, they considered themselves 'very active', 'active', 'somewhat active' or 'sedentary' .

Results and discussion

Cohort differences in height, weight, and adiposity indices

The distributions of height, weight, body mass index (BMI) and triceps skinfold (TSF) of the Chinese American and Singaporean Chinese youths are shown in Figure 1 (a) males and (b) females. Means and medians are presented in Table 2.

Figure 1. Distribution of body mass index and triceps skinfolds in (a) male and (b) female youths. [Solid black] = Singaporean Chinese [Stripes] = Chinese American.



Table 2. Comparison of means and medians for height, weight, BMI and triceps skinfold of Singaporean-Chinese and Chinese-American youths.

MEASURE	FEMALES			MALES
Group	Singaporean Chinese (n+108)	Chinese in California (n=30)	Americans (NHANESII)a	Singaporean Chinese (n= 123)	Chinese in California (n=26)	Americans (NHANES II)a
HEIGHT (cm)
Mean± SD	159.4± 5.3	159.7± 5.4	163.5± 5.6	170.6± 5.6*	173.6± 5.6*	176.5± 6.7
Median	159.5	158.4	1163.7	170.7	173.7	176.9
WEIGHT(kg)
Mean± SD	51.5± 9.0	54.2± 6.0	60.2± 11.0	62.5± 11.6**	70.3± 9.4**	71.7± 11.6
Median	49.5	54.6	57.1	61.1	69.5	69.5
BMI(m/kg2)
Mean± SD	20.2± 3.1*	21.2± 2.2*	22.6± 4.2	21.4± 3.8*	22.6± 2.8*	23.5± 3.6
Median	19.5	21.3	21.6	20.8	22.5	23.0
TSF(mm)
Mean± SD	21.5± 6.9*	18.9± 4.7*	20.7± 8.6	13.1± 6.6	11.9± 4.0	11.6± 6.5
Median	20.8	17.9	19.0	11.4	11.2	10.0

^aNajar and Rowland (1981)¹⁸
*P<0.05(two-tailed t-test): **P<0.01 (two-tailed t-test)

The Chinese American females are almost 3 kg heavier, on average, than the Singaporean Chinese females. However, this difference is not significant and may be due to the smallness of the Californian Cohort. There was no difference in height between the two female cohorts. Among the males, the Chinese Americans are taller by about 3 cm (P<0.05), and heavier by almost 8 kg, than their counterparts in Singapore (P<0.01).

The Chinese American females have a higher mean body mass index (P<0.05) than the Singaporean Chinese. However, their mean triceps skinfold is, unexpectedly, lower (P<0.05). The mean TSF of the Singaporean Chinese females was 21.5± 7 mm, as compared to only 18.9± 5 mm in the Chinese American females. The Chinese American males also have a higher mean BMI value than their counterparts in Singapore (P< 0.05), and a lower mean TSF value, but the difference in TSF does not achieve statistical significance (P=0.3). The median TSF of the Chinese American youths is almost the same as that of the Singaporean Chinese.

When compared with US growth reference data from the National Health and Nutrition Examination Survey¹⁸, mean and median values of BMI of the Chinese Americans are lower, but TSF values are different only among females.

Correlations between BMI and TSF

Spearman's rank correlation coefficients between BMI and TSF are shown in Table 3. These coefficients, rather than Pearson's correlation coefficients, were used since the numbers were relatively small and the distributions of BMI and TSF were not normal. It is noted that the correlations for both males and females are lower in the Chinese American than in the Singaporean Chinese cohort. They are also lower than Pearson correlation coefficients for American youths of European origin²⁰.

Table 3. Spearman's correlation coefficients between body mass index (BMI) and triceps skinfolds (TSF) in Singaporean Chinese and Chinese-American youths.

	Males	Females
Singaporean Chinese, 17-22years, 1992 (n=231)	0.78 (P<0.001) (n=123)	0.64 (P<0.001) (n=108)
Chinese American, 17-22 years, 1992 (n=56)	0.65(P<0.001) (n=26)	0.51(P<0.001) (n=30)
American youths, aged 16-18 years	0.72	0.74

in the Singaporean Chinese, particularly among the females. These observations support the conclusion of previous researchers that environmental factors may have influenced the body composition of the two cohorts^21,22, and that these influences may interact with gender²³. Johnston et al. in an extensive review of the use of equations for predicting estimates of body composition, concluded that such equations can, at best, be used only for genetically and 'environmentally' similar groups²¹. Recently, Hazuda and co-workers²³ found that socio- economic status (SES) and structural assimilation (entrance into the social structure of the host society) predicted body fat distribution in Mexican Americans, and that these associations were different for females and males. In particular, SES was positively associated with waist-to-hip ratio in men but inversely associated with skinfold thickness and waist-to-hip ratio in women. In women only, structural assimilation was inversely associated with BMI and subscapular-to-triceps skinfold ratio (an index of truncal or central body fat distribution). Neither cultural or structural assimilation was related to obesity or body fat distribution in men.

Further, although both BMI and TSF have been widely used to assess obesity in the United States, their validity as indicators of obesity in ethnic Chinese populations must be questioned¹⁷. Garn et al. have noted that since weight is the numerator in BMI, BMI may reflect lean and fat tissue to a comparable degree. Furthermore, BMI may be influenced by relative sitting height (sitting height/stature) to the extent that shorter-legged individuals may have BMI values that are higher by as much as 5 units²⁰. The latter is especially relevant in the comparison of Chinese populations living in different environments. Studies of Chinese children and youths in Hong Kong have shown a secular trend in the relative sitting height of Chinese youths suggesting that body proportions may be influenced by changes in the environment²⁴. Thus, differences in the environment of the Chinese Americans and the Singaporean Chinese may contribute to differences in body composition and body proportion, and suggest a need for defining appropriate standards of obesity for these populations. In an earlier paper^l7, we suggested that the development of these standards should be based not only on population-specific reference data but also on an understanding of how these standards reflect body fat distribution and their association with morbidity and mortality.

Another relevant observation is that the correlations between TSF and BMI were different between the Chinese Americans and the Singaporean Chinese, and were lower in the Chinese American cohort. This suggests that there may be less homogeneity in the 'environment' (including cultural practices relating to food) of the Chinese Americans, almost all of whose parents are first generation immigrants to the United States. Thus, it is postulated that immigrant populations require a period of time to adjust their lifestyles and the rates of acculturation are subject to individual variation.

Acculturation involves continuous and intense contract between two previously autonomous cultures, and often leads to changes in one or both systems^25,26. One obvious change during the a 1000 cculturation process that may impact on growth relates to dietary patterns.

Dietary patterns

Macronutrient intakes. The distribution of the intakes of calories, fat, protein and carbohydrate of both cohorts are shown in Figures 2-5. Mean and median intakes of their macronutrients are given in Table 4. The mean caloric and fat intakes of the Singaporean cohort (2500 kcal and 73 grams, respectively) are lower but not inconsistent with the availability of calories and fat per person for the nation (Food availability is the total quantity of a food produced, imported, and in stock minus the amount exported, put to industrial, or other non-food consumption use, fed to livestock or used for seed, and lost during storage or transportation). Based on FAO statistics, energy availability amounted to 2882 kcal and fat availability to 80.9 grams in 1986-88 in Singapore⁷.

Figure 2. Distribution of energy intake (k/cal).



Figure 3. Distribution of fat intake.



Figure 4. Distribution of protein intake.



Figure 5. Distribution of carbohydrate intake.



Table 4. Means and medians of macronutrient estimates.

				FEMALES			MALES
Nutrient	Singapore (n= 145)	California (n=72)	Sig	Singapore (n=89)	California (n=34)	Sig	Singapore (n=56)	California (n=28)	Sig
Energy (kcal)
Mean	2523± 1128	2031± 840	§	2081± 698	1679± 574	§	3225± 1317	2459± 920	§
Median	2363	1856	--	2070	1666	--	3251	2326	--
Fat (g)
Mean	72.5± 39.5	59.2± 33.8	†	57.9± 25.0	46.6± 19.7	§	95.6± 46.9	74.5± 40.9	‡
Median	663	50.6	--	55.9	39.8	--	91.5	59.0	--
Protein (g)
Mean	3.5± 60.4	75.6± 32.4	§	93.9± 35.9	63.8± 23.6	§	44.7± 76.7	90.3± 35.7	§
Median	105.5	68.1	--	90.5	60.7	--	138.6	84.1	--
CHO (g)
Mean	354.8± 155.4	300.7± 122.0	§	298.2± 106.4	254.1± 98.6	NS	444.7± 178.2	357.4± 125.1	‡
Median	334.2	280.5	--	288.6	232.2	--	406.9	333.0	--
%fat
Mean	25.2± 5.2	25.8± 6.3	NS	24.7± 5.2	25.1± 6.1	NS	26 1± 5.3	26.8± 6.5	NS
Median	25.7	25.2	--	24.7	25.1	--	26.1	25.6	--
% protein
Mean	18.0± 4.3	15.0± 3.2	§	18.2± 4.0	15.2± 3.4	§	17.7± 4.8	14.8± 3.1	§
Median	17.1	14.9	--	17.4	l5.l	--	16.9	14.5	--

*P-value computed using the t-test; † P<0.05; ‡P<0.01; § P<0.001.

Mean and median estimates of energy, protein, fat and carbohydrate intakes are all higher in the Singapore cohort than in the California cohort. While it may seem that the higher macronutrient intakes in the Singapore youths is biologically congruent with the higher adiposity level in this population, as indicated by higher triceps skinfold, there are at least two methodological reasons why this observation cannot be supported.

First, the validation study suggests that the dietary questionnaire may estimate nutrient intakes somewhat differently for the two cohorts. This may be partly due to the higher frequency of consumption of Chinese mixed (stir-fried) dishes among the Singaporean Chinese which makes it difficult to assess the proportion of meat/fish in the 'midst' of vegetables. For example, a person who consumes small amounts of chicken and pork in two separate mixed dishes at one meal is probably more likely to over-estimate the consumption of meat than an individual who is served a 3 oz piece of steak. Certainly, the protein intakes of the Singapore cohort are very high, based on current recommended allowance for protein for the US population²⁷, and also on protein availability figures for Singapore. In 1981-83, FAO estimated that 74.5 grams of protein were available per person. The mean protein estimate for the Singaporean cohort studies was 114 grams. Another reason for the high macronutrient estimates among the Singaporean Chinese may be the use of a plate size (10") that was larger than that commonly used by food stall holders (7-8"), for photographing portion sizes shown in the diet questionnaire. Although subjects were shown an actual sized picture of the plate pictured in the questionnaire, they may have difficulty in assessing the actual portions of food consumed. Second, the validation study of the dietary questionnaire shows that only mean estimates of fat agree with corresponding values from food records.

On the other hand, the findings that Chinese American females consumed a lower mean caloric intake than Singaporean Chinese may also reflect a greater consciousness of dieting and the value of thinness in American society^28-30. The generally higher socio-economic level of the Chinese American cohort may also partly explain their lower caloric intakes. Although it is commonly assumed that the more wealthy eat more, this is true perhaps only in developing societies where food scarcity is a problem. In developed countries, an inverse relationship between caloric intake and socio-economic status has been observed^31,32.

Frequency of consumption of selected food items. The dietary questionnaire elicited information relating to both the frequency of consumption of 120 individual foods and the usual 1000 portions consumed. Since there are apparent problems with the estimation of portion size particularly in the Singapore cohort, a comparison of only the frequencies of consumption of individual food items as well as food groups was made. The food groups were arbitrarily selected to provide a system of classification that reflects broad groups within which individuals tend to substitute food items. These groups were grains (including breads, rice and noodles), fish and shellfish, meats (including poultry), processed meats, dairy products, vegetables, fruit and snacks (including deserts). Sixteen food items which could not be classified into any of these groups were analyzed individually.

Singaporean Chinese youths tended to consume more fish and shellfish products, and grains, and less processed meats, dairy products, and snacks. There were no significant differences in the consumption of meat (including poultry), fruits, and vegetables (see Figure 6). A comparison of individual food items showed significant differences in the mean frequencies of consumption of 68 of the 120 food items. Table 5 (a) provides a list of the items for which the mean frequencies of consumption differed by at least 3 times between the two cohorts, in descending order of magnitude of difference. In contrast, Table 5 (b) provides a list of food items for which the difference in mean frequency of consumption is no more than 1.5-fold.

Figure 6. Mean frequency of consumption of selected food groups.



Table 5(a). Comparison of the mean frequency of consumption of selected food items between Chinese-American and Singaporean Chinese youths.

1000 1000

Food items for which the mean frequency of consumption (per month) differs between the two cohorts, by a factor of more than three.
Food item	Cohort	Mean	SD	Sig1	Ratio2
Mangosteens	Calif3 Spore4	0.0026 0.3280	0.0140 0.8010	P<0.01	126.2
Kaya	Calif Spore	0.0671 4.4441	0.0300 8.5970	P<0.001	66.2
Rambutans	Calif Spore	0.0173 0.6312	0.1060 2.4010	P<0.05	36.5
Jackfruit	Calif Spore	0.0200 0.4237	0.1280 1.1990	P<0.01	21.2
Guava	Calif Spore	0.0306 0.6241	0.0870 1.5930	P<0.01	20.4
Fried fish ball	Calif Spore	0.2729 3.0268	0.4560 4.1830	P<0.001	11.1
Papaya	Calif Spore	0.2295 2.0120	0.6400 3.2200	P<0.001	8.8
Natural cheese	Calif Spore	5.1439 0.6165	7.9120 1.5060	P<0.001	8.3
Condensed milk	Calif Spore	1.4613 11.8610	4.6290 22.1000	P<0.001	8.1
Milk, low fat	Calif Spore	28.4148 3.6637	27.0190 8.3300	P<0.001	7.8
Chinese mustard	Calif Spore	0.8916 6.0909	1.5420 10.9650	P<0.001	6.8
Cereals	Calif Spore	12.3097 1.9003	13.8540 4.4940	P<0.001	6.5
Fishball, plain	Calif Spore	0.8419 4.8747	1.4670 6.6630	P<0.001	5.8
Liver, any kind	Calif Spore	0.1974 1.0813	0.4030 1.9690	P<0.01	5.5
Rice porridge	Calif Spore	1.5239 7.6920	2.2700 10.5510	P<0.001	5.0
Fresh fruit juice	Calif Spore	19.8639 4.3718	25.1620 6.0940	P<0.001	4.5
Soya milk, fresh	Calif Spore	0.9839 4.1397	3.9620 6.3410	P<0.001	4.2
Beef, not mixed	Calif Spore	4.5452 1.0821	5.7950 2.1660	P<0.001	4.2
Hot chocolate	Calif Spore	1.8761 7.8571	3.2590 10.5190	P<0.001	4.2
Rice noodle, fried	Calif Spore	1.0787 4.4781	1.6200 5.709	P<0.001	4.2
Ham, deli meat	Calif Spore	9.6219 2.4794	10.0280 4.2090	P<0.001	3.9
Corn	Calif Spore	4.5077 1.1651	5.060 2.2500	P<0.001	3.9
Durians	Calif Spore	0.1948 0.7494	1.1280 2.2440		3.9
Chinese doughnut	Calif Spore	0.3981 1.5298	0.7410 2.9260	P<0.01	3.8
Beef and veg. mixed	Calif Spore	6.0103 1.8062	6.2810 5.5400	P<0.001	3.3
Processed cheese	Calif Spore	5.8000 1.7653	8.1690 3.6700	P<0.001	3.3
Dried shrimp	Calif Spore	0.6052 1.9294	1.1420 4.1800	P<0.05	3.2
Apple juice	Calif Spore	6.0742 1.9142	10.4450 3.8690	P>0.001	3.2
Dried, salted fish	Calif Spore	0.5806 1.8250	2.5650 3.1000	P<0.01	3.1
Peaches	Calif Spore	3.7942 1.2151	6.0430 2.8720	P<0.001	3.1
Veal, not mixed	Calif Spore	0.1826 0.5693	0.3900 2.5030		3.1
Fish, any kind	Calif Spore	3.9471 12.1311	5.2290 13.6830	P<0.001	3.1
Bananas	Calif Spore	10.3198 3.3796	13.2580 4.7900	P<0.001	3.0

Table 5(b). Similarities in the mean frequency of consumption of selected food items of Chinese-American and Singaporean Chinese youths.

Food items for which the mean frequency of consumption (per month) differs between the two cohorts, by a factor of no more 1000 than 1.5.
Food item	Cohort	Mean	SD	Sig1	Ratio2
Carrots	Calif3 Spore4	6.4606 4.3756	7.3290 5.4620	P<0.05	1.5
Non-carbonated drinks	Calif Spore	14.2839 9.8432	20.1880 16.9620		1.4
White potatoes	Calif Spore	4.5129 3.2152	5.5430 6.4690		1.4
Jams	Calif Spore	5.3794 3.8342	7.6180 7.7840		1.4
Eggs	Calif Spore	9.8465 13.7561	9.7350 22.6530		1.4
Watermelon	Calif Spore	2.2152 3.0240	< 1000 font size="2">4.5000 3.9090		1.4
Nuts	Calif Spore	2.5981 1.9352	3.9270 2.7480		1.3
Pineapple	Calif Spore	0.9276 1.2314	1.3680 2.4660		1.3
Milk, whole	Calif Spore	5.7677 7.544	17.5330 12.6940		1.3
Peas, long beans	Calif Spore	5.2723 4.0845	5.5380 5.1500		1.3
Peanut butter	Calif Spore	4.5903 3.6013	6.6960 6.7910		1.3
Cabbage	Calif Spore	4.3813 5.5840	3.7480 7.6880		1.3
Citrus fruits	Calif Spore	10.3226 8.1251	8.7110 9.9340		1.3
Ice Cream	Calif Spore	4.3839 3.5301	4.9220 5.0870		1.2
Canned meats	Calif Spore	2.7490 2.3201	7.2350 2.9530		1.2
Asparagus	Calif Spore	0.9723 1.1365	1.9020 2.9040		1.2
White bread, roll	Calif Spore	16.1703 13.8469	21.7280 11.2780		1.2
French fries	Calif Spore	3.5019 3.0242	3.7590 4.0000		1.2
Pears	Calif3 Spore4	3.2474 2.8427	4.0350 4.0940		1.1
1000 Glutinous rice	Calif Spore	0.8755 0.7769	1.5980 1.8430		1.1
Chicken and veg. mixed	Calif Spore	5.8439 5.2507	5.2490 9.2890		1.1
Chicken, not mixed	Calif Spore	4.9258 5.3425	3.9100 5.6740		1.1
Apples	Calif Spore	8.6747 8.0125	8.2840 8.5890