Australian and New Zealand Journal of Public Health 2002; 26 (5) : 473-8.
A Comparison of Self-Reported and Measured Height,
Weight and BMI in Australian Adolescents
Zaimin Wang, Carla M. Patterson
Centre for Public Health Research, Queensland University of Technology, Queensland
Andrew P. Hills
School of Human Movements Studies, Queensland University of Technology, Queensland
Abstract
Objective: To explore the relationship between self-reported weight and height to actual weight and
height in older Australian adolescents.
Method: Weights and heights of 572 adolescents aged 15-19 years who participated in the 1995
Australian National Health Survey (NHS) and National Nutrition Survey (NNS) were examined.
Results: Self-reported heights were significantly higher than measured heights in participants. There
were no differences in the accuracy of self-reported heights among the adolescents by gender. Selfreported weights were significantly lower than measured weights among both boys and girls (p<0.01).
There were no differences in the accuracy of self-reported weights among the boys and girls.
Differences between actual weight and self-reported weight were significantly greater for overweight
or obese adolescents compared with normal/underweight adolescents (p<0.01). The use of self-reported
weight and height resulted in the correct classification of overweight or obesity in 69% boys and 70%
of girls.
Conclusions: There was no significant gender difference in reporting weight and height in older
adolescents. Bias in reporting weight and height was much higher in overweight or obese adolescents
than normal underweight/adolescents implications: The percentage of misclassification of overweight
or obesity from self-reported data in this study was 31% for boys and 30% for girls, respectively.
Therefore, the self-reported weight and height of older adolescents needs to be more cautiously utilised.
Efforts to improve the accuracy of self-reporting in older adolescents are needed if this measure is to be
reliable.
Weight and height are common indicators used in both clinical and public health settings to evaluate
the health status of children and adolescents as body weight or relative weight (weight adjusted for
height) is considered to be reasonably correlated with body fatness.1-9 Self-reported heights and weights
are of-ten used in place of actual measurement to assess the health and nutritional status of adolescents
in population surveys. Numerous studies have focused on the accuracy and reliability of self-reported
weights and heights in adults.10-16 Some authors indicated that the correlation typically was over 0.90.1017
Australian investigators examined the accuracy of self-reported weight and height in adults based on
the National Health Survey (NHS) and National Nutrition Survey (NNS) in 1995 and found that 64%
of males and 47% of females were classified as overweight or obese using measured weights and
heights, compared with 52% and 30%, respectively, by self-reporting.18 Other Australian studies in
adults have suggested that the mean difference between self-reported and measured data was a 0.2-2.0
cm over-estimation for height and 0.4-3.0 kg under-estimation for weight in adults.19,20
A limited number of studies have reported on adolescent populations. In American adolescents,
Strauss reported that correlations between actual and reported heights and weights ranged between
0.82-0.94 in a cross-sectional nationally representative sample. 21
The findings of a study on a small group of young adolescents from a well-educated, high-income
community in the US suggested that the correlation was from 0.77 to 0.98.22 An investigation on a subsample of Italian high school students also indicated that correlation was over 0.90.23 There are,
however, few studies on the correlation between the actual and reported heights and weights and the
influences on reporting in older adolescents, i.e. those 15-19 years of age. In order to clarify the
accuracy of self-reported weight and height and examine the influences on reporting in older
adolescents, data were analysed from the 1995 Australian National Health Survey (NHS) and the
Australian and New Zealand Journal of Public Health 2002; 26 (5) : 473-8.
National Nutrition Survey (NNS). Data on individual self-reported height and weight from the NHS
and measured heights and weights from the NNS were merged for the Australian adolescents aged 1519 years.
Methods
The NNS and NHS were conducted from January 1995 to January 1996. For the NHS, a base sample
size approximating one-third of 1% of the population was initially chosen. Some 16,400 private and
non-private dwelling units distributed across all States and Territories were initially selected, bringing
the total active sample size to 53,800 people. The study collected comprehensive information on the
health status of Australians, their use of health services and facilities, and health-related aspects of
their lifestyle as well as a number of demographic characteristics.24 The self-reported heights and
weights were available only for the participants aged 15 years or more in the NHS. The NNS was
conducted using a sub-sample of NHS respondents selected from the base NHS sample of private
dwellings only. Unlike the NHS, the NNS was conducted by approaching a maximum of two inscope people per household in urban areas and three in-scope people in rural areas. These people
were randomly selected from the household. The NNS collected information on food and nutrient
in-take, food habits and attitudes and selected physical measurements for people aged two years or
more.5 Measured heights and weights were available for all participants in the NNS.
The data subsets of the NHS and the NNS were merged on the three unique identifying variables,
which are household indicative number (random), family number and person number. The total
exactly matched samples who had both self-reported and measured heights and weights was 572 for
Australian adolescents aged 15-19 years (boys 295, girls 277). There were 37 families contributing
two children to the sample of 572. Hence, there was some dependency introduced due to this
clustering. This was explored by excluding and including these clusters in the final analysis and
determined to be minimal, and was consequently ignored in the analysis presented here.
For the measurement of weight and height in the NNS, port-able digital scales (Tanita model
1597), a stadiometer and stretch stature method were used. Weight was measured with light clothes
and without shoes. Height was measured without shoes. The precision of weight and height
measurement was to the nearest 0.1 kg and 0.1 cm, respectively.5 In the NHS, participants aged 15
years more were interviewed with trained staff and asked to self-report their heights (without
shoes) and weights (without clothes and shoes).24
Recently developed criteria for overweight and obesity among children and adolescents provide
BMI values for males and females aged 2-18 years in six-month age brackets.25 We applied the
new international BMI cut-off points for overweight and obesity in the participants aged 15-18
years. For each age, the cut-off point at the mid-year value was applied (e.g. for those aged 15
years, we used the cut-off at 15.5 years). BMI cut-off points of 25 and 30 were used in the
definition of overweight and obesity, respectively, for the adolescents aged 19 years.4 Then, the
Australian adolescents aged 15-19 years were categorised into the groups of non-overweight/obese
(i.e. including underweight and normal weight subjects), overweight, obese as well as a combined
overweight/obese grouping.
TABLE 1. Numbers of adolescents aged 15-19 years in study
Age
Males
Number (%)
Females
Number (%)
15y
60
56.1
47
16y
72
55.8
57
43.9
44.2
17y
18y
54
57
47.8
51.8
59
53
52.2
48.2
19y
Total
52
46.0
61
54.0
295
51.6
277
48.4
The SPSS 10.0 for Windows package was used for all data management and statistical analysis.26
The differences between self-reported and measured weights were categorised into ± 2 kgs, ± 5 kgs
and ± 10 kgs while the differences for heights ± 2 ems, ± 4 cms and ± 8 cms. Differences in
continuous variables across categories of weight, height and BMI were assessed using the
independent t-test. Differences in proportions were assessed using Chi-Square. Bland-Altman
Australian and New Zealand Journal of Public Health 2002; 26 (5) : 473-8.
plots were used to display the degree of agreement between measured weight or height and selfreported weight or height.27
TABLE 2. The differences between self-reported weight/height and measured weight/height in
Australian adolescents
Gender
Males (%)
Females (%)
Total (%)
Weight difference (kg)
±2
±5
±10
41
76
91
45
81
95
43
79
93
Height
±2
41
48
48
difference (cm)
±4
±8
70
92
78
94
74
93
Figure 1. Difference against mean for weight (boys and girls)
Results
The total number of respondents aged 15-19 years was 572 (boys 295, girls 277) (see Table 1).
The overall proportion of overweight was 16.8% and obese 5.4% (derived from measured height and
weight) respectively. There was no significant difference in prevalence of overweight or obesity
between boys and girls.
For analysis, nearly half of adolescents (43%) reported their weight within a 2 kg difference of
actual weight, and 79% and 93% within 5 kg and 10 kg difference of actual weight, respectively;
48%, 74% and 93% of adolescents reported their height within 2 cm, 4 cm and 8 cm difference of
actual height, respectively (see Table 2).
On average, the reported weight was 2.0kg lower than measured weight (-2.0 ± 5.2 kg, p<0.01;
for boys -2.0 ± 6.1 kg, p<0.01; for girls -2.1 ± 4.0 kg, p<0.01). The difference between measured
weight and self-reported weight ranged from -15.1 to 43.2 kg. In considering the differences
between self-reported height and measured height, nearly half of the adolescents (48%) reported
within 2 cm, 74% within 4 cm and 93% within 8 cm (see Table 2). Mean self-reported height was
appropriately 1.1 cm higher than measured height (1.1 ± 3.9 cm, p<0.01; for boys 1.2 ± 4.0 cm,
p<0.01; for girls 0.9 ± 3.8 cm, p<0.01). The difference between measured height and self reported
height ranged from -19.9 to 20.7 cm. The mean weight or height difference between measured and
self-reported values was not significant by gender.
Australian and New Zealand Journal of Public Health 2002; 26 (5) : 473-8.
Figure 2. Difference against mean for height (boys and girls)
The Bland-Altman plots (see Figures 1 and 2) display the degree of agreement between measured
weight or height and self-reported weight or height. The ranges of discrepancies are similar for all
weights and heights. There were two extreme values that under-reported their weight more than 30
kg and two extreme values that over- or under-reported their height 20 cm. Although most of
difference values occurred in the range of means 2 SD (for weight difference, the range was from 8.4 to 12.5 kg in boys and girls; for height, it was from -8.8 to 6.7cm), these differences would not
be clinically acceptable.
Differences between measured weight and self-reported weight were significantly greater for
overweight or obese girls and boys compared with non-overweight/obese girls and boys (see Table
3). The differences between measured weights and self-reported weights were also significantly
greater for obese boys and all adolescents compared with overweight boys and all adolescents (see
Table 3). This means overweight or obese adolescents are more likely to report their height and
weight incorrectly compared with their normal/underweight counterparts.
TABLE 3. The differences between measured weights and reported weights by weight classification
Boys
Non-overweight/
obese group vs.
overweight
Mean difference
± SD (kgs)
0.9±3.8
Girls
p
<0.05
2.7±7.2
Non-overweight/
obese group
vs. obese
0.9±3.8
12.2±11.6
Overweight vs
obese
2.7±7.2
12.2±11.6
Mean difference
± SD (kgs)
1.5±3.3
All
p
<0.01
4.5±5.6
<0.01
1.5±3.3
4.5±5.6
4.2±5.6
P
<0.01
3.5±6.6
<0.05
4.2±5.6
<0.01
Mean difference
± SD (kgs)
1.2±3.6
1.2±3.6
<0.01
9.6±10.6
>0.05
3.5±6.6
9.6±10.6
<0.01
Self-reported weight was within 5 kg of measured weight in 85% of the non-overweight/obese
adolescents. For overweight adolescents, this fell to 62% (p<0.01). For obese adolescents, to 36%
(p<0.01). The actual weight, BMI and the proportion of overweight or obesity in adolescents who
under-reported their weights by more than 5 kg were significantly greater than those who reported
their weights within 5 kg of actual weight (see Table 4). There were no significant differences in
gender between those adolescents who under-reported or over-estimated their weight by more than 5
Australian and New Zealand Journal of Public Health 2002; 26 (5) : 473-8.
kg. Additionally, neither age nor measured height affected the numbers under or over-reporting their
weight by more than 5kg.
Based on measured data, 25% boys and 19% girls were classified as overweight or obese,
compared with 18% of boys and 12% girls who were so classified from self-reported data (see Table
5). The percentage of correct classification of overweight or obesity from self-reported data was
69% and 70% in boys and girls, respectively. There was no significant difference of over-weight or
obesity classification from self-reported data between boys and girls.
TABLE 4. Characteristics of adolescents who under-reported or over-estimated their weight by more
than 5 kg
Measured value (mean ± SD)
1. Under-reported weight by more than 5 kg
92
Weight
(kg)
77,4±17.0b
Height
(cm)
172.6±8.5
BMI
(kg/m2)
26.2±4.7b
Overweighta
(%)
33.71b
Obese a
(%)
21.7
2. Reported weight within 5kg of measured
weight
456 64.1±_11.7
170.4±9.0
22.0±3.3
12.9
2.4
3. Over-estimated weight by more than 5 kg
24
168.7±8.6
22.4±3.0
25.0
0
n
63.9±9.9
Notes:
(a) Overweight and obese categories are mutually exclusive.
(b) p<0.01 compared with group 2.
Discussion
Although the results suggested that both boys and girls under-reported their weight and over-estimated
their height, no significant gender difference in reporting occurred. Many previous studies on adults
and children have described that females are more likely to under-report their weight than males. For
example, Strass found that the self-reported weights were significantly lower than measured weight
among girls aged 12-16 years compared with boys.6 With adults, females tend to underestimate their
weight and in contrast, with males only those who are overweight or obese are likely to underestimate
their weight.28-30 Greater concern for weight in women may well be a realistic response to societal
roles, as reflected in advertising that still places the bulk of a woman's status on her appearance.31
The inaccurate information about weight or body image may also influence weight concern in
adolescents.32 Body image was one of the most significant factors associated with trying to change
weight among both male and females adolescents.33 However, much literature reports that young girls
are more likely to report a desire to be thinner compared with similarly aged boys.34-35 It would be
interesting in a further study to explore whether the non-gender difference in reporting found here
remains the same in younger adolescents. Unfortunately, data for this age group are not available in
the NHS and NNS data sets.
The use of self-reported heights and weights in adolescents results in the correct classification
of obesity in 74% of both boys and girls (86% of boys and 67% of girls) while Strass and Epstein
reported that the proportion of correct classification of obesity using self-reported height and weight in
children aged 12-16 years and 6-12 years was 94% and 91 %, respectively.21,35 This may suggest that
self-reported weights and heights of older adolescents remain less reliable for predicting obesity
compared with that of young adolescents. In this study, the self-reported weight was under-estimated
and self-reported height was overestimated. The percentage of misclassification of overweight or obesity
from self-reported data in this study was 31% for boys and 30% for girls. This finding is consistent
with recent results of Australian reports on adults by Flood18 and similar to an over 30%
misclassification reported for overweight in American adults.34 However, the misclassification is
higher than that of Australian adults in the NNS and NHS.19
The greater error in self-reported data in this study may reflect inaccuracy of reporting in older
adolescents. As the self-reported and measured weight and height were collected within a few weeks of
one another from the NHS and NNS, respectively, the weight and height are unlikely to have changed
significantly in such a short time. However, no information is available about how regularly
adolescents weight or measure themselves, so self-reported data by some adolescents may really be
just a guess. If population studies use self-report rather than measured results, it may be an important
health message to encourage the adolescents to weigh or measure themselves prior to completing the
questionnaire if reliable estimates are to be obtained.
Australian and New Zealand Journal of Public Health 2002; 26 (5) : 473-8.
There were some limitations with respect to the data in this study. First, as the weight was measured
with light clothing while self-reported weight was weight without clothing, there could be a slight
under-estimation on self-reported weight. However, as indicated above, there is no information about
how and how often the adolescents weighed themselves, therefore adjustment for clothing is not
justified. Another limitation occurred with this study regarding the unmatched numbers in the NNS.
There were 133 adolescents in the NNS who were unmatched. Comparison of BMI categories between
unmatched and matched records revealed that there were more obese adolescents and fewer
normal/underweight adolescents among the unmatched adolescents compared with those matched
counterparts. As obese adolescents are more likely to under-estimate their weights than their
normal/underweight counterparts, the results of this study may slightly underestimate on the bias
between self-reported and measured weight and height.
TABLE 5. Distribution of measured and self-reported SNII groups (%)a
Self-reported BM1 groups % (n)
Non-overweight
Measured BMI groups % (n)
Boys
Non-overweight/obese
Overweight
Obese
Total
Girls
Non-overweight/obese
Overweight
Obese
__
Total
73.2 (216)
7.1 (21)
1.4 (4)
81.7 (241)
1.7 (5)
10.5 (31)
3.7 (11)
15.9 (47)
0.3 (1)
2.0 (6)
2.4 (7)
74.9 (221)
18.0 (53)
7.1 (21)
100.0 (295)
Overweight
Obese
78.7 (218)
9.4 (26)
__
1.4 (4)
5.4 (15)
0.7 (2)
0.7 (2)
80.9 (224)
15.5 (43)
Total
88.1 (244)
0.7 (2)
7.6 (21)
2.9 (8)
4.3 (12)
3.6 (10)
100.0 (277)
Notes:
(a) Percentage of total boys and percentage of total girls.
In conclusion, this study provides baseline data on the comparison of measured and self-reported
weight and height in a national representative sample of older Australian adolescents, although there
were some limitations of the data. There was no significant gender difference in reporting weight and
height in older adolescents. Bias in reporting weight and height was much higher in overweight or
obese adolescents than non-overweight or obese adolescents.
The self-reported weight and height have to be more cautiously utilised for predicting overweight
or obesity in older adolescents compared with in young children and adolescents.
Acknowledgement
The authors thank Dr Diana Battistutta, statistician in the School of Public Health, Queensland
University of Technology, for her very helpful comments.
References
1. Himes JH, Dietz WH. Guideline for overweight in adolescent preventive services: Recommendations from an
expert committee. Am J Clin Nutr 1994;59307-16.
2. Lazarus R, Baur L, Webb K, Blyth F. et al. Recommended body mass index cut-off values for overweight
screening programs in Australian children and adolescents: Comparisons with North American values. J Paediatr
Child Health 1995;31:143-7.
3. Lazarus R, Baur L, Webb K, Blyth F. Adiposity and body mass indices in children: Bean's index and other weight
for height indices as measures of relative adiposity. 1nt J Obesity 1996;20:406-12.
4. National Health and Medical Research Council. Acting on Australia's Weight: A Strategic Plan for the
Prevention of Overweight and Obesity. Canberra: AGPS; 1997.
5. McLennan W, Rodger A. National Nutrition Survey Users' Guide 1995. Canberra: Australian Bureau of
Statistics; 1998.
6. Hills AP Byrne NM. Body composition, body satisfaction, eating and exercise behaviour of Australian
adolescents. In: Parizkova J, Hills AP, editors. Physical Fitness and Nutrition During Growth: Studies in Children
and Youth in Different Environment. Medicine and Sport Science. Volume 43. Basel: Karger; 1998. p. 44-53.
Australian and New Zealand Journal of Public Health 2002; 26 (5) : 473-8.
7. Pietrobell A, Fath MS, Allison DB, Gallagher D, Chiumello G, Heymsfield SB. Body mass index as a measure of
adiposity among children and adolescents: A validation study. J Pediatr 1998;132:204-10.
8. Schroeder DG, Martorell R. Fatness and body mass index from birth to young adulthood in a rural Guatemalan
population. Am J Clin Nutri 1999;70:137s-44s.
9. Bellizzi MC, Dietz WH. Workshop on childhood obesity; summary of the discussion. Am J Clin Nutri 1999; 70
:173s-5s.
10. Stunkard A.l, Albaum JM. The accuracy of self-reported weight. Am J Clin Nutri 1981;34:1593-9.
11. Schlichting P, Hoilund-Carlsen PF, Quaade F. Comparison of self-reported height and weight with controlled
height and weight in women and men. Int J Obesity 1981;5:67-76.
12. Palta M, Prineas RI, Berman R, Hannan P. Comparison of self-reported and measured height and weight. Am J
Epidemiol 1982;115:223-30.
13. Stewart AW, Jackson RT, Ford MA, Beaglehole R. Underestimation of relative weight by use of self-reported
height and weight. Am J Epidemiol 1987;125:122-6.
14. Kuskowska-Woik A, Karlsson P, Stoll M, Rossner S. The predictive validity of body mass index based on selfreported weight and height. Int J Obesity 1989;13:441-53.
15. Rowland ML. Self-reported weight and height. Am J Clin Nutri 1990;52:1125-33.
16. Nieto-Garcia FJ, Bush TL, Keyl PM. Body mass definitions of obesity: Sensitivity and specificity using selfreported weight and height. Epidemiology 1990;1:146-52.
17. Stevens J, Keil JE, Waid LR, Gazes PC. Accuracy of current, 4-year, and 28-year self-reported weight in an
elderly population. Am J Epidemiol 1990;132:1156-63.
18. Australian Bureau of Statistics. How Australians Measure Up. Canberra: ABS: 1999.
19. Flood V, Webb K, Lazarus R, Pang G. Use of self-report to monitor over-weight and obesity in populations:
Some issues for consideration. Aust NZ Public Health 2000;24:96-9.
20. Waters AM. Assessment of Self-reported Height and Weight and Their Use in the Determination of
Body Mass Index. Canberra: Australian Institute of Health and Welfare; 1995.
21. Strauss RS. Comparison of measured and self-reported weight and height in a cross-sectional sample of
young adolescents. 1nt J Obesity 1999;23:904-8
22. Brooks-Gunn J, Warren MP, Rossa J, Gargiulo J. Validity of self-reported measures of girl's pubertal status.
Child Develop 1987;58:829-41.
23. Giacchi M, Mattei R, Rossi S. Correction of the self-reported BMI in a teen-age population. 1nt J Obesity
1998;22:673-7.
24. McLennan W. National Health Survey Users' Guide 1995. Canberra: AGPS; 1996.
25. Cole T, Bellizzi MC, Flegal KM. Dietz WH. Establishing a standard definition for child overweight and
obesity worldwide: International survey. Br Med J 2000;320:1-6.
26. SPSS: statistical package for the social sciences [computer program]. Version 10.0. New York: SPSS; 2000.
27. Bland JM, Altman DG. Statistical methods for assessing agreement between methods of clinical measurement.
Lancet 1986:2:307-10.
28. Pixie P, Jacobs D, Jeffery R, Hannan P. Distortion in self-reported height and weight data. J Am Diet Assoc
1981;78:601-6.
29. Cash TF, Grant JR, Shovlin JM, Lewis RJ. Are inaccuracies in self-reported weight motivated distortions?
Percept Mot Skills 1992;74:209-10.
30. Paltra M, Pineas RI, Berman R, Hannan P. Comparison of self-reported and measured height and weight, Am J
Epidemiol 1982;115:223-30.
31. Tiggemann M, Pennington B. The development of gender difference in body size dissatisfaction. Aust Psychol
1990;25:306-13.
32. O'Dea JA, Abraham S, Heard R. Food habits, body image and weight control practices of young male and female
adolescents. Aust J Nutr Diet 1996;53:32
33. Middleman AB, Vazquez L, Durant RH. Eating patterns, physical activity, and attempts to change weight among
adolescents. J Adolesc Health 1998;22:37-42.
34. Gustafason-Larson AM, Terry RD. Weight-related behaviors and concerns of fourth-grade children. J Am Diet
Assoc 1992;92:818-22.
35. Maloney M, McGuire J, Daniels S, Specker B. Dieting behaviour and eating attitudes in children. Pediatrics
1989;84:482-7.
36. Epstein LH, McCurley J, Murdock RC. Estimation of percent overweight within families. Addict Behav
1991;16:369.75