Body mass index (BMI) is calculated using body weight and height measurements. Parent-reported measurement of child’s height and weight is often used in public health research because this data collection method is efficient and cost-effective [
Inaccuracies in parent reports of child height may be due to difficulty measuring young children due to their small size, children’s difficulty in standing still during measurement procedures, parent use of incorrect measuring techniques, parent reliance on recall of height data from a doctor’s visit, and parent reporting bias [
Although these inaccuracies make it difficult to have confidence in parent-reported child height data when tracking individual children, parental reports of child height and weight may be acceptably accurate for estimating obesity prevalence in populations. A comparison of measured obesity prevalence in 1,497 school children with overweight prevalence estimated from parent-reported height and weight data found that both were similar, despite the tendency of parents to overestimate overweight by about 17% in boys and 10% in girls [
The obesity epidemic has spawned increased parent interest in child growth as well as obesity prevention intervention studies using BMI as an outcome measure [
This study was approved by the Institutional Review Boards at Rutgers University and the University of Arizona. All participants gave informed consent.
Parents of 2- to 5-year-old children were recruited via flyers posted at community sites and emails sent through workplace listservs at Rutgers University and the University of Arizona. Recruitment notices invited parents to review materials, participate in an interview, and implement instructions for measuring their own and their preschool child’s height. Each participating parent received $25 compensation, and children received a sheet of stickers to compensate them for their time. No parent-child dyad participated in more than one round of testing.
An iterative process was used to develop materials to improve the accuracy with which parents measured and reported child height. The height protocol for parents was created in two phases. Participants in both phases completed a brief survey to gather demographic data (e.g., age, highest education level achieved).
Brief, written instructions for measuring height were developed by nutrition researchers with extensive experience conducting anthropometric measurements. Drafts were iteratively reviewed, tested, and refined by a panel of experts in anthropometric measurements (
Height measurement protocol.
A vast array of commercially available tape measures were reviewed to determine which would have the greatest likelihood of yielding the most accurate measurements. Most were narrow (less than 2.5 cm or 1 in wide) and had measurements (tick marks) in increments of 1/16th in (0.16 cm). These were judged by the research staff as cost-effective but difficult to read accurately when hanging on a wall and had a precision level beyond that needed to calculate body mass index (BMI) percentile. Thus, a 7.6 cm wide (3 in) tape measure similar to those used by visually impaired individuals with increments marked at
The height kits were subjected to three rounds of testing in Phase A. In all rounds, researchers explained to parents that they planned to mail the height kits to parents in an upcoming Internet-based study and wanted to be certain the instructions for measuring height were clear. The researchers further explained that the purpose of parents’ participation was to help researchers improve the kits; thus parents should act as though they were in their own homes and the researchers were not there to answer questions. Each parent then was asked to read and follow the instructions and report the height measurements they obtained for their preschool child who accompanied them to the testing site. Then, with the parent watching, the researcher removed the measuring tape and plumb line, rehung the tape measure and plumb line in the same location used by the parent (if appropriate) or another location (if necessary to avoid obstacles that would result in erroneous measurements such as baseboards and carpet), and measured the child in duplicate. Finally, parents were asked for suggestions for improving the instructions and to use a 5-point scale, with 5 being the best score, to rate the instructions’ clarity, readability, and likelihood that they would improve the way they measure their children’s height.
Two trained researchers observed each parent as they operationalized the instructions to identify errors made in height measurement procedures. These errors were analyzed and used to refine the height kit prior to the next round of testing.
The need for Phase B became apparent with the findings from Phase A (see Results and Discussion). In Phase B, the research team created a brief (8-minute) narrated video (
The video and height kits were iteratively tested and refined in a series of four rounds of testing in Phase B. The video, played at research sites on a standalone laptop computer, was accompanied by a height kit developed and refined in Phase A. The procedures mirrored those used in Phase A. In addition, the survey completed at the start of the data collection protocol asked parents to report their child’s height and indicate how sure they were of the reported height using a 5-point scale. Parents also were interviewed by trained researchers at the end of the session to identify questions that arose as they watched each component of the video and suggestions for improving the understandability of the video. Finally, parents used a 5-point scale, with 5 being the best score, to rate the clarity of the instructions in the video, how well the video held their attention, and likelihood that the video would improve their measurement of child height.
For each participant, the difference between mean height as measured by the parent and the researcher “gold standard” was calculated and compared.
Table
Characteristics of participants and accuracy of child height measurements.
Phase |
Parent’s age in years | Education | Number of children <18 in household | Absolute mean difference in inches (cm) between parent and researcher measured child heights | |
---|---|---|---|---|---|
Mean |
Percent with secondary education or less | Percent with postsecondary education | Mean |
Mean |
|
Phase A | |||||
Round 1 ( |
|
0% | 100% |
|
0.23 ± 0.23 |
Round 2 ( |
|
0% | 100% |
|
0.60 ± 0.65 |
Round 3 ( |
b | 100% | 0% | b | 1.38 ± 2.12 |
Phase B | |||||
Round 1 ( |
|
44% | 66% |
|
1.66 ± 1.97 |
Round 2 ( |
|
54% | 46% |
|
0.71 ± 0.99 |
Round 3 ( |
|
40% | 60% |
|
1.55 ± 1.04 |
Round 4 ( |
|
50% | 50% |
|
0.35 ± 0.40 |
Parent rating of height kit.
Phase |
Clarity of instructions |
Reading ease |
Extent video held viewer’s attention |
Likelihood of improving measurement of child’s height |
---|---|---|---|---|
Phase A | ||||
Round 1 ( |
4.50 ± 0.55 | 4.67 ± 0.52 | — | 4.00 ± 1.55 |
Round 2 ( |
4.50 ± 0.90 | 4.50 ± 0.90 | — | 3.92 ± 1.08 |
Round 3 ( |
4.65 ± 0.49 | 5.00 ± 0.00 | — | 4.60 ± 0.89 |
All Phase A rounds combined | 4.53 ± 0.72 | 4.65 ± 0.71 | 4.09 ± 1.16 | |
Phase B | ||||
Round 1 ( |
4.78 ± 0.44 | — | 3.94 ± 1.07 | 5.00 ± 0.00 |
Round 2 ( |
4.92 ± 0.28 | — | 4.54 ± 0.28 | 5.00 ± 0.00 |
Round 3 ( |
4.80 ± 0.45 | — | 4.00 ± 1.00 | 5.00 ± 0.00 |
Round 4 ( |
4.90 ± 0.32 | — | 4.80 ± 0.42 | 4.95 ± 0.16 |
All Phase B rounds combined | 4.86 ± 0.35d | 4.39 ± 0.86 | 4.98 ± 0.08e |
In Round 2, 13 parent-child dyads recruited from a university-based preschool participated. As in Round 1, parents had a mean age of about 38 years and all had at least some postsecondary education. Similarly, parents rated the clarity of instructions and reading ease highly but were less positive that the height kit would improve the accuracy of the height measurements they made of their children. A comparison of researcher and parent child height measurements revealed close agreement. Minor refinements were made to the instructions based on parent comments and researcher observations.
In Round 3, 5 parent-child dyads were recruited from a community center. None of these parents had any postsecondary education. Although parents felt the instructions were clear, easy to read, and likely to improve their measurements of child height, researcher observations revealed that written instructions were difficult for parents with less education to use with fidelity. Mean researcher and parent child height measurements differed by 1.38 in (3.51 cm).
In Rounds 1 to 4 of Phase B, 9, 13, 5, and 10 parent-child dyads participated, respectively. Of these, about half had no postsecondary education. Common errors were not verifying tape measure straightness, hanging the tape measure over a baseboard instead of finding a flat wall, adjusting straightness of the tape measure by moving the bottom instead of the top which resulted in the tape not remaining flush with the floor, not removing child shoes, not having child breathe in/out before taking measurements, and misreading fractions. Several parents also suggested that the video should end with a summary of the steps and be accompanied with written instructions (i.e., “
Height measurement kit. Video:
Interestingly, on the survey completed at the outset of data collection, 18 (49%) parents reported that they did not know their child’s height. Average absolute differences between the values of the parents who reported child height (
Comparisons, using
This study’s findings indicate that a height kit composed of written instructions, an easy-to-read tape measure, plumb line, and explanatory video can help parents accurately measure child height. Compared to written instructions alone, parents rated the video as having significantly greater clarity and likelihood of improving their measurements of child height. Although no significant differences in accuracy were found between the written and video instructions, researcher observations indicated that the written instructions alone were more difficult for parents to use with fidelity, especially those with less formal education.
Self-report data are considered to be an accurate representation only when the data collection instrument has undergone formative testing such as reliability and validity analysis [
The small sample sizes in this study indicate that statistical comparison should be interpreted with caution. In addition, parents took measurements while under observation and may not be as careful in taking measurements when at home. However, the study has numerous strengths. An iterative process was used to continually modify and refine the measuring tape and written and video instructions and a mixed methods approach was taken for data collection. Quantitative measures of height measurements taken by parents were compared to the researcher “gold standard” and qualitative data collection from parents was used to identify specific areas of confusion and areas for improving and refining the written instructions and video. Future research should investigate the effectiveness of the height kit in improving parental height measurements of older children and determine its utility in more diverse population groups. Pairing the height measurements taken with the kit and child weight could further clarify usefulness of the height kit and video in reporting BMI.
Study findings suggest that an instructional video is perceived by parents as clear, easy to use, and likely to improve their measurements of child height. The height measuring kit developed and validated in this study has the potential to improve parent reports of child height, thereby improving the accuracy of BMI calculations, tracking of childhood obesity prevalence, and confidence in obesity prevention and treatment program outcomes based on parent-reported data.
The authors declare that they have no competing interests.
The following coauthors contributed to the work: Meredith Yorkin to study design, data collection, paper preparation, and paper review; Kim Spaccarotella to paper preparation, data analysis, and paper review; Jennifer Martin-Biggers to study design, data collection, data analysis, and paper review; Carolina Lozada to study design, data collection, and paper review; Nobuko Hongu to data collection and paper preparation and review; Virginia Quick to data analysis, paper preparation, and paper review; Carol Byrd-Bredbenner to study design and paper preparation and paper review. All authors read and approved the final paper.
This work was funded through USDA NIFA #2011-68001-30170.
Meredith Yorkin, Kim Spaccarotella, Jennifer Martin-Biggers, Virginia Quick, and Carol Byrd-Bredbenner received funding from the United States Department of Agriculture, National Institute of Food and Agriculture, Grant no. 2011-68001-30170.