According to research reviewed by Gram, the That’s Me: My Choices survey appears to be a reliable tool for measuring children’s food-related behaviors. Across 5,528 students in grades 3 to 5 over three years, it showed adequate consistency and detected significant program differences each year.
A new study looked at whether a short survey called That’s Me: My Choices can reliably measure how often elementary school children do food-related healthy behaviors. Researchers analyzed answers from 5,528 students in grades 3 to 5 over three years of a school nutrition program. The survey proved consistent and measured one main idea, and it detected meaningful program changes each year. According to research reviewed by Gram, the tool appears to be a dependable way for schools and nutrition programs to check whether their efforts are working. The survey asks children to recall their behaviors from before a program and after it, so some memory limits remain.
Key Statistics
A 2026 study of 5,528 children in grades 3 to 5 found that the That’s Me: My Choices survey showed adequate reliability, with a composite reliability score of 0.803.
A 2026 analysis of the That’s Me: My Choices survey detected statistically significant program differences in children’s food-related behaviors for three consecutive years in the Healthy Schools, Healthy Communities program.
In a 2026 psychometric evaluation of 5,528 elementary students, the survey’s confirmatory fit scores (comparative fit index of 0.977 and root mean square error of approximation of 0.045) indicated a good match to its single-idea structure.
The Quick Take
- What they studied: Can a short, look-back survey accurately measure how often elementary school children do healthy food behaviors, and does it work well enough to show whether a nutrition program is making a difference?
- Who participated: About 5,528 children in grades 3 to 5 at elementary schools. All of them took part in nutrition education programs funded through the Supplemental Nutrition Assistance Program Education, known as SNAP-Ed.
- Key finding: The survey was consistent, with a reliability score of 0.803 where 1.0 is perfect, and it measured one main idea about eating behaviors. It also showed significant program changes for three years in a row. Children in younger grades found it a bit harder to remember what they did before the program.
- What it means for you: If you work with children or run school nutrition programs, this survey may be a practical tool for checking progress. It is not a diagnosis tool for any individual child, and results come from one program, so other settings may differ.
The Research Details
Researchers examined data that had already been collected from children who took part in a school nutrition program called Healthy Schools, Healthy Communities. The survey was given over three school years. Each child answered questions about food-related behaviors and was asked to think back on how often they did those behaviors before the program started, and then how often they did them afterward. This is called a retrospective pre-post design because the ‘before’ answers are gathered after the program ends.
To test whether the survey worked well, the team ran several statistical checks. They looked at whether the questions hung together as a group, whether they seemed to measure a single idea, and whether the survey meant the same thing across different grade levels, different years, and children who found recall easier or harder. They also checked whether questions were appropriate for each grade and whether children seemed tired or rushed by the end.
These steps are the kind used to confirm that a questionnaire is trustworthy before using it to judge results. The study did not test a new food intervention itself. Instead, it checked the measuring tool that programs use to judge their own results.
Surveys about children’s habits are only useful if they measure what they claim to measure. Without careful checks, a program might look successful when the change came from the survey itself, such as children answering differently because of how questions were worded. Testing the tool across several years, several grade levels, and different recall abilities helps show that any change seen is more likely to be real.
The retrospective design is especially practical for schools because it avoids giving a full ‘before’ survey on day one, which can be hard with young children. But it also raises a concern: memories of the past can be off. Testing this design carefully helps researchers understand how much that concern matters.
The study uses a large sample of 5,528 children, which gives the statistical checks solid footing. It used both exploratory and confirmatory factor analysis, which are standard methods for checking whether a survey’s questions group together as expected. The fit scores were good: a comparative fit index of 0.977 and a root mean square error of approximation of 0.045 both fall within commonly accepted ranges for a well-fitting model.
The data came from an existing program evaluation rather than a controlled experiment designed from scratch. That means the study is strong for checking the survey but cannot by itself prove the nutrition program caused the changes. The paper was published in the Journal of Nutrition Education and Behavior, a peer-reviewed journal in its field.
What the Results Show
The main test looked at whether the survey’s questions worked together as one measure. A reliability score of 0.803 was reported, which is considered adequate for a survey of this type. This means the questions were answered in a fairly consistent way and tended to rise and fall together, as they should if they measure the same underlying habits.
Next, the researchers checked the structure of the survey. Exploratory analysis suggested that the questions measure one main idea, which the authors describe as a single dimension of food-related behaviors. A follow-up confirmation step supported this, with fit scores that indicate the one-idea structure matches the data well.
The survey also held steady across different conditions. Its measurement stayed consistent across the three program years, across grade levels, and across children who had more or less trouble recalling their earlier behaviors. In plain terms, the questions appeared to mean the same thing to children regardless of when they took the survey or what grade they were in. This matters because it means changes in scores are more likely to reflect real differences rather than changes in how the survey was understood.
Finally, the survey detected statistically significant program differences in each of three consecutive years. This shows that the tool was sensitive enough to pick up improvements in children’s reported behaviors after the program.
Younger children, in lower grades, reported slightly more difficulty remembering what they did before the program. Those children also tended to give lower ‘before’ answers. This pattern suggests that younger students may have underreported earlier habits, or that the difficulty of recall shaped how they rated themselves. It is a modest effect, but it is important because it could make a program look more effective for younger children than it truly is.
The authors also examined whether questions were appropriate for each grade and whether children showed signs of response fatigue, meaning they became tired and stopped answering carefully. These checks are important for judging whether results from older and younger students can be compared directly.
Many nutrition programs for children rely on surveys that were never formally tested for accuracy. Researchers have long noted that retrospective pre-post designs are useful in schools but can be vulnerable to memory errors. This study adds evidence that such a design can work when the survey is carefully tested. It also fits with broader guidance that psychometric testing, the technical process of checking a survey’s accuracy and consistency, should be part of program evaluation.
The study relied on data that had already been collected, so the survey was not designed from scratch for this analysis. Children were asked to remember their past behaviors, and memory can be imperfect, especially for younger students. The study included only elementary school children in one program, so the results may not apply to older students, adults, or other types of nutrition programs.
The survey measures reported behaviors, not what children actually ate or did. Children could have answered in ways they thought were expected. The study also cannot fully separate the effect of the program from other things happening in children’s lives during the school year. Finally, the reliability score is adequate but not very high, which means individual scores should be read with some caution.
The Bottom Line
Based on this analysis, schools and nutrition educators may reasonably use the That’s Me: My Choices survey to track group-level changes in children’s food-related behaviors. Confidence is moderate to strong that the tool is consistent and measures one main idea, because several independent statistical checks agreed. Confidence is more limited about the exact size of any individual change, because the survey relies on memory and reliability was adequate rather than excellent. Programs should pair survey results with other information when making decisions.
School nutrition staff, SNAP-Ed program coordinators, evaluators, and researchers who work with elementary students will find this study most relevant. Parents and teachers may find it useful to understand how program results are measured. People working with older students, adults, or entirely different health topics should not assume these findings automatically apply to them.
This study focuses on measurement quality rather than a health outcome, so there is no personal timeline for seeing benefits. For programs, the survey was able to detect significant differences within a single school year and repeated that pattern across three consecutive years. Schools that adopt the tool should plan to collect data across a full program cycle to see reliable trends.
Frequently Asked Questions
How accurate is the That’s Me: My Choices survey for measuring kids’ eating habits?
The survey showed adequate reliability, with a composite score of 0.803 in a study of 5,528 students in grades 3 to 5. It measured one main idea consistently across three program years, though children’s memory of past behaviors can introduce some error.
What is the That’s Me: My Choices survey used for?
It is used by school nutrition programs to measure changes in children’s food-related behaviors. Children answer questions about how often they did healthy habits before and after a program, which helps program staff judge whether their efforts are making a difference.
Do younger children have trouble remembering their past eating habits on surveys?
Children in lower elementary grades reported slightly more difficulty recalling behaviors from before a program. Their earlier answers tended to be lower, which suggests recall difficulty may affect results and should be considered when comparing grade levels.
Can schools trust surveys that ask kids to remember what they did before a program?
Schools can use these surveys with reasonable confidence when the tool has been tested for accuracy. This study found the survey stayed consistent across years and grades, but researchers note that memory limits remain and results should be read alongside other program data.
Want to Apply This Research?
- Track the percentage of children who report doing each key healthy behavior at least a few times per week, both before and after a program, using the same survey items each time. Compare the change in these percentages across the school year and across program cohorts.
- Encourage children and families to set one small, specific food goal each week, such as eating a vegetable at dinner three times or choosing water with lunch. Use the app to log completed goals with a simple check-in, which mirrors the kind of behavior the survey measures.
- Run the same short survey at the start and end of each program term, and keep records by grade level so younger and older students can be compared fairly. Review results yearly, watch for changes in recall difficulty or response fatigue, and combine survey trends with attendance and participation data to build a fuller picture over time.
This article summarizes a research study about a survey tool used to measure children’s food-related behaviors. It is not medical advice and does not provide guidance on diagnosing or treating any health condition. The study evaluated a measurement tool within one nutrition education program, so results may not apply to every school or child. Parents with questions about their child’s diet or health should consult a pediatrician or registered dietitian.
This research translation is published by Gram Research, the science division of Gram, an AI-powered nutrition tracking app.