Development and reproducibility of a tool to assess school food-purchasing practices and lifestyle habits of Australian primary school-aged children.

Abstract

Objective: To describe the development and reproducibility of a self-report instrument, for use with children in years 4-6, to identify sources of food eaten during the day, and type and frequency of food purchases at school.

Design: Tool development stages included formulation of content and format, expert review, piloting and a test-retest study.

Subjects/setting: The pilot study included school students (n = 20) in years 4 and 5 (seven girls, mean age 9.7 [+ or -] 0.7 years) attending an Australian public primary school. The test-retest study was performed in a large metropolitan public primary school (n = 245 children, 52% female, mean age 10.7 [+ or -] 0.91 years) including children from years 4 (n = 88), 5 (n = 84) and 6 (n = 73).

Statistical analysis: A Kappa statistic was used to assess level of agreement between the two time periods separated by 1 week. The results were analysed using SAS version 8.2 with each question compared at time 1 and 2.

Results: The mean kappa was 0.529 using pairings from 17 questions. Values ranged from 0.18 to 0.71 (CI 0.46-0.60).

Conclusions and applications: The School Eating Habits and Lifestyle Survey has been developed and pilot-tested in primary school-aged children and shown to have moderate stability over time. The results show that each phase of development, particularly those spent in consultation and testing, led to progressive improvement of this instrument. This process improved the quality of information produced and gave insights to self-report of dietary intake and behaviours among children.

Key words: children, food habit, nutrition survey, purchasing behaviour, school canteen.

INTRODUCTION

Changes in the physical eating environment of schools together with the increasing prevalence of childhood obesity have prompted many nutrition experts and public health advocates to question whether the current school environment supports and promotes the development of healthy eating as normative childhood behaviour. (1-3) In schools, efforts can be directed to improving food intake through a multisectorial approach, incorporating the school food service and catering facilities. (4) Recognition of this opportunity has resulted in a worldwide trend towards the development of guidelines and standards for provision of healthy food choices in schools. (5-8)

Unlike the USA where a National School Lunch program provides lunch to approximately 59% of all public school children, most children in Australia (86%) bring their lunch from home. (9,10) Australian school food services are usually limited to an on-campus canteen that provides the opportunity to purchase food and drinks without leaving the school grounds. (9) While they most commonly operate during the morning tea and lunch breaks, each differs in their management systems, operating hours, price and range of products offered for sale. (9)

In Australia there is considerable government commitment to improving foods sold in schools, even though there are limited published data about children's school food-purchasing practices. (11) A recent study has indicated that 91% of primary school children use the canteen. (12) Although trend data are lacking it has been suggested that more children are purchasing lunch, recess and afternoon snacks from the school canteen because of busy family lifestyles and the competing demands on parents' time. (12) Data from a small number of studies conducted in Australia suggest that children most commonly use the canteen to purchase less healthy convenience foods like hot chips, pies and pastries. (12,13) Evidence also suggests that although Australian canteen food contributes less to overall energy intake than food brought from home, the types of foods commonly purchased from the canteen are those that are strongly associated with obesity. (9)

Development of effective strategies to improve children's food choices requires an understanding of their eating behaviours and factors that influence this in the school environment. (14) A more comprehensive understanding of school food-purchasing patterns could provide a means for informing development of intervention strategies, monitoring trends and evaluating the effectiveness of healthy school food programs. To do this specific tools are required to measure canteen purchasing practices. In addition, the validity and reliability of the assessment method must be reported in order to provide an assessment of reliability. (15) Currently, evaluation of school-based interventions for primary school children is hampered by the dearth of valid, reliable and acceptable methods for gathering self-reports of dietary intake and behaviours within a school setting. (16)

The present study describes the development, piloting and reliability testing of the School Food Eating Habits and Lifestyle Survey (SEHLS). The aim of the study was to report on the development and validation of a research tool capable of measuring influences on children's food choices and eating behaviours in the school environment, focusing on school food purchases.

METHODS

Phase 1 -- content and format

The initial content of the research tool was guided by a set of research questions generated from a review of relevant literature. A self-report instrument was required and this guided format and terminology in order to aid comprehension and interpretation by the target age groups. Questions were developed to assess practices and behaviours associated with key aspects of the Australian primary school food-purchasing environment. In the present survey the school food-purchasing environment is defined as foods eaten or purchased during the school day from the canteen, vending machines and external sources such as local shops and takeaway outlets.

Given there were no existing published tools with content relating to food-purchasing practices of school-aged children, the development of new questions was necessitated. Items relating to food-purchasing habits were structured to follow a logical sequence throughout the day. The response options for items related to canteen food purchases were included based on a review of 50 local primary school canteen menus. Common menu items were identified and grouped in the review and used as the basis for both the type of foods included and the terminology used to describe them. Similarly, items relating to vending machines and food purchases from sources outside school grounds were identified via a review of common options available. Items were included in the questionnaire relating to frequency of food purchasing from the canteen, vending machine use and food from sources external to the school. The time of day most likely to purchase food was based on the known available windows of opportunity for students.

Phase 2 -- expert review

This phase involved content validation both from a methodological and practical perspective by field experts. This included two experienced nutrition researchers (CC, RS), a statistician and data entry programmer all of whom have expertise in the field of survey development, measurement and evaluation. A primary school teacher also reviewed the tool to ensure that the survey would be easily understood, did not appear leading and was appropriately structured to facilitate interpretation.

The reviewers provided comments on the formulation of specific items, exclusiveness of items, presentation and wording, and the choice of response options. Items were also checked to ensure they were written clearly, complied with the format criteria, used familiar words, offered a plausible range of response options and were suitable for data entry.

Phase 3 -- pilot test

After expert review the revised survey was pilot-tested to assess the suitability of the tool for what it was intended to measure; to assess the feasibility of the proposed method of data collection and to provide experience in logistics of completion such as, time taken for administration and level of assistance required, as well as data management and analysis. The final aim of the pilot was to ensure that all questions were easily comprehensible to all respondents.

The pilot test was conducted at a local government primary school (n = 20, seven girls, mean age 9.7 [+ or -] 0.7 years). School selection and recruitment of the subjects were opportunistic. Subjects were recruited by their teacher who identified eligible students giving consent with permission from the school Principal and parents. The survey was administered in class time by the teacher. Subjects were in years 4 and 5, with at least half representing the lowest level of English language literacy estimated to be able to functionally self-complete the questionnaire. The subjects also represented the lowest age group intended to self-complete the survey. After completion, the teacher was interviewed about perceptions of student comprehension, time taken for administration and statements that the students found difficult to respond to or that needed clarification.

Phase 4 -- reliability testing

Reliability testing measured the stability of responses over time using a test-retest methodology. This included 245 children in years 4 (n = 88), 5 (n = 84) and 6 (n = 73) from a large metropolitan primary school (mean age 10.7 [+ or -] 0.91 years, 52% female). The survey was completed in class by children whose parents gave informed consent, with teacher supervision on two occasions, separated by one week. Approval to conduct the research was granted by the Hunter Area Research Ethics Committee, the University of Newcastle Human Research Ethics Committee and the NSW Department of Education and Training. These school years represented the full range of ages for future studies. Results were analysed using SAS statistical software, Version 8.2 of the SAS System for Windows, 2005, SAS Institute Inc., Cary, NC, USA. The frequencies of the categorical data from each question were compared between the first and second administration of the survey. A Kappa statistic was used to assess level of agreement between the two time periods. The resulting reliability coefficient is commonly expressed as a number between 0 and 1, with 0 indicating no reliability and 1 indicating perfect reliability. (17,18) For survey tools, values of 0.4 or above are considered to indicate moderate agreement and values of 0.8 or higher indicate excellent agreement. (19,20)

RESULTS

Phase 2 -- expert review

Feedback on the initial version of the SEHLS resulted in a number of modifications to achieve content validity. Table 1 presents a summary of modifications and rationale for the changes. An item relating to motivation for student food purchasing was added (question 12).

Phase 3 -- pilot testing

Further revisions were made after reviewing responses made in the pilot study and feedback from teachers. Teacher recommended modifications related to the interpretability and comprehension of items requiring changes to some wording and layout of items in the tool. The younger students and those with lower literacy ability also took longer to fill out the survey and required support, that is, someone was required to introduce and explain how to answer it and to be available to answer any questions. Time taken to complete the questionnaire was longer than anticipated and took about half an hour at the most. Modifications as a result of student responses led to the addition of new response options identified from responses written under the 'other, please specify' category. Inclusion of these was based on the likelihood of these options being used by the larger study target population. Table 1 summarises the resulting survey modifications.

Version 3.0 of the SEHLS comprises a total of 35 items, of which 25 focus on assessing food habits (questions 4-12), seven relate to activity and sedentary pursuits (questions 13-19) and three identify demographic variables including self-reported gender and date of birth (questions 1-3). Items relating to food habits assess usual canteen purchasing times and purchasing frequency, canteen spending, frequency of canteen use, sources of food eaten at school and at breakfast, and, types of foods purchased. One item is included for children to suggest influences on canteen purchases (question 12, 'If you buy food at school, what is the main reason why you would choose a particular food?'). All questions are multiple choice and categorical in nature. For some of the questions there is also the option for respondents to specify a response that is not given as an option. A full copy of version 3.0 of the SEHLS is available from the authors (MF).

Phase 4 -- reliability testing

Table 2 presents the results of the test-retest reliability assessment. The mean value was analysed excluding the questions allowing multiple responses due to the presence of very small sample sizes leading to unstable values. A small number of responses with perfect agreement (k = 1), due to small samples, were also removed. The resulting mean kappa value is thought to be a conservative estimate. From the remaining correlations using pairings from 17 questions, the kappa values ranged from 0.18 to 0.71. Table 2 presents the results for each of the 17 included questions.

DISCUSSION

The aim of the present study was to report on the development and validation of a research tool capable of measuring influences on children's food choices and eating behaviours in the school environment, focusing on school food purchases. The development of the SEHLS incorporated a methodology, adapted from similar studies addressing content validity and reliability, in the context of operational and practical efficiency. (15-17) The results clearly show that each phase of development, particularly those spent in consultation and testing, provided valuable feedback, progressively enhancing the instrument and increasing the viability of its future use.

The results from the developmental phase confirm the importance of these steps in survey tool development. Indeed the modifications made at this stage addressed several limitations inherent in the initial version. These included limited incorporation of data entry coding, unnecessary respondent burden, and ambiguity and lack of clarity in some items. Modifications improved the quality of data collected and generalisibility of results through addition and modification of questions and response options. The expert review phase, and specifically the inclusion of reviewers, with expertise spanning a variety of areas, improved the content validity both from a methodological and practical perspective. The resulting modifications give a clear indication that this was achieved.

The results of the pilot provided valuable information about potentially weak areas of the tool with respect to its administration and student interpretation. Teacher recommended modifications related to the interpretability and comprehension of items (see Table 1). Conversely, student responses led to the addition of new response options within the survey items (see Table 1). These findings highlight the importance of obtaining feedback from the tool administrator, in this case the teacher, in addition to subjects, in this case students, whose comments are likely to relate to different but important aspects of a tool.

The review of student responses and teacher interview, post pilot, gave specific insight into subjects' comprehension and understanding of what was being asked of them. Both age and cognitive ability influence the child's understanding of the method used and the thought processes that contribute to self-reporting of food choices. (15) Central to this was the inclusion of subjects in the pilot representing the lowest level of English language literacy estimated to be able to functionally self-complete the questionnaire. This increased the authors' confidence that the changes made to the SEHLS, version two, were appropriate for children at relevant stages of their curriculum.

Subar, in a commentary on the development of dietary tools, suggests that extensive piloting or cognitive testing is useful to be sure that the target population both understands what is being queried and is capable of providing the answers sought by investigators. (21) The pilot process, undertaken here, provided valuable information relating to the practical administration of the SEHLS. Information on completion time and level of instruction required is essential for the successful future application of the tool in a school classroom or formal research setting. Additionally, rehearsing and testing of surveys and tools can enhance the feasibility of future studies. (21,22)

A test-retest technique was employed to measure the temporal stability of the SEHLS to provide evidence of its consistency. (17,18,21-23) The SEHLS questionnaire was readministered after one week. The amount of time allowed between two measurements was important because of memory effects or potential dietary changes that can occur in the interim period between administrations. (24) The shorter the time gap, the higher the correlation, the longer the time gap, the lower the correlation. (15,17,18) The time interval (seven days) between the two administrations was long enough to avoid recall bias and is sufficiently short to avoid changes in the studied attributes. (18)

The SEHLS questionnaire produced a mean one-week test-retest correlation of 0.529 (CI 0.462-0.596) with a range from 0.18 to 0.71. Values of 0.4 or above are considered to indicate moderate agreement. (19) Within the range of results for the 17 questions assessed, three had values less than 0.4. The lowest value of 0.18 applied to the question asking about usual recess consumption, followed by the question about usual lunch consumption (0.35) and the question relating to vending machine use (0.36). The low kappa values indicate that these questions may need to be removed or modified in future versions.

Test-retest correlation coefficients from comparative published studies in child populations using self-completed surveys of dietary habits range between 0.06 and 0.91. (15,25-28)

Overall, the reproducibility of the SEHLS is similar to those of other questionnaires for school-aged children and represents an acceptable level of agreement in this population.

LIMITATIONS

The instrument is designed to identify school food-purchasing practices and lifestyle habits of school-aged children based on self-report. Ideally, the reliability study would be complemented with a validation study to give an assessment of the accuracy of the tool; however, that was beyond the scope of the present study. Finally, it is acknowledged that the survey is not comprehensive in assessing all factors associated with the school food environment. Other important measures of relevance may include: the types and quantities of food sold; price differentials between more healthy and less healthy foods; attitudes to nutrition; and assessment of indicators of the school policy environment. Furthermore, the proportion of the average child's food intake sourced from the school canteen in the context of overall dietary intake was not measured in the present study. This information is currently lacking in the literature and would provide vital insight into the wide range of factors influencing food choices of school-aged children in the context of their overall dietary intake.

CONCLUSIONS AND APPLICATION

The eating behaviour of children in the school environment has come increasingly under the spotlight in recent times. This is against the background of rising prevalence of overweight and obesity and a search for settings in which to target interventions. Development of effective strategies requires an understanding of children's eating behaviour and the factors that influence behaviour in specific environments including schools. Insight into food consumption patterns at school is valuable because it allows public health and education professionals to develop interventions that are sensitive to the school food setting. Evaluation of school-based interventions for primary school children is hampered by the availability of valid, reliable and acceptable methods for gathering self-reports of dietary intake in a school setting. Furthermore, the use of traditional dietary assessment methods such as 24-hour recalls, dietary histories and dietary records are often not suitable. The results from each stage of development of the SEHLS provide evidence for the importance of incorporating comparable methodologies in the development of similar tools in order to avoid many of the common pitfalls when such developmental work is not undertaken. This process not only provides gains in the quality of information produced but also contributes to a wider understanding of the nature of self-reports of dietary intake and behaviours among children. Finally, this work allows the authors to conclude that the SEHLS is ready to be used in a representative survey of school food canteen use in Australian primary schools.

ACKNOWLEDGEMENT

The present study was supported by funding from an Allied Health Research Grant from Hunter New England Health.

REFERENCES

1 Kubick M, Lytle LA, Hannan PJ, Perry CL, Story M. The association of the school food environment with dietary behaviours of young adolescents. Am J Public Health 2003; 93: 1168-73.

2 Dietz W. The obesity epidemic in young children. Reduce television viewing and promote playing. BMJ 2001; 322: 313-14.

3 Dowda M, Ainsworth BE, Addy CL, Saunders R, Riner W. Environmental influences, physical activity, and weight status in 8 to 16-year-olds. Arch Pediatr Adolesc Med 2001; 155: 711-17.

4 Davis SP, Davis M, Northington L, Moll G, Kolar K. Childhood obesity reduction by school based programs. ABNF J 2002; 13: 145-9.

5 Zive MM, Pelletier RL, Sallis JF, Elder JP. An environmental intervention to improve a la carte foods at middle schools. J Am Diet Assoc 2002; 102 (3 Suppl.): S76-8.

6 Baur L. Child and adolescent obesity in the 21st century: an Australian perspective. Asia Pac J Clin Nutr 2002; 11 (Suppl.): S524-8.

7 WHO. Obesity: Preventing and Managing the Global Epidemic. Geneva: World Health Organization, 1997. World Health Organization Technical Report Series No. 894.

8 Mckenna ML. Issues in implementing school nutrition policies. Can J Diet Pract Res 2003; 64: 208-13.

9 Bell AC, Swinburn BA. What are the key food groups to target for preventing obesity and improving nutrition in schools? Eur J Clin Nutr 2004; 58: 258-63.

10 Nicklas T, Johnson R, American Dietetic Association. ADA Reports. Position of the American Dietetic Association: dietary guidance for healthy children ages 2-11 years. J Am Diet Assoc 2004; 104: 660-67.

11 NSW Health. Prevention of Obesity in Children and Young People: NSW Government Action Plan 2003-2007. Sydney: NSW Department of Health, 2003.

12 Cleland V, Worsley A, Crawford D. What are grade 5 and 6 children buying from school canteens and what do parents and teachers think about it? Nutr Diet 2004; 61: 145-50.

13 Scarlett G. What are children eating at school? In: Food Matters at School: Proceedings of the Towards a Healthier Canteen Conference. Sydney: Nutrition Education Service, 1993; 5-17.

14 Shannon SM, Fulkerson JA, French SA. Factors in the school cafeteria influencing food choices by high school students. J Sch Health 2002; 72: 229-34.

15 Mcpherson SR, Hoelscher DM, Alexander M, Scanlon KS, Serdula MK. Dietary assessment methods among school-aged children: validity and reliability. Prev Med 2000; 31: S11-33.

16 Edmunds LD, Ziebland S. Development and validation of the Day in the Life Questionnaire (DILQ) as a measure of fruit and vegetable questionnaire for 7-9 year olds. Health Educ Res 2002; 17: 211-20.

17 Huck SW, Cormier WH. Reliability. In: Reading Statistics and Research. New York: Harper Collins, 1996; 75-88.

18 Turconi G, Celsa M, Rezzani C, Blino G, Sartirana MA, Roggi C. Reliability of a dietary questionnaire on food habits, eating behaviour and nutritional knowledge of adolescents. Eur J Clin Nutr 2003; 57: 753-63.

19 Stokes ME, Davis CS, Koch GG. Categorical Data Analysis Using the SAS System, 2nd edn. Cary, NC: SAS Institute, 2000.

20 Boynton PM. Selecting, designing and developing your questionnaire. BMJ 2004; 328: 1312-15.

21 Subar. A. Developing dietary assessment tools. J Am Diet Assoc 2004; 104: 769-70.

22 Duxbury J. Testing a new tool: the management of aggression and violence attitude scale. Nurse Res 2003; 10: 39-52.

23 Howell DC. Statistical Methods for Psychology, 4th edn. Belmont, CA: Wadsworth Publishing Company, 1997; 160-61.

24 Hoelscher DM, Day RS, Kelder SH, Ward JL. Reproducibility and validity of the secondary level school-based nutrition monitoring student questionnaire. J Am Diet Assoc 2003; 103: 186-95.

25 Rockett HR, Wolf AM, Colditz GA. Development and reproducibility of a food frequency questionnaire to assess diets of older children and adolescents. J Am Diet Assoc 1995; 95: 336-40.

26 Anderson AS, Bell A, Adamson A, Moynihan P. A questionnaire assessment of nutrition knowledge--validity and reliability issues. Public Health Nutr 2001; 5: 497-503.

27 Stevens J, Cornell CE, Story M et al. Development of a questionnaire to assess knowledge, attitudes, and behaviors in American Indian children. Am J Clin Nutr 1999; 69 (Suppl.): 773S-81S.

28 Smith K, Hoelscher D, Lytle L et al. Reliability and validity of the Child and Adolescent Trail for Cardiovascular Health (CATCH) Food Checklist: a self report instrument to measure fat and sodium intake by middle school students. J Am Diet Assoc 2001; 101: 635-47.

Meghan FINCH, (1,2) Andrea BEGLEY, (3) Rachel SUTHERLAND, (1,2) Michelle HARRISON (4) and Clare COLLINS (2)

(1) Hunter Population Health, Wallsend, Schools of (2) Health Sciences and (4) Mathematics and Physical Sciences, University of Newcastle, Newcastle, New South Wales, and (3) Nutrition and Dietetics Program, School of Public Health, Curtin University of Technology, Perth, Western Australia, Australia

M. Finch, MPH, Public Health Nutritionist, Conjoint Lecturer in Nutrition and Dietetics

A. Begley, MPH, APD, Lecturer

R. Sutherland, MPH, APD, Nutrition Program Manager, Conjoint Lecturer in Nutrition and Dietetics

M. Harrison, GradDipMedStat, Lecturer/Tutor

C. Collins, PhD, APD, Senior Lecturer in Nutrition and Dietetics

Correspondence: C. Collins, Mail Box 38 Hunter Building, Callaghan, Newcastle, NSW 2308, Australia. Email: [e-mail]

Table 1 Summary of the types of modifications to the School Eating
Habits and Lifestyle Survey

                                         Item(s) or      Item(s) or
Reason                                   response(s)     response(s)
underlying                               prior to        post
change            Example of change      modification    modification

Expedite data     Pre-coding added       NA              NA
  entry purposes  alongside response
                  options. For example
                  respondents now asked
                  to indicate chosen
                  response by circling
                  number instead of
                  shading in circle
Address           Layout of questions    NA              NA
  formatting      on the page changed
                  from having up to
                  three running from
                  left to right across
                  a page to single
                  questions running
                  down the page
Ease respondent   Addition of filter     Where do you    Do you usually
  burden          questions for          usually eat     eat breakfast
                  consumption of         breakfast?      on school
                  breakfast, recess and  How often do    mornings?
                  lunch. Prior to items  you usually     Where do you
                  asking about           eat breakfast?  usually eat
                  particulars of                         breakfast?
                  breakfast consumption                  How often do
                                                         you usually eat
                                                         breakfast?
Minimise          Addition of            'Yes'           'Yes'
  confusion       additional filter      'No'            'No'
                  options in the item                    'No, my school
                  asking about school                    doesn't have a
                  vending machine                        vending
                  purchasing                             machine'#
                                                         'I don't know
                                                         what a vending
                                                         machine is'#
Enhance data      Addition of please-    'Other'         'Other, please
  quality         specify handwritten                    specify ______'
                  option for 'other'
                  responses
Enhance           Rewording and          'Yes'           'No, my school
  interpretation  addition of extra      'No'            does not allow
  of results      filter options for                     it'
                  item relating to                       'My school
                  purchase of food                       allows it and I
                  outside school                         don't buy any'
                  grounds                                'Yes, my school
                                                         allows this and
                                                         I do buy some'
                                                         'My school
                                                         doesn't allow
                                                         it and I do buy
                                                         some'
Enhance study     Items relating to      NA              How much time
  quality         physical activity                      do you usually
                  and sedentary                          spend each day
                  behaviours                             on the computer
                                                         or playing
                                                         video games?
                                                         On week days
                                                         how much time
                                                         do you spend
                                                         watching
                                                         television?
                                                         How many times
                                                         a week do you
                                                         usually eat
                                                         your meal at
                                                         night in front
                                                         of the
                                                         television?
                                                         How much time
                                                         do you usually
                                                         spend each week
                                                         playing games
                                                         or doing
                                                         activities that
                                                         make you run
                                                         around or huff
                                                         and puff?
                                                         What do you
                                                         usually do at
                                                         recess?
                                                         What do you
                                                         usually do at
                                                         lunch time?

Boldface indicates changes.

Note: The changes indicated with #.

Table 2 Results of test-retest study of the School Eating Habits and
Lifestyle Survey by question

                                                         Lower   Upper
Question                                          Kappa  95% CI  95% CI

Q4 -- Do you ever buy food from the school        0.53    0.279  0.78
  canteen?
Q4a -- What time of day do you usually buy food   0.54    0.45   0.63
  from the school canteen?
Q4b -- How often do you buy food from the school  0.57    0.48   0.66
  canteen?
Q5 -- Do you usually eat breakfast on school      0.58    0.37   0.79
  mornings?
Q6 -- Do you usually eat recess when you are at   0.18   -0.07   0.43
  school?
Q7 -- Do you ever buy recess from the school      0.68    0.58   0.78
  canteen?
Q8 -- Do you usually eat lunch when you are at    0.35    0.08   0.63
  school?
Q9 -- How often do you buy lunch from the school  0.52    0.34   0.70
  canteen?
Q10 -- Do you ever buy food or drinks from a      0.36    0.19   0.53
  vending machine that is in your school?
Q11 -- Do you ever buy food or drinks from        0.48    0.26   0.69
  places outside school grounds during school
  time?
Q13 -- How much time do you usually spend each    0.59    0.48   0.68
  day on the computer playing video games?
Q14 -- How much pocket money are you usually      0.57    0.46   0.67
  given each week to buy food, including snacks
  and drinks?
Q15 -- On week days how much time do you usually  0.51    0.38   0.64
  spend watching television (TV)?
Q16 -- How many times a week do you usually eat   0.63    0.56   0.71
  your meal at night in front of the television
  (TV)?
Q17 -- How much time do you usually spend each    0.63    0.53   0.73
  week playing games or doing activities that
  make you run around or huff and puff?
Q18 -- What do you usually do at recess?          0.71    0.62   0.79
Q19 -- What do you usually do at lunch?           0.57    0.46   0.68
Overall mean                                      0.53    0.46   0.60
SD                                                0.13