Research Methods, test two (Ch 10-11, 13-16)
What are Qualitative questionnaires
They are more descriptive, used by epidemiologists, may use to assess dietary behavior (ex dietary supplement use, self-imposed or prof recommended dieting) or environmental factors (ex personal characteristics, cultural influences, involvement in private and public nutrition programs, interventions, or treatments) that may influence dietary exposures and their health-related outcomes.
When conducting research, why would someone want to use FERRET (Federal Electronic Research and Review Extraction Tool)
This government-developed computer search tool: 1. Enables users to access and manipulate large demographic and economic data sets online 2. Quickly locates current and historical info from sources 3. Get tabulations for specific info needed 4. Make comparisons between different data sets 5. Create simple tables 6. Download large amounts of data for custom reports
What is the goal of human nutrition survey research
To produce data on a sample of individuals drawn from a well-defined population of interest that relates to one or more of the following: nutrition-related beliefs, attitudes, and/or behaviors; their environmental influences, contexts,or determinants; and related health outcomes or consequences.
The highest intake level that would likely pose no risk of adverse health affects to almost all individuals in the population
Tolerable Upper Intake Level (UL)
The Interactive CDC Growth Chart is an example of...
Training manuals & how NHANES increased efforts to instruct users on how to access, process, and interpret data. -Survey documentation on methods and quality control procedures -Data user conferences -Listservs
True or False: Each nutrient with a UL has had an adverse effect identified
True -If no UL= data lacking, does not mean a person won't have a reaction
True or False: the mean intake of a healthy population could be well above the RDA
True (referring to the AI)
In regard to knowledge, attitude, and behavior assessments, what does the Behavioral Risk Factor Surveillance System focus on
-Focuses on personal behaviors (diet, physical activity, weight-loss practices) and nutrition/health status -Helps create and monitor programs: Youth Risk Behavior Survey
What does the National Nutrient Data Bank do
-It derives representative nutrient values for more than 6,000 foods and as many as 80 components consumed -Released as part of USDA Nutrient Data Base for Stndard Reference (updated annually) -FDA's Total Diet Study provides annual food composition analysis based on foods consumed most frequently in CSFII/NHANES
What is unique about the NHANES survey
-It used a complex multistage probability sampling and weighting -It required highly qualified and extensively trained staff -It collected data with mobile examination units and in-home interviews
What are some applications of survey methods
1. Assess various dimensions of individual nutritional status 2. Monitor controlled clinical trials 3. Design effective preventative nutrition interventions and MNT 4. Extend investigations of programs, communities, or systems (type, quality, impact, efficacy) 5. Monitor the nature and quality of food supply and food production systems
From the public access resources and data sets that are available and relevant, what are some pros and cons
1. Benefits: no time or money to collect data & recognized as credible data 2. Trade-offs: limited to variables that do not meet your needs, may not include your particular population of interest *If good data is out there, there is no need to re-create it. *It's important to know your audience
What were the overall goals of the Framingham study
1. Quantify trends in Framingham population's food and nutrient intake 2. Develop traditional and innovative nutritional risk assessment methods 3. Examine the relationships between dietary behaviors, disease risk, and health outcomes 4. Identify opportunities for preventative nutrition intervention strategies
What were some of the survey methods used in the Framingham study
1. RD's were employed in all interview-administered nutrition protocols 2. Their food frequency questionnaire & behavior questionnaire were developed and validated 3. They used 2-dimentional food portion visual for estimating portion sizes that were also developed and validated
What are some ways we can put our research to use
1. Show trends in diet and nutritional status or national incidences or estimates 2. Cite published findings or reference data 3. Allow data comparisons using national data collection and survey tools
What is needed for estimating statistical power
1. Sufficient numbers to answer research questions confidently- adequate study power. 2. Use published data relating to your research questions to make formal calculations of statistical power
What are some of the instruments used by RDs during the CSFII data collection/survey
1. Two 24hr diet recall surveys: -quick list -forgotten foods -time and occasion -detail cycle -final probe 2. Household questionnaire 3. Self-administered 2-day diet recall 4. Several weeks later, another survey by phone
What are some ways to format questions and surveys that would make the survey/interview be more successful
1. Use clear, easy to understand instructions, avoid items likely to be misinterpreted 2. Use questions with well-defined terms and conditions, underline the qualifying statement 3. Using open-ended questions allows respondents to provide information in their own forms.
What is the risk of inadequacy if a person's intake is equal to EAR
50%
Can be used when the Standard Deviations are unavailable for requirements of a nutrient
Coefficient of variation (CV) -CV= 10% was assumed for most nutrients in creating DRIs -Protein, niacin, and Vitamin A had different CV levels -A different assessment method was used in general for Iron (it varies too much) [CV=(SD/Mean x100]
Which sampling bias can be introduced by the mechanical procedures used to select units from the frame into the sample
Consistent sampling -If the sample of telephone numbers only contacted weekly between 9am-5pm, this may underrepresent employed persons and students who would unlikely be at home during those hours
What do you call the ratio of true variance of a statistic to the variance of a statistic for a simple random sample within the same number of cases
Design effect >No single design effect is universally applicable (in Nutri. it has to be specific) >Uses the assumptions of simple random sampling for the complexities of the sample design to adjust estimates and statistics >Make sure to compare the subsample to what to assure it is free of bias
What does the USDA compile data on
US food supply estimates -US Food and Nutrient Supply Series -Updated and published annually -Used to assess the potential of the US food supply to meet population needs and track changes over time
Give one example of a program that the USDA's Food and Nutrition Service Web-site gives vital statistics on
WIC School lunch program, Food Stamp Program
What are Semi-Quantitative questionnaires, give an example
Food Frequency instruments are an example. They can be used to characterized food intake behavior or estimate subjects' nutrient intake.
What is an example of sampling bias
Frame: sampling list used when the listing of the sampling units in the population is too difficult or tedious -Frame bias: can be caused by the use of an incomplete list -a sample based on a listing of telephone numbers for a particular city or geographic area would include only those households that have telephones or telephone numbers in a directory. -This may underrepresent low-income households and people who prefer unlisted numbers
Gender is answered as male or female, and employment status can be characterized as employed or unemployed. Which response category would be used in that example
Mutually exclusive: the responses can be classified in only one way
The cornerstone of the NNMRRP measurement component that also provides national data on the 1. nutritional status 2. dietary intake, and 3. numerous health indexes of the US pop is called what
NHANES *it also provides national pop -reference distributions -national prevalence of diseases and risk factors -trends in nutrition and health status over time.
What is the name of the survey that provides information about self-reported health conditions
NHIS National Health Information Survey -Large sample sizes enable data to be reported for the major racial/ethnic subgroups in the US pop in addition to age group, gender, and income level.
What is the NHIS survey and how often is it updated
NHIS is a continuous cross-sectional survey of the US pop, updated every 10-15 yrs It monitors trends in the health and health-status of Americans -Units are households (1 adult + 1 child), randomly selected by geographic probability sampling (provides info about Vit/mineral supp use, youth risk behavior, aging, food program participation, cancer, food prep)
This organization is composed of interconnected federal and state activities. It also provides info on: 1. Dietary and nutritional status of the US pop 2. Conditions that affect the dietary and nutritional status of people 3. Relationships between diet and health
NNMRRP
What is considered to be one of the best nutrition monitoring systems in the world?
NNMRRP
How is NNMRRP involved in public policy decisions
NNMRRP data collection is also used for public policy decisions related to the 1. Regulation of fortification, safety, and labeling of the food supply 2. Food production and pre marketing approvals 3. Food safety programs 4. WIC 5. US Dietary Guidelines for Americans
Which variable would this be an example of: Environmental characteristics that influence nutrition behavior/outcomes -Two or more categorical variables -Qualitative EXAMPLES: -"Do you eat breakfast daily? Yes or No" -"Check all that apply"
Nominal
What type of bias is it that fails to include elements in the sample that should be included; usually to reduce costs associated with collecting the data
Noncoverage bias (the failure to include elements in the sample that would properly belong in the sample) Ex. Hispanic Health and Nutrition Examinatin survey excluded areas with low Hispanic pops -May underrepresent more affluent Hispanics in the sample
Data collected on participants (women and children) from publicly funded food assistance, nutrition, and/or health programs involve a self-selected population is called what
Program Participants Data >Does not represent the community at large >BUT it's helpful in planning and evaluating these programs
Which DRI is used to plan people's diets, and from what is it derived
RDA is used to plan people's diets because the risk of inadequacy is so low -It is derived from EAR
-Achieve all characteristics of ordinal, but with meaningful distances between variables EXAMPLES: -Body weight in pounds or kg -Nutrient intakes in grams (or established unit) -Nutritional biomarkers/labs
Ratio Measurement/Interval
Intake that would meet the nutrient needs for ~98% of healthy people
Recommended Dietary Allowance (RDA)
What is another name for the population of interest
Referent population -Specific types of individuals, families, or households -Consumer or prof sub-groups -Communities or pops with unique characteristics -Other groups with traits that may affect human nutrition
When considering whether survey and surveillance data may be biased, errors are described for two areas...
Sampling bias, which includes sampling and non coverage errors AND Nonsampling bias, which includes nonresponse and measurement errors as well as data processing errors.
How do sampling errors and non-sampling errors differ
Sampling errors happen before you have your sample or during the process, and non-sampling errors happen once you already have the sample.
Which data collection method allows the subject time to think about their response, and search for needed info or check records, and is cost-effective
Self-administered
Which method of data collection methods would have less quality control, cannot assure completion, requires higher literacy levels/cognitive functioning, and carries the risk of non-response bias (results from failure to obtain observations on some elements selected and designed for the sample)
Self-administered -Mail or web-based/electronic -In person, individually or group -Can choose to hand-deliver and pick-up
What are the two data collection methods
Self-administered and Interviewer-administered
What is the CSFII survey
The Continuing Survey of Food Intakes by Individuals, of the USDA & NHANES, were established to provide national estimates of food and nutrient intakes in the general US population and subgroups (gen pop & low-income pop). -Environmental determinants of dietary behavior, nutrition knowledge and attitudes, food and nutrient intake -The survey was designed to determine what we eat in America, using a complex multistage probability sampling strategy: entire US> PSUs (Population Sampling Units) at county level> DUs (Dwelling Units) -Team approach nutrition professionals, survey methodologists, statisticians, food specialists, and project managers
What are the two response categories for survey questionnaires that capture the variable meaningfully and completely
Mutually exclusive and Exhaustive
What is oversampling
Certain groups are sampled with higher probabilities than others to provide adequate sample sizes for reliable estimates for these groups
What is the Framingham Nutrition Study
-Longest-standing prospective epidemiological investigations ever conducted originally designed to study the natural history and determinants of CVD since expanded to include: aging, familial genetics, other chronic disease endpoints, etc. -Two major cohorts 5,209 original men and women (18-60 years baseline) 5,135 FOS (Framingham Offspring Study); & spouses Framingham was chosen because it was rep of the US very well at the time of the study, it has a good response during the 90's because they built a good report with the community. It was a microcosm. Chosen as the study pop of interest after a national investigation ID communities with pop profiles that provided a suitable microcosm of the US pop in the mid-20th century. It was considered a pop representative of a typical American community, largely non minority individuals
How is NNMRRP used for research
-Nutritional requirements throughout the life cycle -Development of DRI's and their applications -Food composition, nutrient content, and bioavailability -Nutrition education research -Relationship of knowledge and attitudes to dietary and health behavior -Economic aspects of food consumption -Role of nutrition in the etiology, prevention and treatment of chronic diseases and conditions -Improve survey methods -Measure food insecurity -Increase the capability to capture state and local nutrition info
What should be considered when hiring staff for testing
-Staff needs to be suitably experienced and fully trained, training should focus on any problems identified in pre-testing -Quality control and assurance at all levels: Sampling, data collection, coding, data analysis, reporting
What are two surveys that deal with food and nutrient consumption
1. CSFII (continuing survey of food intakes by individuals) -Food and nutrient intake of the general population and low-income pop. 2. NHANES -Dietary intake is related to health status in the same individuals and racial-ethnic determinants of health are emphasized 3. CPS (Current Pop Survey) -By US Census Bureau -Measures the extent of food insecurity and hunger 4. School Nutrition Dietary Assessment Study -Assesses diets of American school children and contribution of the National School Lunch program to overall nutrient intake
Nutrition surveillance data are collected in more than 4,000 public health clinics across the country, and each data set has its own quality issues. What are important aspects that ensure high quality data
1. Calibrated equipment 2. Periodic training of staff 3. Quality control/assurance programs 4. Data collected through personal or proxy interviews are self-reported data and thus subject to error *Mom reporting birth wt of baby is what they have found to be accurate, not much else
What does the operationalizing process involve
1. Clearly defining each research question 2. Breaking it into its discrete component variables 3. Identifying a final set of research domains and questions 4. Linking each to a method of measurement 5. Data analysis for each measurement
What information does NCHS data systems (CDC) provide
1. Data on vital statistics 2. Information on health status, lifestyle, exposure to unhealthy influences, and the use of health care -National Center for Chronic Disease Prevention and Health Promotion (Maternal and child health) -CDC's Network of Prevention Research Centers -FDA's Center for Food Safety and Applied Nutrition Web-site -USDA's Food and Nutrition Service Web-site (Food Stamp Program, food security, WIC)
What are the 5 stages of survey research planning
1. Determine the research problem area 2. Create and pre-test survey plan 3. Finalize survey plan 4. Data collection 5. Manage data analysis
What is the NHANES survey, give an example of a survey linked to it
1. Determines population prevalence of various diseases and biological risk factors 2. Monitors trends in behavioral and environmental determinants of disease risk 3. Test hypotheses and model nutrition-disease risk relationships 4. Explore public health policy concerns -Provides national data on the nutritional status, dietary intake, and numerous health indexes of the US pop, national prevalences of diseases and risk factors, attends in nutrition and health status over time.
What is the purpose of NNMRRP
1. Help establish research priorities of significance to public health 2. Identify nutrition-monitoring needs (ex. Healthy People 2010 objectives for Improving Health) 3. Develop reference standards for nutritional status: CDC Growth Charts
What are limitations and other factors in survey and surveillance data analysis that should be considered, what questions should we ask ourselves?
1. Is the data suitable for the questions being asked 2. Consider survey design, strengths and weaknesses of methods used, and assess for bias. *Bias in survey or surveillance data can arise from errors in the sampling process and coverage of the target pop, from the lack of response to the survey by respondents and/or particular groups of respondents, and from measurement error.
What are the three aspects of the NHANES survey instrument
1. It used a home interview to assess for family/household info (food security, insurance, income, smoking) and for individual info (Medical condition, disease, wt, vision) 2. It gave a full examination in their Mobile Examination Unit (MEC) 3. They used both interview and self-administered surveys, using computer-assisted personal self-interview, and also assessed for behavioral risk factors
Why is survey research planning important
1. May answer one or more research question 2. Test hypotheses 3. Estimates pop characteristics 4. Model relationships between nutritional characteristic to other variables
What are two data bases that provide human nutrition research
1. NIH (Human Nutrition Research and Information Management system database) -Includes NNM research and applied methodologies supported in whole or in part by the fed govt 2. Health Information Index -A subject-word guide to diseases and conditions under investigation
Variables have qualities and characteristics that necessitate them being measured in certain ways. What are the ways variables can be measured
1. Nominal 2. Ordinal 3. Interval, or ratio scales
What are the 5 NNMRRP measurement component areas
1. Nutrition and health measurements 2. Food and nutrient consumption 3. Knowledge, attitudes, and behavior assessments 4. Food composition and nutrient databases 5. Food supply determinations *Shows the relationship between diet & health, diet & disease
When considering population sampling, random, probability, or clustered sampling strategies should be used and guided by what
1. Published research, found in the literature review (what studies had the best results?) 2. Advanced technologies are available to facilitate random sampling (Random-digit dialing or random email access) *Can have more than one pop if sufficient numbers to enable meaningful comparisons
What is the name of the scientific journal that gives citations for journal articles, monographs, theses, audiovisuals materials, and technical reports relating to all aspects of agriculture, and how is this different from MEDLINE/PubMed
AGRICOLA (AGRICultural Online Access) is the journal. MEDLINE is a biomedical database with citations from journal articles only. Both databases can be used to retrieve journal abstracts.
Intake assumed to be adequate to meet the nutrient needs for a group of healthy people, used when no RDA has been set/determined, the mean nutrient intake of a healthy population
Adequate Intake (AI)
-Ideally the mean intake of a population with a low prevalence of inadequacy -The standard for almost all nutrients for infants (based on the mean intake supplied by human milk for healthy, exclusively breastfed babies
Adequate Intake (AI) -expected to exceed the RDA for that population if one could be set
What were the alpha and beta errors for the Framingham study
Alpha (type I) error set at less than P < .05 The risk of finding an association between disease and diet risk when none exists Beta (type II) error was found to be ≤15% The risk of concluding that diet is unrelated to the development of CVD risk when it is actually related
What type of error considers the following question: What is the risk of rejecting the null hypothesis if it is true?
Alpha error, type I Ex: What is the risk of concluding that a relationship exists between diet and a biological risk factor when it does not actually exist?
A statistical multistage technique used to identify survey samples, but introduces complexities into the analysis of the data. It minimizes costs but with a trade-off in randomness
Area Probability Sampling -To produce estimates for a nation based on individual observations >Inflate the individual data by their probability of selection >Adjust for nonresponse >Post-stratify to bring the pop set into close agreement with US Census Bureau estimates >Special computer software for estimating variances for complex samples (SUDAAN, Wes VarPC, STATA) *This would not be a simple random sample, but it can be adjusted
What is a suggestion for making a self-administered questionnaire more engaging
Begin the survey with the most interesting items to capture their attention
What is a suggestion for helping the Interviewer-administered questionnaire person build rapport with the person being interviewed, make them feel more comfortable
Begin the survey with the simplest items/questions
What type of error considers the risk of accepting the null hypothesis when it is false?
Beta error, type II Ex: What is the risk of concluding that diet is unrelated to the development of a disease when it actually is related?
A survey's response rate can indicate...
Bias, but does not prove it
How can alpha error be minimized
By applying appropriate criteria in your statistical testing *Typical approach set the error limit at 5% or less (P < .05) Chance of rejecting the null hypothesis when it is true is 1 in 20 or less
How can beta error be reduced
By recruiting sufficient numbers of research subjects *Typical approach set the power level at ≤15% 15% or lower chance of accepting the null hypothesis when it is false Determine sufficient number of subjects to answer your question (s) with 85% or greater power
What are Quantitative questionnaires
Can be used to determine an individual's level of food and nutrient intake or related characteristics and behaviors, how numerical are we going to be.
When considering the ecological framework, what is one of the PRIMARY modifiable human behaviors of research interest
Diet: -Food and nutrient intake -Dietary quality -Relationship between diet and other human behaviors *There is a complex array of environmental & ecological influences on these behaviors:Knowledge, attitudes, & other biological and demographic factors; the psychological, social or cultural, organizational, a& physical environments in which the individual lives, functions, or is served or treated; and public and private policies, programs, and incentives that may influence diet and nutrition (such as Food Stamps or Medicare reimbursement of medical nutrition therapies)
-Set by several expert panels, including Food and Nutrition Board of the IOM -Six reports were published between 1997-2005 -Includes: EAR, RDA, AI, UL
Dietary Reference Intakes (DRIs)
Intake that would meet the requirement of 50% of healthy people
Estimated Average Requirement (EAR)
Responses that provide all suitable options, every possible answer is mentioned, are called what ("other" category can be used as convention to provide this dimension).
Exhaustive
Which data collection method uses the telephone or face-to-face, which allows for clarification of questions, interaction with participants might increase participation, uses computer technology that would increase quality control, ability to check for consistency or erroneous responses, can also help build rapport
Interviewer-Administered
Which data collection method is more costly, time consuming, has a greater potential for staff to introduce bias, and can contain field management challenges (poss hazardous setting)
Interviewer-Administered Telephone or face-to-face
Surveys addressing specific topics have ben periodically conducted. What does NCI's 5-A-Day for Better Health Baseline Survey assess
It assesses knowledge, behavior and attitudes about fruits and vegetables Also: -FDA's study of consumer food-handling practices -FDA's studies to evaluate the features and usability by consumers of Nutrition Facts Label
Name two scientific journals that contain information about federal and nonfederal nutrition research
MEDLINE/PubMed *they are a biomedical database with citations from journal articles only
What is it referred to when consistent errors arise in the interview or lab method used to obtain the data
Measurement bias or measurement error Examples: -Probing questions -Coding assumptions -Errors in food composition database used -Selection of the day of the week to interview *24-hr dietary recalls, measurement bias can be introduced through the interview methods
Nonsampling bias arises from systematic errors related to nonresponse, measurement, or data processing. How is nonresponse bias different
Nonresponse bias results from failure to obtain observations on some elements selected and designed for the sample. -Occurs because people are 1. not at home despite repeated attempts to contact them 2. they refuse to participate, they are incapacitated & unable to participate 3. Lost data (interviews/lab) 4. Incomplete reporting Incomplete reporting is a potential for bias in survey & surveillance data. Missing data elements should be excluded from analyses. Statisticians often use the survey response rate as an overall indicator of the quality of a survey. (non sampling bias)
When a lg proportion of the sample selected for a survey does not participate, a potential for bias exists if the nonrespondents differ from the respondents in some systematic way. The greater the nonresponse, the greater the potential for bias. What type of bias would this be
Nonsampling bias because it arises from systematic errors related to nonresponse, but more accurately would be nonresponse bias because of the failure to obtain observations on some elements selected and designed for the sample. Non-response bias is a sub-type of a non-sampling bias, so technically it would be both
An ongoing description of nutrition conditions in the population, with particular attention to sub-groups defined in socioeconomic terms, for purposes of planning, analyzing the effects of policies and programs on nutrition problems, and predicting future trends, is the definition for what?
Nutrition Monitoring
What is the online publication that provides information on all surveys, surveillance systems, and selected research activities
Nutrition Monitoring in the US: The Directory of Federal and State Nutrition Monitoring and Related Research Activities
An application of survey methods in human nutrition research that tests hypotheses relating to specific food or nutrient exposures to health outcomes of interest, and assess relationships between health and exposure methods and OVERALL diet quality
Nutritional Epidemiology
-Logical and rank-ordered -Categorical EXAMPLES: -"Do you consider your health to be: Excellent, Very Good, Good, Fair or Poor?" -"How often do you consume breakfast? Rarely or Never, About Once a Month, Several Times a Month, etc."
Ordinal
Which surveillance system focuses on nutrition-related problems and behavioral risk factors associated with low birth weight among high-risk prenatal populations
PNSS Pregnancy Nutrition Surveillance System
Which surveillance system has indicators on nutritional status among low-income, high-risk infants and children who participate in publicly funded programs
PedNSS Pediatric Nutrition Surveillance System