Parental self-efficacy questionnaire: Development and psychometric testing Jonathan W. Decker, PhD, ARNP, FNP-BC Assistant Professor University of Central Florida College of Nursing Orlando, FL Supported by a grant from the Florida Nurses Foundation
Background Childhood overweight and obesity is an epidemic in the US BMI ≥ 85th percentile for age & gender Obese BMI ≥ 95th percentile for age & gender 2008 data >1/3 US 6-11 year olds overweight/obese 20% obese 5x increase in obesity from 1974-2008!
Background Parents are primary agents of change for children Parents should be targeted for intervention to help curb childhood obesity epidemic
Background Many parents claim to have knowledge of healthy diet and physical activity behaviors for their children... BUT Do not feel able or lack the confidence to enact that knowledge...
Background Lack self-efficacy Bandura’s self-efficacy theory Self-efficacy is confidence or belief in one’s ability to perform a behavior and to persevere and overcome any barriers that may arise in doing so.
Background Self-Efficacy Outcome Expectations Physical Social Self-Evaluated Environmental Factors Facilitators Barriers Goals Behavior
Purpose Interventions to increase parental self-efficacy...but need to assess change... To develop a reliable and valid instrument to measure parental self-efficacy for enacting healthy diet and physical activity behaviors in their children ages 6-11 years
Research Design
Sample US parents of children 6-11 years old Eligibility requirements Parent of a child 6-11 years old Able to read and write in English Available computer with internet access Convenience sample recruited
Sample Recruitment Internet posting to parenting discussion groups and websites E-mails to parental, professional and healthcare organization membership lists Fliers posted at several local pediatrician and pediatric dentist’s offices Internet posting to social networking site Word-of-mouth $5 electronic gift card incentive
Sample Internet usage for research 2008 data of internet users >200 million users in the US 70.2% of US population Environmental control Same as mailed surveys In-person cost prohibitive Multiple responses Restriction by IP address Specific demographic information
Sample Pilot 15 participants Main 146 participants Female 88% Non-Hispanic/Latino ethnicity 91% Caucasian race 82% Married 84% Employed full-time 64% Educated 97% Test-retest 25 participants
Data collection UCF IRB approval Waiver of documentation of consent Via internet No identifying information required Completed on SurveyMonkey Anonymous Encrypted environment
Measures 35-item questionnaire developed from: US Department of Agriculture guidelines MyPyramid Review of the literature Review of previous tools in similar domains Rate confidence on scale from 0 not at all confident 10 mostly or totally confident Pilot test α = 0.95
Measures Content validity Face validity 8 experts in relevant fields Assessed using content validity index (CVI) Rated items on a scale of 1 (totally irrelevant) to 4 (totally relevant) All items rated ≥3 CVI = 0.97 Face validity Same 8 experts Judged to be satisfactory
Demographic questionnaire Measures Demographic questionnaire Age Race Ethnicity Gender Marital status Education Work status Household income Zip code of primary residence Parental contact Number of children Ages Height Weight
Measures Self-efficacy for Eating Behaviors Survey (SEB-Eat) 5-point scale 1 I know I cannot 5 I know I can 61 items 5 factors α = 0.85 – 0.93 Test-retest r = 0.43 – 0.64
Measures Self-efficacy for Exercise Behaviors Survey (SEB-Ex) 5-point scale 1 I know I cannot 5 I know I can 12 items 2 factors α = 0.85 and 0.83 Test-retest r = 0.68
Data Analysis Reliability Internal consistency Test-retest reliability Cronbach’s alpha (α) Total score Subscale scores Test-retest reliability Pearson correlation coefficient (r) Individual items
Data Analysis Construct Validity Factor analysis Item analysis Forced 2-factor Item analysis Correlation of each item with Its subscale (with item removed) Other subscale Concurrent Pearson correlation coefficient with SEB-Eat and subscales SEB-Ex and subscales
Results
Data Analysis Internal consistency reliability (α) Total score 35 item α = 0.94 34 item α = 0.94 Dietary Behaviors (DB) subscale 27 items α = 0.93 Physical Activity Behaviors (PAB) subscale 8 items α = 0.92 7 items α = 0.94
Data Analysis Test-retest reliability Individual items Scores DB subscale r = 0.50 – 0.95 PAB subscale r = 0.53 – 0.92 Scores Total r = 0.94 DB subscale r = 0.89 PAB subscale r = 0.93
Data Analysis Factor analysis 2 interpretable factors DB PAB Item #33 25.3% of variance PAB 16.8% of variance Item #33 Did not load strongly onto either factor (conceptualized PAB item) Removed for further analysis
Data Analysis Item analysis Items correlated more strongly with their own subscale DB items DB subscale r = 0.31 – 0.70 PAB subscale r = 0.12 – 0.43 PAB items PAB subscale r = 0.67 – 0.90 DB subscale r = 0.36 – 0.44
Data Analysis Concurrent validity Total scores DB subscale scores SEB-Eat total r = 0.51 SEB-Eat subscales r = 0.32 – 0.48 SEB-Ex total r = 0.35 SEB-Ex subscales r = 0.32 & 0.34 DB subscale scores SEB-Eat total r = 0.55 SEB-Eat subscales r = 0.38 – 0.50 PAB subscale scores All <0.06, not significant
Discussion
Discussion Responses from 146 parents of children 6-11 years old resulted in 34-item questionnaire Total score + 2 subscale scores Dietary behaviors Physical activity behaviors Reliable Valid
Discussion Sample Not as diverse as anticipated Groups not well represented Racial minorities Ethnic minorities Lower income Less educated Males Non-married Sample homogeneity made analysis of difference between demographic groups impossible
Discussion Item #33 “How confident are you that you can limit your child’s screen time (i.e. T.V., video games, computer) to no more than 2 hours per day?” Did not load strongly onto DB or PAB Generated as a Physical Activity item Should convey association between screen time and physical activity Item removed
Discussion Concurrent validity Significant but moderate correlations Similar concept Novel concept Total score correlations SEB-Eat > SEB-Ex Physical activity not as consistent within a household as dietary behavior
Discussion – Future Research Expand demographic sample Racial and ethnic minorities Lower socioeconomic status Translate Spanish Other languages
Discussion – Future Research Test in other age groups Children 2-5 years old Children 12-17 years old Test outside the US Ensure meaning of questions consistent Canada Mexico Europe
Discussion – Future Research Qualitative investigation Do the questionnaire items mirror parents’ healthy lifestyle beliefs and practices? Do the questionnaire items get to the “heart” of the issue for parents?
Discussion – Future Research Uses for questionnaire Interventional studies Examine relationship of scores Children’s weight status Demographic subgroups Clinical screening tool Develop/refine models of childhood obesity
Discussion – Implications Education Healthcare practitioners Patients Policy Government spending Government programs Practice Treatment/prevention
Thank you Questions?