Parental food selection questionnaire - Infant version

Abstract

Purpose:

Complementary feeding practices (CFPs) are associated with health outcomes (e.g., obesity and food allergies). Understanding how parents select foods for their infant is limited. This study’s purpose was to develop a psychometrically sound measure of parents’ food selection motives for their infant during the complementary feeding period.

Methods:

Development and testing of the Parental Food Slection Questionnaire-Infant Version (PFSQ-I) occurred in three phases. English-speaking, U.S. mothers of healthy infants, aged 6–19 months old participated in a semi-structured, face-to-face interview (Phase 1) or a web-based survey (Phases 2 & 3). Phase 1 was a qualitative study of maternal beliefs and motives surrounding complementary feeding. Phase 2 involved adaptation and exploratory factor analysis of the original Food Choice Questionnaire (Steptoe et al., 1995). Phase 3 involved validity testing of the relationships among PFSQ-I factors and CFPs (timing/type of complementary food introduction, frequency of feeding method, usual texture intake, and allergenic food introduction) using bivariate analyses, and multiple linear and logistic regression analyses.

Results:

Mean maternal age was 30.4 years and infant age was 14.1 months (n = 381). The final structure of the PFSQ-I included 30 items and 7 factors: Behavioral Influence, Health Promotion, Ingredients, Affordability, Sensory Appeal, Convenience, and Perceived Threats (Cronbach’s α = 0.68–0.83). Associations of factors with CFPs supported construct validity.

Discussion:

The PFSQ-I demonstrated strong initial psychometric properties in a sample of mothers from the U.S. Mothers who rated Behavioral Influence as more important were more likely to report suboptimal CFPs (e.g., earlier than recommended complementary food introduction, delayed allergenic food introduction, and prolonged use of spoon-feeding). Additional psychometric testing in a larger, more heterogenous sample is needed, along with examination of relationships between PFSQ-I factors and health outcomes.

1. Introduction

Suboptimal complementary feeding practices (CFPs), which refer to when, what, and how solid foods are introduced to infants, are associated with adverse health outcomes across the lifespan (Du Toit et al., 2015; Gingras et al., 2019; Koplin et al., 2010; Mennella et al., 2021; Peters et al., 2019). In high income countries, these adverse health outcomes include obesity and food allergies (Du Toit et al., 2015; Gingras et al., 2019; Koplin et al., 2010; Mennella et al., 2021; Peters et al., 2019). For example, early introduction of complementary foods (prior to 4 months old) and feeding of juice, nutrient-poor snacks and desserts/sweets during infancy have been associated with increased risk of obesity (Auerbach et al., 2017; Gingras et al., 2019; Mennella et al., 2021; Moore et al., 2019; Pearce et al., 2005). Additionally, delayed introduction of allergenic foods (i.e., after 12 months old, compared to prior to 6 months old) has been associated with increased risk of developing a food allergy (Du Toit et al., 2008, 2015; Field, 2017; Koplin et al., 2010; Nwaru et al., 2010; Peters et al., 2019). These adverse feeding-related health outcomes are increasing in prevalence (Fryar et al., 2020; Gupta et al., 2019; Tang & Mullins, 2017), can persist across the lifespan (Gupta et al., 2018; Pandita et al., 2016; Park et al., 2015), and can have significant impacts on quality of life and productivity (Abrams et al., 2020; Hamilton et al., 2018; Puhl & Heuer, 2009; Umer et al., 2017; Warren et al., 2016). However, interventions have been relatively unsuccessful in reducing trends in obesity and food allergy rates in high income countries, possibly because they have not broadly addressed the complex factors and motives that influence parents’ choices behind food selection for their infant (Du Toit et al., 2015; Koplin et al., 2016; Redsell et al., 2016). Therefore, a measure of parental motives for food selection on behalf of their infant is critically needed in order to understand and address risk factors for sub-optimal CFPs that may impact lifelong health risks for obesity and food allergies.

Worldwide, professional guidelines such as those published by the World Health Organization (WHO) (World Health Organization, 2021), the European Society for Pediatric Gastroenterology, Hepatology, and Nutrition (ESPGHAN) (Fewtrell, 2017), and the U.S. Department of Health and Human Services (USDHHS) and U.S. Department of Agriculture (USDA) (U.S. Department of Health and Human Services & United States Department of Agriculture, 2020) acknowledge the relationships between CFPs and health outcomes. Based on the evidence, key complementary feeding recommendations are to exclusively breastfeed until around 6 months old (Fewtrell et al., 2017; USDHHS & U.S. Department of Health and Human Services and U.S. Department of Agriculture, December 2020; WHO, 2021), at which time complementary foods, such as protein foods (e.g., meats), fruits, vegetables, and grains should be introduced (Fewtrell et al., 2017; USDHHS & U.S. Department of Health and Human Services and U.S. Department of Agriculture, December 2020; WHO, 2021). Recommendations in the United States (U.S.) are to also introduce the eight most common allergenic foods (e.g., peanuts, tree nuts, fish, shellfish, eggs, wheat, soy, and cow’s milk) along with the introduction of other complementary foods (USDHHS & U.S. Department of Health and Human Services and U.S. Department of Agriculture, December 2020). Further professional complementary feeding recommendations are to use a responsive feeding method, in which parents respond sensitively to infant hunger and fullness cues (USDHHS & U.S. Department of Health and Human Services and U.S. Department of Agriculture, December 2020; Fewtrell et al., 2017). Guidelines also recommend that complementary foods are offered as textures that are simple and easy for the infant to swallow, slowly advancing the texture and allowing for finger feeding as tolerated by the infant (from around eight months old) (Fewtrell et al., 2017; USDHHS & U.S. Department of Health and Human Services and U.S. Department of Agriculture, December 2020; WHO, 2021). Professional guidelines for CFPs help parents, healthcare providers, and policy makers understand best practices to promote health. However, health, although important, is unlikely to be the only factor influencing food selection during the complementary feeding period.

Healthcare providers often focus advice narrowly on health outcomes when communicating with their patients (Kneipp, 2023); however, many factors aside from potential health outcomes, such as safety, environment, and social support, contribute to health behaviors (Reis et al., 2000; Sallis et al., 2015). Indeed, multiple models have been used to demonstrate relationships of contextual and personal factors with food choices in adults (Furst et al., 1996; Steptoe et al., 1995). For example, using the Food Choice Model (Furst et al., 1996), researchers found personal and contextual factors, including sensory perceptions, monetary considerations, convenience, time, health and nutrition beliefs, social relationships, and perceived quality to influence adults’ food choices (Ares & Gámbaro, 2007; Jabs & Devine, 2006; Zenk et al., 2005). Similarly, developers of the Food Choice Questionnaire (Steptoe et al., 1995) found price, sensory appeal, convenience, ethical considerations, health, mood, natural content, familiarity, and weight control to be important priorities in adults’ food selection choices.

More recently, the Food Choice Questionnaire has been used to study general feeding motivation and feeding practices in parents of preschool and school-aged children; however, findings are limited and inconsistent. One study in Finland found parental motive of convenience to be positively associated with their pre-teens’ consumption of nutrientdense foods, while motives related to health, natural content, and sensory appeal demonstrated negative associations with those same foods (Roos et al., 2012). A study in Norway had similar findings in 12–13-year-olds (Oellingrath et al., 2013). A U.S. study with parents of 3–6-year-old children found motives of health and nutrition to be positively associated with fruit and vegetable intake (Kiefner-Burmeister et al., 2014). Conversely, in Australian parents of 2–5-year-old children, while health and nutrition were rated as important motivators for food selection, they were not associated with child fruit and vegetable preferences (Russell et al., 2015). In this same study, child preferences played a large role in parents’ choices about foods for their child (Russell et al., 2015).

There is a dearth of literature examining parents’ food selection motives on behalf of infants. One Norwegian study examined associations of parental food selection motives (on behalf of themselves) with infants’ fruit and vegetable intake (Røed et al., 2020). However, this study did not examine parental feeding motives on behalf of the infant.

Given that personal and contextual factors have been found to influence food selection choices in adults and parents of young children, it is conceivable that similar factors influence parents’ motives for food selection on behalf of their infant. However, there are likely differences in motivation behind parents’ food selection choices in the infant age group. For example, pickiness (i.e., food neophobia) peaks around 3 years old (Taylor et al., 2015). Therefore, the findings of Russell et al., that child preferences heavily influenced parents’ food selection motives for their 2–5 year olds, are likely different in the infant age group, who are not as likely to demonstrate pickiness (Russell et al., 2015). Additionally, complementary feeding marks a significant transition from a highly controlled diet of infant formula or breastmilk to one of diverse foods and textures, and therefore, motives behind parental food selection in this age group might be different than for adults selecting foods for themselves. Yet, there is no instrument available to measure parents’ food selection motives during the complementary feeding period. Furthermore, it remains unclear how and if personal and contextual factors are incorporated into clinical education practices by pediatric healthcare providers.

Gaps also exist in the ability to measure relationships of parents’ food choice motives with health outcomes in young children. The majority of instruments related to parents’ food selection during complementary feeding have been designed and used to study relationships of caregiver attitudes and motives with growth outcomes such as overweight or obesity (Baughcum et al., 2001; Thompson et al., 2009). For example, the Infant Feeding Styles Questionnaire (IFSQ) is a psychometrically valid measure that has the ability to identify caregiver feeding styles during the complementary feeding period (Thompson et al., 2009). However, the IFSQ only includes beliefs that map onto a feeding style framework that has been used to better understand obesity risk (Thompson et al., 2009). Similarly, the Infant Feeding Questionnaire (IFQ) has been used to measure maternal beliefs related to obesity and associated CFPs (Baughcum et al., 2001). However, recommendations related to the introduction of allergenic foods (DHHS and U.S. Department of Health and Human Services and U.S. Department of Agriculture, December 2020), along with an emphasis on responsive feeding, texture variety, and progression to finger feeding (Fewtrell et al., 2017; USDHHS & U.S. Department of Health and Human Services and U.S. Department of Agriculture, December 2020; WHO, 2021), necessitates the development of a measure to account for a broader set of food choice motives. This broader measure can be used to assess relationships to other potential health outcomes, as they may influence caregivers’ choices surrounding complementary feeding.

The purpose of this study was to develop an instrument to guide future research and inform broader clinical education practices related to complementary feeding. The aims of this study were to a) explore mothers’ beliefs and motives surrounding complementary feeding, b) develop an instrument to measure a broad set of parents’ motives for food selection on behalf of their infant, and c) analyze relationships of the instrument subscales with specific CFPs.

2. Phase 1

2.1. Methods

Phase 1 was a qualitative study of maternal beliefs and motives surrounding complementary feeding and the factors that influenced mothers’ selected CFPs.

2.1.1. Sample and setting

Given that mothers make the majority of early feeding decisions (Radzyminski & Callister, 2016; Thullen et al., 2016), this study exclusively sampled mothers, rather than other caregivers. Mothers were recruited via three methods. First, we recruited in person from the Special Supplemental Nutrition Assistance Program for Women, Infants, and Children (WIC) clinics located in county health departments in Virginia, U.S. Second, we recruited mothers online via Craigslist volunteer advertisements. Craigslist is an online classified advertisement platform located in the U.S. and Canada in which users can advertise or browse for jobs, houses, items for sale and available services by geographical region (washingtondc.craigslist.org). The third method of recruitment was via flyers posted at community sites, such as public libraries, coffee shops, laundromats, and community boards in neighborhoods. Participants were compensated for their time with a $25 electronic gift card. English-speaking mothers were eligible to participate if they were between 18 and 40-years-old with an infant between 6 and 19-months-old. Exclusion criteria included a medical issue of the infant that could impact infant feeding abilities and/or maternal complementary feeding motives and practices, such as Down syndrome, cleft lip/palate, diagnosed food allergy, metabolic disorder, or birth prior to 35 weeks estimated gestational age. The infant age ranges were selected to include infants who presented for WIC clinic visits during which complementary feeding is frequently discussed ( ± 1 month).

2.1.2. Procedures

A full description of the qualitative methods is available in a previously published manuscript that focused solely on maternal perceived threats surrounding complementary feeding and the strategies mothers used to mitigate such threats (Graf et al., 2022). In brief, the Vanderbilt University Institutional Review Board (IRB) and the Virginia Department of Health IRB approved all study procedures (study numbers 201179 and 4027, respectively). Between August and November 2019, eligible participants (n 27) took part in a private interview with the principal investigator (PI) (a registered nurse, family nurse practitioner, and PhD in nursing science student with prior qualitative research experience). A semi-structured interview guide, developed by a review of the literature and in collaboration with a team of experts, guided the interview. Digital audio files of the interviews were transcribed verbatim by a professional transcription service. The study team used qualitative inductive and deductive content analysis to explore maternal beliefs and motives surrounding complementary feeding and factors behind food selection during complementary feeding.

2.2. Results

Seven primary themes (including perceived threats) emerged as factors important to mothers’ complementary feeding choices: nutrition/health, limiting unhealthy foods, natural/organic, convenience, liked by baby, behavioral influence, and perceived threats. Table 1 includes a description of these themes and illustrative quotes. Based on these findings, the team reviewed currently existing measures of beliefs, motives, and factors related to food selection. The team identified a measure, the Food Choice Questionnaire (FCQ) (Steptoe et al., 1995), which has been used to explore food selection motives in adults. Given the overlap between the qualitative findings and the factors included in the Food Choice Questionnaire, the team chose to adapt and test the previously validated measure for use in mothers of infants and toddlers.

Table 1.

Themes and illustrative quotes for complementary feeding motives.

Themes and Description	Illustrative Quotes
Promoting Health and Nutrition  Participant discussed the importance of selecting foods that were high in nutritional content and/or foods that supported optimal growth and development.	• That’s very important …, making sure she’s getting the nutrition that she needs-lots of fruits and vegetables. (P3–13 months) • As long as they’re getting all of their recommended vitamins and nutrients … from those foods … that’s what’s most important. (P4–16 months)
Limiting Unhealthy Foods  Participant discussed the importance of avoiding foods that contained ingredients they considered to be unhealthy, such as salt, sugar, or fat.	• I just don’t think it’s healthy for babies to have a whole lot of salt or sugar. (P21–11 months) • I just try not to feed him obvious sweets like cookies and ice cream. (P18–16 months)
Natural/Organic  Participant discussed the importance of selecting foods that were natural or organic.	• We’ve been really adamant about if it’s organic. And if it’s not organic, we’re always looking at the ingredients on the labels now. (P13–6 months) • I don’t like the Gerber food like chicken and rice … If I’m going to give him chicken and rice, it’s going to be real chicken and rice. So I’m just sticking more to … stuff that’s just really natural. (P24–8 months)
Convenience  Participant discussed convenience as an important factor in food selection for their infant.	• Ease and convenience is something that is … very important to me. (P12–9 months) • Convenience is very important. It makes life easier. (P4–16 months)
Liked by Baby  Participant discussed the importance of selecting foods that the baby enjoyed eating.	• Taste is of course number one. If he doesn’t like it, it’s not even worth it because he’s not going to eat it. (P24–8 months) • I was raised in a house where you only get one meal. Like mommy is not making two and three meals. So I’ll try it again. But if I really know like he doesn’t like it, then I won’t try it again. (P26–11 months)
Behavioral Influence  Participant discussed the importance of selecting foods that would promote infant sleep, positive mood, contentedness. Examples included using food to cheer the infant up or to help with sleep.	• I have to feed her solids otherwise she’s cranky and hungry … breast milk isn’t just enough. (P15–9 months) • We were considering, “Well, should we give her oatmeal at night … to make her more full? Is that why she’s waking up?” … “Should we feed her to get her to sleep more?” (P16–12 months)
Perceived Threats  Participant discussed adverse outcomes they perceived to be associated with complementary feeding.	• That’s scary … just thinking about the food allergies and the choking hazard. (P20–13 months) • I wanted to avoid having potentially, a picky eater … so we just slowly started introducing … what seemed normal for us. (P16–12 months)

3. Phase 2

3.1. Methods

Phase 2 involved adaptation of the Food Choice Questionnaire, named the Parental Food Selection Questionnaire-Infant Version (PFSQ-I) after adaptation. Phase 2 also involved an anonymous web-based survey of U.S. mothers of infants, and exploratory factor analysis with principal components analysis of the PFSQ-I. Rather than confirmatory factor analysis, which is used to test an existing theory, exploratory factor analysis with principal components analysis was selected to facilitate conceptualization and categorization of complementary feeding motives (Zeynivandnezhad et al., 2019).

3.1.1. Sample and setting

The study team recruited mothers from across the U.S. to complete a survey about beliefs and motives related to complementary feeding. Inclusion and exclusion criteria were the same in Phase 2 as Phase 1, with the exception of infant age, which was increased from 6 to 19-months old (Phase 1) to 9 to 19-months old (Phase 2) to include only mothers of infants who had several months of experience with complementary feeding. Data collection took place over four days in July 2020. The Vanderbilt University Institutional Review Board reviewed the protocol for this study and deemed it exempt from IRB review.

Participants were recruited using three online recruitment platforms: Facebook, Vanderbilt University Research Notifications Distribution List (RNDL), and Cloud Research. For Facebook, the PI posted a brief study description and link in private parenting groups, following approval from the group’s moderator. For the RNDL, the PI sent an approved mass email with a study description and link to research volunteers at Vanderbilt University. The PI contracted with Cloud Research (formerly TurkPrime) (Litman et al., 2017), a research platform that includes a participant pool of over 50 million people, and with the ability to focus recruitment to participants based on specific demographic characteristics. Cloud Research shared the study description and survey link with participants who may have met eligibility criteria based on their reported demographic profiles.

3.1.2. Procedures

Participants used their personal device (e.g., computer, phone, or tablet) to complete eligibility screening and access the study online via Remote Electronic Data Capture (REDCap), a secure, online data collection system, located on a password protected server. Following eligibility screening, eligible participants received a Vanderbilt University IRB approved information sheet that explained the study purpose and procedures and informed participants that they could withdraw at any time. By beginning the survey, participants agreed to have read the information sheet and voluntarily agreed to participate.

Participants who enrolled from Facebook and RNDL received a $25 electronic gift card as compensation for their time. Cloud Research paid participants after completing the survey, according to the terms they agreed upon prior to the start. Compensation for Cloud Research participants is pre-determined by the participant, so researchers cannot calculate the amount of compensation each participant receives.

3.1.3. Questionnaire

Data were collected via an author-developed, anonymous, online, cross-sectional, 59- to 61-item questionnaire. The number of items varied depending on responses to some items about breastfeeding history and duration. Questions included personal and socio-demographic information and adapted versions of standardized measures to assess CFP and motives related to food selection. Prior to administration, the PI and four senior researchers with experience in survey design, infant nutrition, and maternal and child health research reviewed and refined the questionnaire (see section 3.1.3.1 for examples of changes), considering findings from the Phase 1 qualitative research to enhance content validity.

3.1.3.1. Adapted food choice questionnaire.

Phase 2 also involved adaptation and psychometric testing of the Food Choice Questionnaire (Steptoe et al., 1995). In the original development of the Food Choice Questionnaire, exploratory factor analysis with varimax rotation identified 36 items which clustered into nine factors: health, mood, convenience, sensory appeal, natural content, price, weight control, familiarity, and ethical concern. In the original Food Choice Questionnaire, subjects were asked to endorse the statement, “it is important to me that the food I eat on a typical day …” for each of the 36 items. Responses ranged from 1 = not at all important, 2 = a little important, 3 = moderately important, and 4 very important. Cronbach’s alpha for each of the factors ranged from 0.70 to 0.87. Repeat administration of the original Food Choice Questionnaire at two separate times found that correlations between scores at two administrations were >0.70 for all factors, suggesting acceptable test-retest reliability.

To adapt this measure, mothers were asked to endorse a statement, “it is important to me that the food I feed my baby on a typical day …” The context of the items was changed from “me” to “my baby”. Response options remained the same as for the original Food Choice Questionnaire. Because ethical concern did not emerge as a motive in food selection for complementary infant feeding in Phase 1, items from that factor were omitted. Additionally, items from the familiarity factor were omitted because they did not translate well into the adapted version. The wording of one item was changed from “helps me relax” to “helps him/her sleep.” Two items within the mood factor were omitted: “helps me cope with stress” and “helps me to cope with life,” and instead, the following item was included: “helps him/her fuss less.” These omissions left 27 of the original items. Because avoiding sugar was frequently discussed as important in the qualitative interviews, one item, “is low in sugar” was added to the measure. Because mothers frequently reported concerns about feeding related health outcomes, one item from The Experiences of Mealtimes Tool was included in the PFSQ-I (Brown & Lee, 2011). The item was worded as, “I worry about my baby choking on food,” with response options ranging from 0 = strongly disagree, to 5 = strongly agree. Two additional items were added: “I worry about my baby having an allergic reaction when s/he eats a new food” and “I worry about my baby becoming a picky eater when s/he is older.” The final adapted version of the PFSQ-I included 33 items. Although price was not reported as an important motive in food selection in Phase 1, the study team chose to retain this factor from the original measure, given that mothers may have been hesitant to discuss this topic with the interviewer.

3.2. Data analysis

All analyses were performed using SPSS (version 28.0). Descriptive statistics (i.e., means, medians, and frequencies) were used to examine distributions and appropriately summarize the sample characteristics and all variables. An exploratory factor analysis was run on the 33 items of the PFSQ-I using principal components analysis with varimax rotation.

3.3. Results

3.3.1. Participant characteristics

A total of 401 mothers from a wide geographic area across the U.S. completed the survey. After removal of data provided by participants who did not meet inclusion criteria, the final sample included 381 mothers. Mean maternal age was 30.4 years (Standard Deviation (SD) – 5.1) and infant age was 14.1 months (SD 3.4). About one-third of infants were less than 12 months old (35.4%), while about two-thirds were 12–19 months old (64.6%). Median years of maternal education was 16 (Interquartile Range (IQR) 13–18), and median maternal body mass index was slightly overweight (Median 26, IQR 23–31). Most mothers were multiparas (57.5%) and breastfed this infant 6 months or longer (63.3%). A majority were employed (54.1%), had commercial health insurance (64.2%), were married or living with a partner (87.1%), and reported an annual household income of $60 K or greater (51.0%). In comparison, the U.S. median household income was $67,521 in 2020 (Shrider et al., 2021). A majority identified as non-Hispanic/non-Latinx (88.7%) and white/Caucasian (78.2%). About half of participants were from the U.S. South (53.0%), with the remaining half of the sample living in the Northeast (12.2%), Midwest (18.1%), and West (15.7%).

3.3.2. Exploratory factor analysis of adapted food choice questionnaire

Table 2 summarizes the results of the exploratory factor analysis that was run on the 33 items of the PFSQ-I using principal components analysis with varimax rotation. The Kaiser-Meyer-Olkin statistic was 0.865 indicating an adequate sample size. Bartlett’s test of sphericity was statistically significant (x² – 428.3, p < 0.001, df – 528), indicating the sample was suitable for factoring. Items with factor loadings >0.50 were retained. Items that did not load clearly on a single factor (i.e., <0.15 difference between two separate factors loadings) were discarded, with a total of three items removed. Seven factors had Eigen values > 1.0. Examination of the scree plot revealed that a 7- to 9-factor solution best represented the data, therefore, 6-, 7-, 8-, and 9-factor solutions were examined for variance explained, cross loadings of items, and conceptual clarity of the derived items. Through this process, it was determined that a 7-factor solution was the best solution. The final structure included 30 items and 7 factors that together accounted for 61.0% of the total variance. Eigen values ranged from 1.15 to 8.09. Factor 1 contains five items representing behavioral influence as a motive underlying food selection, such as to help with sleep, fussiness, or alertness. Factor 2 contains five items representing health promotion motives for food selection, such as keeps him/her healthy or is nutritious. Factor 3 contains 7 items representing food selection motives related to specific ingredients, such as additives, sugar, fat, fiber, and calories. Factor 4 contains 4 items related to the affordability of food, such as good value for the money. Factor 5 contains 3 items related to the sensory appeal of food, such as the smell, taste, or appearance. Factor 6 contains 3 items related to the convenience of food preparation, such as can be cooked very simply. Factor 7 contains 3 items about perceived threats related to complementary feeding, such as choking, allergic reaction, and pickiness. Supplemental File A contains all abbreviated items and factor loadings of items included in the exploratory factor analysis.

Table 2.

Final PFSQ-I factors in order of loading, sample items, number of items, and reliability estimates of the factors (n = 381).

Factor and Item Examples	Number of Items	Mean (SD)	Cronbach’s α
1 Behavioral Influence Helps him/her with sleep Helps him/her fuss less	5	2.6 (0.8)	0.83
2 Health Promotion Keeps him/her healthy Is nutritious	5	3.6 (0.5)	0.76
3 Ingredients Contains no additives Contains no artificial ingredients	7	2.8 (0.6)	0.80
4 Affordability Is not expensive Is cheap	4	2.9 (0.7)	0.80
5 Sensory Appeal Smells nice Looks nice	3	2.8 (0.070)	0.70
6 Convenience easy to prepare Can be cooked very simply	3	2.8 (0.7)	0.78
7 Perceived Threats Having an allergic reaction when s/he eats a new food Becoming a picky eater when s/he is older	3	2.9 (1.2)	0.68

3.3.3. Internal consistency reliability

For each factor, henceforth referred to as subscale, scores were computed by calculating the mean of unweighted ratings for the individual items. Cronbach’s alpha (α) values assessed internal consistency reliability of the total PFSQ-I and the subscales. Table 2 depicts the mean subscale scores and Cronbach’s α value for each subscale. The overall PFSQ-I demonstrated good internal consistency reliability (α = 0.87), and the seven subscales demonstrated moderate to good internal consistency reliability. The highest rated subscale was Health Promotion (Mean = 3.60), while the lowest rated subscale was Behavioral Influence (Mean = 2.56). Table 3 depicts the bivariate correlations between the subscales. Positive correlations existed between the subscales, except for health and perceived threats, for which no statistically significant correlation existed. The correlation among subscales is expected in social sciences research, given that overlap of thoughts and behaviors is common in humans (Costello & Osborne, 2005). The small-to-moderate correlations suggest that these subscales indeed captured unique constructs and that the sample is acceptable for factoring.

Table 3.

Pearson correlations of the PFSQ-I subscales (n = 381).

s	1	2	3	4	5	6
1. Behavioral Influence	–	–	–	–	–	–
2. Health Promotion	0.38	–	–	–	–	–
3. Ingredients	0.49	0.38	–	–	–	–
4. Affordability	0.40	0.14	0.19	–	–	–
5. Sensory Appeal	0.49	0.35	0.37	0.34	–	–
6. Convenience	0.34	0.14	0.21	0.55	0.30	–
7. Perceived Threats	0.23	0.07*	0.15	0.18	0.12	0.17

^*p > 0.05.

4. Phase 3

4.1. Methods

To assess validity of the new measure, Phase 3 involved examination of the PFSQ-I in relation to complementary feeding practices (CFPs). Data for this phase was collected along with the data for Phase 2, thus the setting, sample, and procedures were the same for Phase 3 as for Phase 2.

4.1.1. Instrument

Four CFPs were selected based on prior complementary feeding research (Brown & Lee, 2011; Huh et al., 2011; Seiverling et al., 2011): a) timing/type of complementary food introduction, and b) usual texture intake, c) frequency of feeding method (e.g., spoon feeding and finger feeding), and d) allergenic food introduction. For timing/type of complementary food introduction, participants selected from four retrospective response options to indicate the age at which they first fed infant cereal, other solid foods (e.g., fruits, vegetables, eggs, or meats), or drinks other than milk or water to their infant. Response options were 0 = I have not fed this to my baby, 1 = < 2 months, 2 = 2–3 months, 3 = 4–5 months, and 4 = 6+ months. For analyses, these responses were compressed into 1 <4 months, 2 4–5 months, and 3 6 months (Huh et al., 2011). To assess feeding method, participants responded to two items, for which they selected on a sliding scale ranging from 0 to 100 percent of the time, indicating the percentage of time a) they fed their infant with a spoon, and b) the infant self-fed with their fingers (Brown & Lee, 2011). To assess usual texture intake, participants responded to a dichotomous, yes/no response item indicating if, on a daily or usual basis, their infant ate each of four food textures: a) pureed, b) ground/lumpy, c) cut up/chunky, and d) dry/crispy) (Seiverling et al., 2011). Developmentally appropriate examples were included for each texture. To assess allergenic food introduction, participants checked a box (either yes/no) to indicate if their infants had ever eaten each of the eight foods considered highly allergenic in the U.S.: peanuts, eggs, fish, shellfish, wheat, soy, tree nuts, and cow’s milk (DHHS and U.S. Department of Health and Human Services and U.S. Department of Agriculture, December 2020).

4.1.2. Data analysis

For Phase 3, all analyses were performed using SPSS (version 28.0). Descriptive statistics were used to examine and summarize distributions of the CFPs. Given that CFPs tend to progress according to infant age and development (DHHS and U.S. Department of Health and Human Services and U.S. Department of Agriculture, December 2020), bivariate correlations (e.g., point biserial and Spearman’s Rho) were used to examine associations of usual texture intake, allergenic food introduction, and feeding method with infant age. Spearman’s Rho was used to examine associations of PFSQ-I subscales with timing/type of complementary food introduction. Skewed variables were transformed using rank transformation. Pearson’s r was used to examine associations of PFSQ-I subscales with feeding method; multiple linear regression analyses controlled for infant age. Simple logistic regression analyses were used to examine associations of PFSQ-I subscales with usual texture intake and allergenic food introduction; multiple logistic regression analyses controlled for infant age. An alpha value of p < 0.05 was considered statistically significant.

4.2. Results

4.2.1. Summary of complementary feeding practices

A total of 381 mothers provided complete data needed for the analyses. For timing/type of complementary food introduction, most mothers retrospectively reported first introducing any complementary foods to their infant between 4 and 5 months old (44.6%), with 14.2% introducing complementary foods prior to 4 months old. Infant cereal was the item most commonly introduced prior to 4 months old (12.6%). A majority of mothers reported introducing other solid foods (e.g., fruits, vegetables, eggs, and meats) and drinks after 6 months old (71.4% and 71.1%, respectively).

Bivariate associations were used to examine relationships among infant age and the other CFPs (usual texture intake, allergenic food introduction, and feeding mothed). Each of the CFPs was associated with infant age, except for ground/lumpy texture consumption, for which there was no statistically significant association. The majority of associations were positive, except for the regular consumption of pureed foods and the percentage of time utilizing a spoon-feeding method, which demonstrated inverse associations with infant age. These findings were as expected, given that feeding of pureed foods and utilization of a spoon-feeding method should decline, as infants reach developmental milestones over time. For a full summary of complementary feeding practices and bivariate associations with infant age, see Supplementary File B.

4.2.2. Summary of unadjusted and adjusted associations of complementary feeding practices with infant food choice questionnaire subscales

Fig. 1 summarizes the associations of the PFSQ-I subscales with timing/type of complementary food and drink introduction. Since this was a retrospective report of the timing and type of complementary food introduction, and the age of the infant at entry into the study varied from 9 to 19 months old, adjusted associations for infant age were not calculated for this CFP. The strongest associations were between the behavioral influence subscale and timing of infant cereal and drink introduction (inverse for both). This indicates that mothers who rated the importance of behavioral influence as higher tended to report introducing infant cereal and drinks at an earlier age. The ingredients and perceived threats subscales were inversely associated with the timing of infant cereal introduction, while the affordability and sensory appeal subscales were inversely associated with the timing of introduction of drinks. This indicates that mothers who rated ingredients and/or perceived threats as more important tended to have introduced infant cereal at an earlier age, and mothers who rated affordability and/or sensory appeal as more important tended to have introduced non-milk/non-water drinks to their infant at an earlier age. Analyses found no statistically significant associations of the health promotion or convenience subscales with the timing of introduction of any of the complementary feeding items, nor of the PFSQ-I subscales with the introduction of other solid foods. (Fig. 1).

Fig. 1.

Associations of the PFSQ-I subscales with timing and type of complementary food and drink introduction (n = 381) using Spearman’s rho. *p < 0.05, **p < 0.001.

Unadjusted associations were calculated for the PFSQ-I with feeding method, texture type, and allergenic food introduction. Adjusted associations were also calculated, controlling for infant age at the time of the study. There were no changes in relationships or statistical significance when adjusting for infant age; therefore, only the adjusted associations are reported in Table 4.

Table 4.

Associations of the PFSQ-I subscales with complementary feeding practices adjusted for infant age.

PFSQ-I Subscales	Complementary Feeding Practices
	Usual Texture Intake (yes/no) (n = 377)				Feeding Method (%) (n = 380)		Allergenic Food Introduction (yes/no) (n = 380)
	Pureed Food	Ground/Lumpy	Cut-up/Chunky	Dry/Crispy	Spoon	Fingera	Peanut	Eggs	Wheat	Shellfish	Fish	Soy	Cow’s Milk	Tree Nuts
	OR 95% CI	OR 95% CI	OR 95% CI	OR 95% CI	β	β	OR 95% CI	OR 95% CI	OR 95% CI	OR 95% CI	OR 95% CI	OR 95% CI	OR 95% CI	OR 95% CI
Behavioral Influence	1.39* 1.05–1.84	1.20 0.84–1.72	0.59* 0.38–0.92	0.77 0.50–1.19	0.35**	− 0.21**	0.47** 0.34–0.66	0.54* 0.35–0.83	0.49** 0.32–0.73	0.56** 0.41–0.75	0.47** 0.35–0.63	0.57** 0.43–0.75	0.64 0.41–1.01	0.64* 0.49–0.86
Health Promotiona	1.35 0.84–2.16	1.92* 1.11–3.33	1.24 0.61–2.52	1.62 0.83–3.16	0.13*	−0.05	0.90 0.53–1.54	1.15 0.58–2.29	1.05 0.55–2.00	1.19 0.73–1.93	0.90 0.57–1.43	0.92 0.59–1.44	1.10 0.53–2.30	0.99 0.62–1.57
Ingredients	1.78* 1.23–2.60	1.37 0.85–2.21	0.44* 0.24–0.78	0.84 0.47–1.	.20**	−0.09	0.45** 0.29–0.69	0.59 0.34–1.03	0.51* 0.31–0.85	0.84 0.58–1.21	0.66* 0.46–0.94	0.65* 0.46–0.92	0.53* 0.29–0.95	0.91 0.64–1.30
Affordability	1.21 0.90–1.63	1.07 0.73–1.58	0.74 0.47–1.17	1.48 0.93–2.34	0.14*	−0.03	0.89 0.64–1.24	1.13 0.73–1.76	0.68 0.44–1.03	0.78 0.58–1.06	0.66* 0.49–0.89	0.70* 0.52–0.94	1.04 0.65–1.65	0.86 0.64–1.16
Sensory Appeal	0.92 0.68–1.25	1.14 0.77–1.71	1.08 0.68–1.74	1.01 0.62–1.63	0.18**	−0.00	0.77 0.55–1.09	1.28 0.80–2.04	0.74 0.48–1.13	1.05 0.76–1.43	0.77 0.57–1.04	0.80 0.60–1.09	0.79 0.48–1.28	1.04 0.76–1.41
Conveniencea	1.46* 1.08–1.97	1.01 0.69–1.47	1.00 0.63–1.57	0.98 0.61–1.55	0.11*	−0.05	0.75 0.53–1.04	0.98 0.63–1.52	0.61* 0.40–0.95	0.76 0.57–1.02	0.67* 0.50–0.90	0.84 0.64–1.12	0.70 0.43–1.14	0.93 0.69–1.24
Perceived Threats	1.47** 1.22–1.77	0.92 0.73–1.17	0.91 0.68–1.21	1.05 0.79–1.39	0.09	−0.10*	0.86 0.70–1.06	0.84 0.63–1.12	0.85 0.66–1.11	0.78* 0.65–0.94	0.77* 0.64–0.93	0.87 0.73–1.04	0.92 0.68–1.25	0.90 0.75–1.07

Note: Multiple linear regression analyses and multiple logistic regression analyses, controlling for infant age. OR is an abbreviation for Odds Ratio; CI is an abbreviation for Confidence Interval.

^*p < 0.05

^**p < 0.001.

^aSkewed variable transformed to normal distribution via rank transformation.

For feeding method, the behavioral influence subscale was positively associated with spoon feeding and inversely associated with finger feeding. This indicates that mothers who rated behavioral influence as being more important utilized a spoon-feeding method more often and a finger feeding method less often to feed their infant. The health promotion, ingredients, affordability, sensory appeal, and convenience subscales were positively associated with spoon feeding, indicating that mothers who found these subscales to be more important tended to report utilizing a spoon-feeding method more often. Conversely, the perceived threats subscale was inversely associated with finger feeding, indicating that mothers who found perceived threats to be more important in their food selection choices tended to use a finger feeding method less often.

For usual texture intake, the behavioral influence and ingredients subscales were statistically significantly positively associated with usual or regular feeding of pureed foods and inversely associated with feeding of cut-up/chunky foods. These associations indicate that mothers who rated behavioral influence and ingredients as being more important were more likely to feed pureed foods and less likely to feed cut-up/chunky foods, compared to mothers who considered behavioral influence and ingredients to be less important. The subscale of health promotion was positively associated with usual feeding of ground/lumpy texture. Mothers who rated health promotion as more important were more likely to regularly feed ground/lumpy foods, compared to mothers who rated health promotion as less important. The subscales of convenience and perceived threats were positively associated with usual intake of pureed foods, indicating that mothers who found these subscales to be more important were more likely to report regularly feeding pureed foods, compared to mothers who found them to be less important. There were no associations of any of the PFSQ-I subscales with feeding of dry/crispy texture.

For allergenic food introduction, the behavioral influence subscale was inversely associated with the introduction of each of the allergenic foods, except for cow’s milk. This indicates that mothers who found behavioral importance to be more important in their food selection choices were less likely to have introduced allergenic foods to their infant, compared to mothers who found behavioral influence to be more important. The ingredients subscale was inversely associated with the introduction of peanuts, wheat, fish, soy, and cow’s milk, indicating that mothers who rated ingredients as more important, were more likely to have introduced those allergenic foods, compared to mothers who rated it as less important. The affordability subscale was inversely associated with the introduction of fish and soy, the convenience subscale with wheat and fish (both inverse), and the perceived threats subscale with shellfish and fish (both inverse). There were no associations of the health promotion or sensory appeal subscales with the introduction of any allergenic foods. (Table 4).

5. Discussion

This work provides a new instrument to measure a broad set of factors influencing caregivers’ food selection choices for their infant. The exploratory factor analysis of the PFSQ-I showed seven distinct factors to be important in mothers’ reported food selection choices for their infant: behavioral influence, health promotion, ingredients, affordability, sensory appeal, convenience, and perceived threats. Findings showed good internal consistency reliability. The methods used in development of the PFSQ-I also support initial validity of the instrument. The systematic process of qualitative inquiry in Phase 1 grounded the early development of the instrument and justified adaptation of the original Food Choice Questionnaire, rather than development of an entirely new instrument. Coupled with the systematic process of item adaptation, development and testing, this study gathered evidence to support the content validity of the PFSQ-I.

Comparison of PFSQ-I subscales with CPFs also supported construct validity. The findings of relationships among PFSQ-I subscales and CFPs was consistent with the literature. For example, some of the items that clustered within the behavioral influence subscale (e.g., helps him/her sleep and helps him/her fuss less) have been included in prior work examining feeding practices and styles (Haycraft, 2020; Thompson et al., 2013, 2021). Pressuring or controlling feeding practices, which include the behavior of using food to soothe or help with sleep are associated with early introduction of complementary foods (Haycraft, 2020; Thompson et al., 2013, 2021). Similarly, mothers who rated behavioral influence as more important also tended to introduce infant cereal and drinks to their infant at an earlier age.

Also supporting construct validity, many of the relationships between PFSQ-I subscales and CFPs were in expected directions. For example, we predicted that mothers who reported convenience to be more important, tended to also select pureed foods and a spoon-feeding method more often. In the U.S., where this research took place, many commercially prepared baby foods come in pureed form, require minimal preparation, and are known to be safe and convenient sources of nutrition for infants. This is especially true for baby foods provided by WIC (i.e., nutrition education and support programs for families demonstrating financial need), which only provides payment for infant cereal and fruits, vegetables, and meats in pureed form for infants in the 6–12-month-old age range. Similarly, when convenience is important, mothers may strive to avoid messiness associated with infant finger feeding, and therefore select a spoon-feeding method in order to have more control over the environment. This thought also potentially explains why the behavioral influence subscale was correlated with the provision of pureed food and use of a spoon-feeding method, while inversely related to the feeding of cut-up chunky foods and the use of a finger feeding method. To elaborate, mothers who rated behavioral influence as more important responded to individual items indicating that it was important to them to select foods that achieved improved sleep and decreased fussiness in their infant. It is likely that mothers who reported behavioral influence as more important were likely under more stress and therefore, selected CFPs that required less thought, preparation, and mess (e.g., pureed food and spoon feeding, rather than cut-up/chunky foods and finger feeding). Similarly, it is probable that mothers who rated behavioral influence as more important were less likely to have introduced the allergenic foods because they were more concerned with controlling their environment, and less able to plan ahead or deal with the potential for an allergic reaction. These relationships could be assessed in future studies.

As predicted, mothers who rated affordability as more important tended to use a spoon-feeding method more often. For infants 6–12-months-old, the state and federal nutrition assistance program for families with financial need, WIC provides monetary assistance only for purchasing of commercially prepared pureed foods, rather than fresh fruits and vegetables (USDA, 2022). Mothers who were recipients of WIC were likely more concerned with the affordability of food, compared to mothers with higher annual incomes, and also more likely to feed pureed foods (which require a spoon-feeding method), rather than fresh fruits and vegetables, because of the WIC purchasing guidelines in this age group. However, it is unclear why the affordability subscale was not associated with pureed food intake.

Of note, despite health being rated as an important motive in food selection for adults (Furst et al., 1996; Steptoe et al., 1995), and the health promotion subscale having the highest mean score of importance in this sample, the health promotion subscale was only associated with two of the CPFs examined (positive associations with ground/lumpy texture intake and spoon feeding). Given this finding, it may be useful to tailor clinical education and interventions to include the broader contextual factors that we found to influence CFPs, such as behavioral influence, ingredients, and convenience, rather than focusing on health implications alone.

Given the frequency of discussion of perceived threats in the qualitative phase of this study, it was expected that the perceived threats subscale would be associated with a greater number of CFPs than found in the quantitative analyses. It is possible that choking, food allergies, and pickiness are conceptually different concerns, and they clustered together due to the way they were scored. Additional analyses are needed to investigate whether these items perform better as individual items, rather than part of the perceived threats subscale.

Given the qualitative findings in Phase 1 that led to adaptation of the original Food Choice Questionnaire, it was expected that there would be some differences in clustering of factors important in food selection, compared to factors in the original Food Choice Questionnaire. While some of the subscales clustered similarly to the original Food Choice Questionnaire (e.g., sensory appeal and convenience) (Steptoe et al., 1995), other subscales were entirely unique (e.g., behavioral influence, perceived threats, and ingredients). Additionally, this measure provides new items within the perceived threats subscale (i.e., concern about pickiness, allergic reaction, and choking) that can be used individually to study maternal decision making surrounding complementary feeding in future research. Within Phase 3, the items that assessed allergenic food introduction were also newly developed and should be used in future research to further explore CFPs and their relationship with additional health outcomes, such as food allergies.

5.1. Limitations

As with every research study, this study was not without limitations. One limitation was the racial and ethnic distribution of the sample of participants who completed the questionnaire. A high percentage of participants identified as white/Caucasian and non-Hispanic/non-Latinx, thus limiting generalizability of the findings. Additionally, this sample only included mothers. Additional psychometric testing in a larger and more racially and ethnically heterogenous sample, along with the inclusion of perspectives of other caregivers, such as fathers, grandparents, daycare providers, or extended family members, would further support internal and external validity of the instrument.

This study also presented several potential limitations in measurement. First, we did not collect contact information from participants, and therefore, we were unable to conduct test-retest reliability analyses. However, the original Food Choice Questionnaire showed good test-retest reliability. Future studies would benefit from examining test-retest reliability of the PFSQ-I. Second, this study involved retrospective data collection for some of the complementary feeding practices (e. g., timing of complementary food introduction and history of allergenic food introduction). It is possible that recall bias was present. However, two recent studies using parental recall of complementary feeding practices found consistent results across samples, suggesting that parental recall was an effective way to measure these behaviors (Fu et al., 2018; Komninou et al., 2019). Third, the health promotion subscale was not associated with many of the measured complementary feeding practices, which was not expected. It is possible that other complementary feeding practices may be more closely related to the PFSQ-I subscales.

To further test the validity of the PFSQ-I, it would be useful to analyze the relationship of the PFSQ-I with additional complementary feeding practices, such as fruit and vegetable intake, which have been associated with parental health motives in prior research (Røed et al., 2020). Various sociodemographic and psychosocial factors have been associated with infant feeding practices. For example, higher maternal stress and depressive symptoms have been associated with lower rates of breastfeeding initiation and duration (Oyetunji & Chandra, 2020). Similarly, elevated household chaos and decreased maternal emotional responsiveness during mealtimes have been linked to child eating behaviors, such as increased emotional overeating and food responsiveness (Saltzman et al., 2019). Perceived infant fussiness has also been linked to initiation, duration, and discontinuation of breastfeeding, as well as timing of complementary food introduction (Vilar-Compte et al., 2022). Although beyond the scope of the present study to measure and report, subgroup analyses could be helpful to examine independent and interactive effects of these factors, including parental stress, household chaos level, infant fussiness, and/or socioeconomic status on complementary feeding motives and CFPs.

The wording of the PFSQ-I is a potential limitation of the measure. The wording may suggest that feeding motives are stable when they likely vary by time of day and context, and age of the infant. However, this was an adaptation of the original Food Choice Questionnaire, aiming to capture general motives related to complementary feeding; therefore, the wording of the questions remained as similar to the original as possible. Similarly, because the same rater responded to all of the items, common method biases were possible, especially related to social desirability and transient mood states (Podsakoff et al., 2003). However, the study was anonymous, thus limiting the potential for social desirability biases. Furthermore, varied formatting of the items assessing the complementary feeding practices limited the potential for method effects produced by item characteristics (Podsakoff et al., 2003). Finally, this was a cross-sectional study and therefore, no causal relationships can be assumed. Longitudinal analyses are needed to explore whether parental feeding motives may predict infant health outcomes, such as obesity, pickiness, and food allergies.

5.2. Conclusion

Despite the reported limitations, this PFSQ-I provides a psychometrically sound measure of mothers’ food selection motives for their infant. This study reports the early stages of development and testing of the PFSQ-I. Preliminary evidence supports reliability and validity of this new measure of a broad set of factors influencing caregivers’ complementary feeding decisions. The documented associations between the PFSQ-I subscales and CFPs indicates that the PFSQ-I is an effective instrument in measuring complementary feeding motives; however, more testing is needed to determine their linkage to complementary feeding behaviors. Findings present opportunities for clinical education and intervention. These motives are likely established prior to feeding practices (i.e., behaviors), and may be more amenable to intervention prior to the onset of adverse health outcomes. Next steps for the PFSQ-I include additional testing in a more diverse sample and examination of relationships of PFSQ-I subscales with additional CFPs and health outcomes, also utilizing longitudinal methods.

Abbreviations

CFPs

Complementary Feeding Practices

DHHS

U.S. Department of Health and Human Services

USDA

U.S. Department of Agriculture

CFPs

Complementary Feeding Practices

FCQ

Food Choice Questionnaire

PFSQ-I

Parental Food Selection Questionnaire-Infant Version

RNDL

Research Notifications Distribution List

REDCap

Remote Electronic Data Capture