Research Article Technology Use and Interest Among Low-Income Parents of Young Children: Differences by Age Group and Ethnicity Taren M. Swindle, PhD1; Wendy L. Ward, PhD2; Leanne Whiteside-Mansell, EdD1; Patti Bokony, PhD1; Dawn Pettit, MSW1 ABSTRACT Objective: To examine demographic differences in frequency of use of technologies and interest in receiving nutrition information via technology by low-income parents and caregivers. Design: Descriptive, cross-sectional study. Setting: Head Start and state-funded child care programs. Participants: A total of 806 parents and caregivers from low-income families. Variables Measured: A 20-item survey assessed frequency of use and interest in technologies (dependent variables) and collected participant age and ethnicity (independent variables). Analysis: Multivariate ANOVA analysis investigated whether age, ethnicity, and their interactions were related to frequency of use and interest in technology types. Results: Daily rates of usage for Internet, text messaging, and cell phone use were over 60%. However, Twitter and blogs were accessed daily by < 13% of respondents. The omnibus 2-way interaction of ethnicity and age was nonsignificant. However, main effects for ethnicity (Wilks’ l ¼ .85; F ¼ 3.13; P < .001) and age (Wilks’ l ¼ .89; F ¼ 2.29; P < .001) were observed. Conclusions and Implications: Facebook, e-mail, texting, and smartphone applications may be innovative modalities to engage with low-income parents and caregivers aged # 45. However, some strategies may be ineffective for reaching Hispanic families as they reported less use of the Internet, Facebook, and e-mail as well as less interest in e-mail. Key Words: technology, child nutrition sciences, electronic e-mail, social media, low-income population (J Nutr Educ Behav. 2014;46:484-490.) Accepted June 7, 2014. Published online July 23, 2014.
INTRODUCTION Early nutrition predicts long-term cognitive, social, and physical health outcomes.1-3 However, many children aged < 5 are not meeting daily recommendations for minimum nutri-tion, particularly those in lowincome families.4 Parents are the gatekeepers for the food that is purchased and prepared for their children but may need support to gain the knowledge and skills necessary to make changes. Thus, interventions aiming to improve nutrition for young
children must target parents also. In accordance with the Health Belief Model,5 parents will be more likely to follow recommendations when they understand the benefits of the recommendation and perceive the seriousness of failing to follow recommendations. Thus, intervention programs can have an important role in increasing adherence to dietary recommendations for children by raising parents' awareness of the associated health risks of a poor diet for children and advocating the health benefits of a quality diet.
1 Department of Family and Preventive Medicine, University of Arkansas for Medical Sciences, Little Rock, AR 2 Department of Pediatrics, University of Arkansas for Medical Sciences, Little Rock, AR Address for correspondence: Taren M. Swindle, PhD, Department of Family and Preventive Medicine, University of Arkansas for Medical Sciences, 4301 W Markham St, #530, Little Rock, AR 72205-7199; Phone: (870) 236-0997; Fax: (501) 686-8421; E-mail: tswindle@ uams.edu PUBLISHED BY ELSEVIER INC. ON BEHALF OF THE SOCIETY FOR NUTRITION EDUCATION AND BEHAVIOR http://dx.doi.org/10.1016/j.jneb.2014.06.004
484
Despite the potential positive outcomes, low-income parents and caregivers face many barriers to enrolling and attending intervention programs.6,7 Time constraints, child care needs, transportation issues, and work conflicts are among the challenges parents report regarding enrolling and remaining engaged in a parent education program. Recruitment rates for families of low socioeconomic status are below 31%8 and attrition rates are high even if parents are enrolled successfully and provided with transportation and child care.9 Innovative intervention delivery and engagement modalities are needed to overcome barriers to participation and connect with parents despite their limiting circumstances. In the current digital age, technology holds potential to address this need. Nutritionists seeking to reach lowincome parents and caregivers must know whether and how families use technology if it is to be a feasible intervention modality. By 2009, 35% of US
Journal of Nutrition Education and Behavior Volume 46, Number 6, 2014
Journal of Nutrition Education and Behavior Volume 46, Number 6, 2014 families making < $20,000 reported broadband Internet connections in their homes.10 Another recent report showed that 44% of all age groups and 61% of young Americans (aged 14–24 years) living in poverty used the public library for computer and Internet access.11 In addition to Internet access on computers, modern technology allows for the access and delivery of information via cell phones. In 2009, 32% of American adults reported accessing the Internet with a cell phone or other mobile device.12 Over 70% of American adults indicated that they used their cell phone to text.13 These studies suggest that many low-income families access information via technology, but these studies do not indicate how often low-income families use various technologies or their preferences for receiving information. Effective efforts to reach lowincome families need to be informed by the types of Internet sites (eg, blogs, e-mail, social media) most typically visited by these families as well as the rate at which they are viewed (eg, daily, weekly). For example, data suggest that 47% of US adults and 72% of young adults (aged 18–29 years) use Facebook for social networking.14 Another social media site, Twitter, is used by 13% of adults, with the highest rates of use by younger adults (18%) and African Americans (25%).15 The Pew Research group reported that 74.3% of US adults use e-mail and 25.3% read blogs.16 Yet results from these studies do not consider patterns of use within lowincome families. Despite widespread technology use, there are documented digital discrepancies in the general population. Whites in the US are more likely to be daily Internet users, but this effect is confounded with a higher average income. However, African Americans are the most likely group to use mobile Internet, offsetting the inequity in traditional modes of access.12 African Americans and English-speaking Hispanic adults are more likely (7%) than whites to own a cell phone and are more intense users of their cell phones, making calls and texting more frequently.13 Adults aged > 65 are the least likely to have cell phones,13 and age is negatively related to having a computer in the home and use
of the Internet in any location.17 Demographic differences in access and use of technology within a sample of low-income families are unknown. Two exceptions are noted.18,19 Among Supplemental Nutrition Assistance Program recipients in Indiana, whites and younger adults (aged # 50) were the most likely to have a computer and Internet access in the home.19 In another study, usage of the Internet by Supplemental Nutrition Assistance Program recipients in New Jersey was related to age but not ethnicity.18 However, the Indiana sample was overwhelmingly white (85%) and the New Jersey sample, although more racially balanced than the Indiana sample, was modest in size (n ¼ 93). Replication and extension of these findings are needed in additional locations and among large and diverse samples of low-income families. It is possible that participant characteristics could interact in unique ways to influence use of technologies. For example, the impact of age may vary across ethnicity. No studies were identified exploring such possibilities. A final consideration is warranted if technology is to be a viable option for intervention and information delivery for low-income families. Access and use of technology do not necessarily overlap with interest in receipt of information from intervention and education programs. However, a few studies suggest families may be open to using technology in this way. For example, a Pew survey of US adults found that 46% of respondents agreed that mobile Internet was important for ‘‘getting information on the go.’’12 Hesse and colleagues20 reported that the Internet is already the most used resource for health information for US adults. Before designing technology-based nutrition programs for low-income parents, research is warranted to determine whether these interests in technology extend to parents for learning information on nutrition. In light of this gap, the current study sought to determine whether technology is a viable and preferred avenue of information delivery among low-income parents and caregivers. Specifically, this study examined the following questions: (1) How often are various technologies (eg, cell phones, Facebook) accessed by low-income families? (2) Do low-income families
Swindle et al 485 express interest in technology as a mode of information delivery? (3) How do frequency of use and interest differ by age and ethnicity within a low-income sample? (4) Do demographic factors interact to have a unique impact on frequency of or interest in technology use?
METHODS Study Design The researchers recruited sites for this descriptive, cross-sectional study in a Southern state from early childhood programs targeting low-income families in 2013. Recruitment sought to reflect the geographic region (urban vs rural) and ethnic composition of children typically served by these programs in the state. Urban was defined using the US Census Bureau classification of an urbanized area of $50,000 people.21 Of 20 total sites, 11 were in urbanized areas (population range: 65,934– 195,314) and 9 were in rural areas (population range: 1,745–36,295).22 A total of 806 parents and caregivers were surveyed with 65.5% from urban and 34.5% from rural sites. Data collectors completed standardized training to minimize procedural differences. To collect surveys, trained data collectors positioned themselves during drop-off and pickup times to request completion of a survey from parents and caregivers. Participants were paid a $1 coin after completing the survey. Data collectors offered assistance to every individual and aided (ie, read aloud and recorded responses) any participant who indicated interest in help. Data collectors were trained to monitor for signs of confusion and repeat offers of assistance. The survey was approved and deemed as minimal risk by the University of Arkansas for Medical Sciences Institutional Review Board. Consent was not required.
Setting Early child care programs included in this study were Head Start (n ¼ 5 urban, 4 rural) or state-funded early child care programs (n ¼ 6 urban, 5 rural). Head Start is a governmentfunded early child care program that serves low-income families (#100% of poverty) with children birth to age 5. As a part of this state's efforts to
486 Swindle et al provide quality child care and education to children, children in lowincome families can qualify to attend state-funded child care. This program serves children in families that do not qualify for Head Start but fall at # 200% of poverty. Between Head Start and state-funded child care, 25,950 children affected by poverty are served. The 20 sites selected represent a convenience sample based on directors' willingness to participate and accessibility by data collectors. However, sites were included from every geographic quadrant of the state.
Participants All parents and caregivers entering and exiting the center were asked to complete a survey. Participation was voluntary and anonymous; 806 parents and caregivers elected to complete the survey. Signatures were collected from participants on a separate log sheet to account for expenditure of the $1 coins and minimize duplication. At larger centers ($5 classrooms) that required $ 2 survey collection periods, data collectors were positioned at the same entry and exit points as at previous collections to further minimize duplication. When several parents entered or exited at the same time, data collectors announced a general invitation to participate in the survey. Multiple data collectors with clipboards were available to facilitate simultaneous completion of surveys. Although most parents who were invited participated, some children were picked up by after-school programs, which precluded parental invitation. As such, it was not possible to estimate an accurate rate of refusal to the survey. When parents and caregivers declined, the most common reason was time constraints. The total rate of participation at these centers was 29% based on enrollment statistics. Because Head Start and statefunded classrooms were often in the same building with the same pickup and drop-off locations, distinction between the types of classrooms could not be made reliably. Thus, results were combined for the groups.
Variables Measured The authors developed a 1-page, 20-item survey to assess parents' and
Journal of Nutrition Education and Behavior Volume 46, Number 6, 2014 caregivers' frequency of use and interest in various technologies (ie, dependent variables) and collect demographic information (ie, independent variables). A panel of earlychildhood teachers and parent service coordinators reviewed the survey for readability and ease of completion. In addition, the survey was piloted in 2012 with over 500 respondents and then modified for the 2013 survey to improve clarity, readability, and formatting based on participant and data collector feedback. Pilot survey participants were recruited using the same criteria as the full sample and were similar in terms of race, geography, age, and income. The final version of the survey was at a 3.2 Flesch-Kincaid reading level. The extensive preliminary testing and review of the instrument with both parents and teachers attest to the face and content validity of the instrument. Furthermore, refinements made to the instrument in light of pilot feedback served to reduce participant confusion (ie, error variance) and thus increase the reliable variance in participant responses. Typical assessment of internal consistency (eg, Cronbach alpha) was not appropriate to assess reliability because the items were checklists or single item constructs. Participants were asked, ‘‘How often do you use each of these?’’ Responses were given for Internet, cell phone (for texting), smartphone (for Internet access), Facebook, Twitter, e-mail, and blogs. Participants indicated frequency of use on a 1 (‘‘never’’) to 6 (‘‘daily’’) scale. In addition, participants were asked to ‘‘Indicate all the ways (yes or no) that you would like to receive information on the following topics: (a) what's going on at your child's center (b) help with parenting, and (c) healthy eating.’’ The former 2 topics (eg, center activities and parenting help) served as a reference for interest in nutrition information. For each topic, participants indicated whether they were interested in using each technology (Facebook, text messaging, phone calls, twitter, e-mail, and blogs) for receiving information (‘‘yes’’) or if they were not (‘‘no’’). These questions reflect 3 changes that were made based on feedback from the pilot survey: (1) the response scale for frequency of technology use was expanded to 6 intervals instead
of 4 to reflect participant request for greater refinement; (2) the instructions and response options for indicating interest in methods of information delivery were modified to have parents indicate ‘‘yes’’ or ‘‘no’’ for each topic for each technology rather than asking them to ‘‘mark all that apply,’’ to address confusion in pilot testing; and (3) Internet and smartphone were not included in the list of methods of information delivery through which parents indicated interest in receiving new information. This final change was made because pilot participants consistently indicated confusion about the intent of these methods. Recurring comments from pilot participants were that they were unsure whether the Internet and smartphone options were intended to be self-directed (eg, a Google search) or externally driven (eg, a Web page or phone application housing information).
Statistical Analyses The researchers calculated descriptive statistics for individual items to document rates of use and interest for each technology type (frequencies). An overall interest variable was created for each technology mode by summing across the positive responses for each technology. For example, a ‘‘yes’’ response for interest in receiving Facebook updates on parenting and healthy eating, but not activities at the center, would correspond to a Facebook Interest score of 2. Thereafter, a MANOVA approach was selected to account for the intercorrelations of the dependent variables (ie, frequency of use and overall interest across technologies). Follow-up ANOVAs and Bonferroni post hoc comparisons were conducted when the omnibus effects were at statistically significant levels. Wilks' lambda values for the omnibus analysis were considered an indication of effect size. Because of the number of tests, a conservative P ¼ .01 was adopted for omnibus tests.
RESULTS Descriptives Most respondents were female (76.1%). About a third (34.2%) identified as white, 41.2% identified as African American, 19.5% identified
Journal of Nutrition Education and Behavior Volume 46, Number 6, 2014
Swindle et al 487
Table 1. Frequency of Use and Interest in Delivery of Information, by Technology Type (n ¼ 806)
Frequency: ‘‘How Often Do You Use Each of These?’’
Internet
A Few 1–3 1–2 3–6 Never Times/y Times/mo Times/wk Times/wk Daily 6.1 3.1 5.1 6.9 9 69.9
Interest: ‘‘Indicate All the Ways (Yes or No) That You Would Like to Receive Information on the Following Topics’’ What Is Going on at the Center –
Parenting Help –
Healthy Eating –
Text messaging
4.5
2.5
2.8
3.3
6.3
80.6
64.6
49.4
53.8
Cell phone (for calls)
1.3
1.3
0.9
2.2
3.5
90.9
58.5
47.7
43.1
Smartphone (for Internet)
21.1
1.7
2.2
2.2
4.3
68.4
–
–
–
Facebook
13.7
3.5
6.9
7.6
11.3
57.0
52.3
47.1
53.2
Twitter
72.2
4.9
4.4
3.2
2.8
12.5
6.6
4.8
6.4
E-mail
15.4
4.1
8.1
8.5
10.9
53.0
68.8
62.2
64.5
Blogs
75.6
5.6
4.5
2.7
2.4
9.2
6.8
9.3
11.4
Note: Numbers presented are percentages. For Frequency, rows total 100%. For Interest, participants were encouraged to select as many or as few ‘‘yes’’ answers as they liked. Thus, interest columns or rows do not total to 100%. Rather, percentages represent the number of participants who endorsed the options or interest. Participants were not asked about interest in Internet and smartphone because of ambiguity about initiation. Additional rationale is offered in the text. as Hispanic, and 4.8% identifed as multiracial or ‘‘other.’’ Of the entire sample, 17.6% requested and completed the survey in Spanish. On average, participants were 33.14 years of age (SD, 10.08 years; range. 16–74 years). To examine for generational differences in use and interest in technology, age was grouped into 4 categories for inclusion in the analysis: 16–25 (n ¼ 150), 26–35 (n ¼ 285), 36–45 (n ¼ 100), and $46 years (n ¼ 76). Of those surveyed, 77.9% were parents of the child, 12.8% were grandparents of the child, 7.8 % were other relatives, and 1.5% were partners of a parent. Table 1 presents the proportions of participants responding in each frequency and interest category for each technology type. Most technologies (6 of 8) were used daily by $ 50% of the sample. The most popular daily technology uses were cell phones for calls (90.9%), cell phones for texting (80.6%), the Internet (69.9%), smartphones for Internet (68.4%), and Facebook (57%). More than half of participants used e-mail on a daily basis (53.0%). When weekly usage was considered ($1–2 times/wk), $ 72% of participants used most technologies (6 of 8) on a regular basis. However, most families surveyed did not access Twitter or blogs regularly.
Approximately 75% said they never used those options at all. Table 1 also presents the percentage of respondents indicating each technology type as a preferred information delivery mode on the topics presented. E-mail was the technology of most interest overall, with over 60% indicating interest across topics. Text messaging and Facebook were also of interest for $ 47% of participants across topics, similar to rates of indicated interest for phone calls. Interest in Facebook and text messaging for receiving information on healthy eating were within 11% of the most preferred technology for that topic, e-mail. As seen in Table 1, interest in receiving information on healthy eating varied by technology type, ranging from being of lowest comparative interest via texts and phone calls and of highest comparative interest via Facebook.
Demographic Differences The omnibus 2-way interaction of ethnicity and age was nonsignificant. However, main effects for age (Wilks' l ¼ .87; F ¼ 1.98; P < .001) and ethnicity (Wilks' l ¼ .79; F ¼ 3.39; P < .001) were observed. Table 2 presents mean usage rates by ethnicity and age with significant group differences denoted. The significant
omnibus main effect of age was explored to reveal a significant univariate effect on 5 dependent variables: use of Internet (F ¼ 8.82; P < .001), Facebook (F ¼ 9.26; P < .001), e-mail (F ¼ 4.73; P ¼ .003), texting (F ¼ 4.98; P ¼ .002), and smartphone for Internet use (F ¼ 10.66; P < .001). Where group differences were observed, the oldest age ($45 years) used technology least frequently. However, this group was not significantly less likely to indicate interest in any type of technology for receiving information. Because of the significance of the overall test for ethnicity, the authors examined the univariate main effects. Significant univariate main effects for ethnicity were obtained for use of Internet (F ¼ 11.18; P < .001), Facebook (F ¼ 6.07; P < .001), Twitter (F ¼ 6.20; P < .001), e-mail (F ¼ 11.99; P < .001) and interest in e-mail (F ¼ 3.57; P ¼ .014), and phone call (F ¼ 6.39; P < .001). Hispanics used Internet, Facebook, Twitter, and e-mail significantly less than at least 1 other group and were significantly less interested in the use of e-mail than all other groups (Table 2). African Americans used Twitter significantly more than all other ethnic groups and were more interested in receiving phone calls than were whites.
Journal of Nutrition Education and Behavior Volume 46, Number 6, 2014
488 Swindle et al
Table 2. Mean Frequency of Parent Technology Use and Preference, by Age and Ethnicity 17–25 y (n ¼ 150)
26–35 y (n ¼ 285)
36–45 y (n ¼ 100)
45–55 y (n ¼ 76)
Total (n ¼ 611)
Age
Mean
SD
Mean
SD
Mean
SD
Mean
SD
Mean
SD
P
Internet use
5.5a
1.1
5.4a
1.3
5.5a
1.2
4.6b
2.0
5.3
1.4
< .001
Facebook use
4.9a–c
1.6
5.1ac
1.6
4.5bc
2.0
3.7d
2.2
4.8
1.8
.02
Twitter use
2.2
1.9
2.00
1.8
1.9
1.7
1.5
1.4
2.0
1.8
NS
E-mail use
4.4a–c
1.9
4.9ab
1.7
4.9ab
1.8
4.0c
2.1
4.7
1.8
.002
Blog use
1.6
1.4
1.8
1.6
2.1
1.9
1.3
1.0
1.7
1.6
NS
Cell phone use for texting
5.7a
0.9
5.6a
1.2
5.7a
1.0
4.6b
1.9
5.5
1.3
< .001
Cell phone use for calls
5.9
0.6
5.8
0.8
5.9
0.5
5.6
1.1
5.8
0.8
NS
Smartphone use for Internet
5.2ab
1.7
5.1a–c
1.9
4.5bc
2.3
3.2d
2.4
4.8
2.1
.03
Facebook interest
1.0
1.3
1.2
1.3
0.8
1.2
0.8
1.2
1.0
1.3
NS
Twitter interest
0.1
0.6
0.1
0.5
0.8
0.4
0.9
0.1
0.1
0.5
NS
E-mail interest
1.4
1.4
1.4
1.4
1.4
1.4
0.9
1.3
1.3
1.4
NS
Blog interest
0.2
0.6
0.2
0.6
0.2
0.7
0.1
0.4
0.2
0.6
NS
Texting interest
1.2
1.3
1.2
1.3
1.1
1.3
0.8
1.1
1.1
1.3
NS
Phone call interest
1.0
1.2
0.9
1.2
0.9
1.2
1.0
1.2
0.9
1.2
NS
White (n ¼ 209) Ethnicity Internet use
Mean
SD
African American (n ¼ 254) Mean SD
5.5a
1.3
5.5a
Facebook use
a
5.1
1.7
4.7
Twitter use
1.6a
1.4
2.5b a
a
ab
Hispanic (n ¼ 119)
Multiracial/ Other (n ¼ 29)
Total (n ¼ 611)
Mean
SD
Mean
SD
Mean
SD
1.3
4.7b
1.7
5.5a
1.1
5.3
1.4
.02
1.8
b
4.3
2.0
5.1
1.5
4.8
1.8
.001
2.1
1.6a
1.4
1.6a
1.5
2.0
1.8
.03
b
a
ab
E-mail use
5.0
1.7
4.9
1.7
3.6
2.1
4.7
2.0
4.7
1.8
.02
Blog use
1.8
1.6
1.8
1.6
1.6
1.5
1.7
1.4
1.7
1.6
NS
Cell phone use for texting
5.6ab
1.3
5.7a
1.1
5.2b
1.5
5.3ab
1.5
5.5
1.3
.005
Cell phone use for calls
5.9
0.5
5.8
0.8
5.8
0.8
5.6
1.4
5.9
0.8
NS
Smartphone use for Internet
4.6a
2.2
5.2b
1.8
4.5a
2.2
4.6ab
2.2
4.8
2.1
Facebook interest
1.1
1.3
1.0
1.2
1.0
1.3
1.2
1.3
1.0
1.3
NS.
Twitter interest
0.1
0.4
0.1
0.6
0.1
0.4
0.03
0.2
0.1
0.5
NS
E-mail interest
1.4a
1.4
1.5a
1.4
0.9b
1.3
1.4a
1.4
1.3
1.4
.01
.008
Blog interest
0.2
0.7
0.1
0.6
0.1
0.5
0.1
0.5
0.2
0.6
NS
Texting interest
1.0
1.2
1.3
1.3
1.0
1.2
0.8
1.0
1.1
1.3
NS
Phone call interest
0.7a
1.1
1.1b
1.3
0.9ab
1.2
0.8ab
1.0
0.9
1.2
< .001
NS indicates comparisons that had no significantly different group means. Note: Where significant group differences were found based on Bonferroni post hoc mean comparisons (adjusted P < .05), homogenous groups are indicated by superscripts. The largest significant adjusted P for the set of comparisons is given.
DISCUSSION Rates of technology use by this lowincome sample exceeded previous study estimates12,13,15,16,18,23 and rates observed in the 2012 pilot survey.24 Results suggest that lowincome families find means to access the Internet and that cell phone use by low-income families is similar to
rates of cell phone use by those of higher socioeconomic status.25 Additionally, this study documented how often low-income families use technology and the types of social media outlets they do (and do not) frequent. Text messaging, Facebook, and e-mail reached over 70% of this sample on a weekly basis. Rates of use for these technologies among this low-income
sample are consistent with or higher than rates found in broader samples of the US public.13,14,16,23 Furthermore, the technologies least likely to be used by low-income parents and caregivers in this sample seem to match those least used by the general public. Twitter and blogs, specifically, were used by < 20% of participants regularly, which is similar to reported
Journal of Nutrition Education and Behavior Volume 46, Number 6, 2014 rates of general public use from the Pew Research center.15,16 Previous research on demographic differences in technology use suggested that white adults used the internet19 and Facebook26 more than other ethnic groups; African Americans used Twitter15 and mobile Internet12 more than other ethnic groups; young adults used Facebook14 and Twitter15 more than older age groups; and older adults used cell phones less.27 Consistent with previous research, African Americans in this sample were the most frequent users of Twitter and used mobile Internet (ie, smartphones) more than other ethnic groups. In addition, the oldest age group (> 45 years) used Facebook, cell phones for texting, and mobile Internet significantly less than younger adults. Contrary to previous findings, whites used Internet at rates similar to African American and multiracial participants and did not differ from African Americans in Facebook use. Furthermore, there were no age group differences in use of Twitter and cell phones for calls. In addition to providing information on rates and types of technology used by demographic subgroup, this study assessed parents' and caregivers' interest in receiving information via technology. According to the Health Belief Model,5 the individual's value of a health goal is an important predictor of health behavior. Thus, interest in receiving nutrition information is a marker for potential to improve behaviors. In the present study, interest in receiving parenting, child care, or nutrition information was > 50% for at least 1 topic for 4 technology types. Furthermore, only 2 differences in interest were observed across demographic subgroups despite the observed differences in use. Less use of a technology corresponded to less interest for only 1 ethnic group (Hispanics) on 1 mode of delivery (e-mail). In this sample, groups using technologies less frequently were still interested in using these avenues to receive information. Furthermore, the similarity in rates of interest between the topics of healthy eating, parenting help, and activities at the center suggests that nutrition is an area in which many parents were interested in receiving information.
This study recruited to approximate a racial balance similar to low-income families in the state, but the sampling procedure was not adequate to ensure a representative sample. Replications in other states are needed to provide additional support for these findings and to extend understanding of technology use to additional ethnic groups (eg, Native Americans). In addition, the self-selective nature of this survey may have created a selection bias that affected findings. Although the authors were unable to collect data to estimate how parents and caregivers who refused to participate or were absent from the center were different from participating parents, these differences could impact demographic patterns.
IMPLICATIONS FOR RESEARCH AND PRACTICE Data from the present study suggest that technology may be an important tool in closing the gap between those served by nutrition education programs and those interested in service who are not reached. In a survey of 68 Head Start mothers, the greatest reported barriers to involvement were schedules that conflicted with Head Start activities and having another child at home who required care.28 Through technology, parents and caregivers could engage in an online community on their own schedules. Technology could be used to reach parents with new material (eg, recipes), increase program participation (eg, virtual discussions), and obtain needed information (eg, updates on progress or barriers to goals). Technology-based programs also promote an environmentally protective practice (ie, ‘‘green’’) and require less expenditure on materials for the intervention. With successful technologybased campaigns such as text4baby29 emerging, the current study's findings can inform the selection of information delivery modalities that fit the targeted population's use of and interest in technology types. Findings on ethnic and age differences, for example, can inform the development of interventions for specific populations. Although technology-based programs seem to be a
Swindle et al 489 viable option for many low-income families based on access and use, the current study suggests that there are barriers in certain subsamples. For example, older caregivers in this sample reported using several technologies less frequently. Technology efforts to reach this population may require a combination of participant training and supplementation with other approaches (eg, phone calls). Similarly, programs working with Hispanic families should consider using text messaging rather than Internet-based efforts. Further research is needed to better understand the use of technology to reach low-income parents and caregivers. Respondents' reported level of interest in technologies for receiving information may not translate to interest in participation in full-blown interventions. Efficacy trials with participation estimates are needed to determine whether interest levels in this study are consistent with parent and caregiver involvement in implemented technology-based programs. Most important, the successes with and barriers to using technology to reach low-income parents and caregivers must be documented, as well as the impact of these technologybased interventions on targeted behaviors. The current study attests to the potential feasibility, but not utility, of such programs. It is possible that effectiveness of technology interventions would be greater with low-income families compared with programs delivered through in-person contacts because of the many barriers families face to recruitment and retention in onsite programs. The mobile nature of a technology intervention may be more flexible to the frequent physical moves of low-income families.30 In addition, research is needed to determine whether interest in technology as an information pathway increases with exposure and experience. Surveying low-income parents and caregivers about their preference for technology information delivery before and after a technology-based intervention could provide evidence on this issue. In addition, there are other ways to use technology for intervention (eg, Web sites, smartphone applications) that were not assessed in the current survey. Future studies could provide additional detail on the sites that
490 Swindle et al low-income parents and caregivers visit and how smartphones are used. More study is also warranted to determine the optimal ways to reach families who are the hardest to access (ie, Hispanic and older caregivers).
ACKNOWLEDGMENTS This project was supported by Agriculture and Food Research Initiative Competitive Grant 2011-6800130014 from the US Department of Agriculture, National Institute of Food and Agriculture.
REFERENCES 1. Golley RK, Smithers LG, Mittinty MN, Emmitt P, Northstone K, Lynch JW. Diet quality of UK infants is associated with dietary, adiposity, cardiovascular, and cognitive outcomes measured at 78 years. J Nutr. 2013;143:1611-1617. 2. Slopen N, Fitzmaurice G, Williams DR, Gilman SE. Poverty, food insecurity, and childhood internalizing and externalizing behaviors. J Am Acad Child Adolesc Psychiatry. 2010;49:444-452. 3. Zeisel SH. Epigenetic mechanisms for nutrition determinants of later health outcomes. Am J Clin Nutr. 2009;89: 14885-14935. 4. Lorson BA, Melgar-Quinonez HR, Taylor CA. Correlates of fruit and vegetable intake in children. J Am Diet Assoc. 2009;109:474-478. 5. Becker MH, Maiman LA. Sociobehavioral determinants of compliance with health and medical care recommendations. Med Care. 1975;13:10-24. 6. Baker CN, Arnold D, Meagher S. Enrollment and attendance in a parent training prevention program for conduct problems. Prev Sci. 2011;12:126-138. 7. Winslow EB, Bonds D, Wolchik S, Sandler I, Braver S. Predictors of enrollment and retention in a preventive parenting intervention for divorced families. J Prim Prev. 2009;30:151-172. 8. Heinrichs N, Bertram H, Kuschel A, Hahlweg K. Parent recruitment and retention in a universal prevention program for child behavior and emotional problems: barriers to research and program participation. Prev Sci. 2005;6:275-286.
Journal of Nutrition Education and Behavior Volume 46, Number 6, 2014 9. Huebner CE. Evaluation of a clinicbased parent education program to reduce the risk of infant and toddler maltreatment. Public Health Nurs. 2002;19:377-389. 10. Rainie L. Internet, broadband, and cell phone statistics. http://www.pewinternet .org/Reports/2010/Internet-broadbandand-cell-phone-statistics.aspx. Accessed July 10, 2014. 11. Becker S. First-ever national study: millions of people rely on library computers for employment, health, and education. Lib Times Int. 2010;26:47. 12. Horrigan J. Wireless Internet use. http:// www.pewinternet.org/Reports/2009/12Wireless-Internet-Use.aspx. Accessed July 10, 2014. 13. Lenhart A. Cell phones and American adults. http://www.pewinternet.org/ Reports/2010/Cell-Phones-and-AmericanAdults.aspx. Accessed July 10, 2014. 14. Lenhart A, Purcell K, Smith A, Zickuhr K. Social media and young adults. http://www.pewinternet.org/2010/02/ 03/social-media-and-young-adults/. Accessed July 10, 2014. 15. Smith A. Twitter update 2011. http:// www.pewinternet.org/Reports/2011/ Twitter-Update-2011.aspx. Accessed July 10, 2014. 16. Zickuhr K. Generations 2010. http:// www.pewinternet.org/2010/12/16/gener ations-2010/. Accessed July 10, 2014. 17. Ono H, Zavodny M. Digital inequality: a five country comparison using microdata. Soc Sci Res. 2007;36: 1135-1155. 18. Corda K, Palmer D, Adler A. New Jersey EFNEP and FSNE program participant Internet usage: a study to examine the efficacy of using the Internet to deliver nutrition education. J Nutr Educ Behav. 2008;40:S25. 19. Neuenschwander LM, Abbott A, Mobley AR. Assessment of lowincome adults’ access to technology: implications for nutrition education. J Nutr Educ Behav. 2012;44:60-65. 20. Hesse BW, Nelson DE, Kreps GL, Croyle RT, Arora NK, Rimer BK. Trust and sources of health information: the impact of the Internet and its implications for health care providers. Findings from the first Health Information National Trends survey. Arch Intern Med. 2005;165:2618.
21. Department of Commerce. Federal Register: Urban criteria for the 2010 census 2011. http://www.census.gov/ geo/www/ua/fedregv76n164.pdf. Accessed July 10, 2014. 22. US Census Bureau. Current Population Survey (CPS): a joint effort between the Bureau of Labor Statistics and the Census Bureau. http://www.census. gov/cps/about/cpsdef.html. Accessed July 10, 2014. 23. Lenhart A. Social & mobile Internet use among teens and young adults. http:// pewinternet.org/Reports/2010/SocialMedia-and-Young-Adults.aspx. Accessed July 10, 2014. 24. Swindle T, Whiteside-Mansell L, Ward W, Bokony PA, Pettit D. Technology use and preference by low-income parents of young children: demographic patterns and implications for intervention. Presented to the 2013 biennial meeting of the Society for Research in Child Development; Seattle, WA; April, 2013. 25. Lane W, Manner C. The impact of personality traits on smartphone ownership and use. Int J Business Soc Sci. 2011;17:22-28. 26. Hampton KN, Goulet LS, Purcell, K. Social networking sites and our lives: how people’s trust, personal relationships, and civic and political involvement are connected to their use of social networking sites and other technologies. http://www.pewinternet.org/ 2011/06/16/social-networking-sites-andour-lives/. Accessed July 10, 2014. 27. Lenhart A. Cell phones and American adults. http://www.pewinternet.org/ Reports/2010/Cell-Phones-and-AmericanAdults.aspx. Accessed July 10, 2014. 28. Lamb-Parker F, Piotrkowski CS, Baker AJL, Kessler-Sklar S, Clark B, Peay L. Understanding barriers to parent involvement in Head Start: a research–community partnership. Early Child Res Q. 2001;16:35-51. 29. Evans WD, Wallace JL, Snider J. Pilot evaluation of the text4baby mobile health program. BMC Public Health. 2010;12:1031. 30. Ma T, Gee L, Kushel MB. Associations between housing instability and food insecurity with health care access in low-income children. Ambulatory Pediatrics. 2008;8:50-57.
INTRODUCTION Early nutrition predicts long-term cognitive, social, and physical health outcomes.1-3 However, many children aged < 5 are not meeting daily recommendations for minimum nutri-tion, particularly those in lowincome families.4 Parents are the gatekeepers for the food that is purchased and prepared for their children but may need support to gain the knowledge and skills necessary to make changes. Thus, interventions aiming to improve nutrition for young
children must target parents also. In accordance with the Health Belief Model,5 parents will be more likely to follow recommendations when they understand the benefits of the recommendation and perceive the seriousness of failing to follow recommendations. Thus, intervention programs can have an important role in increasing adherence to dietary recommendations for children by raising parents' awareness of the associated health risks of a poor diet for children and advocating the health benefits of a quality diet.
1 Department of Family and Preventive Medicine, University of Arkansas for Medical Sciences, Little Rock, AR 2 Department of Pediatrics, University of Arkansas for Medical Sciences, Little Rock, AR Address for correspondence: Taren M. Swindle, PhD, Department of Family and Preventive Medicine, University of Arkansas for Medical Sciences, 4301 W Markham St, #530, Little Rock, AR 72205-7199; Phone: (870) 236-0997; Fax: (501) 686-8421; E-mail: tswindle@ uams.edu PUBLISHED BY ELSEVIER INC. ON BEHALF OF THE SOCIETY FOR NUTRITION EDUCATION AND BEHAVIOR http://dx.doi.org/10.1016/j.jneb.2014.06.004
484
Despite the potential positive outcomes, low-income parents and caregivers face many barriers to enrolling and attending intervention programs.6,7 Time constraints, child care needs, transportation issues, and work conflicts are among the challenges parents report regarding enrolling and remaining engaged in a parent education program. Recruitment rates for families of low socioeconomic status are below 31%8 and attrition rates are high even if parents are enrolled successfully and provided with transportation and child care.9 Innovative intervention delivery and engagement modalities are needed to overcome barriers to participation and connect with parents despite their limiting circumstances. In the current digital age, technology holds potential to address this need. Nutritionists seeking to reach lowincome parents and caregivers must know whether and how families use technology if it is to be a feasible intervention modality. By 2009, 35% of US
Journal of Nutrition Education and Behavior Volume 46, Number 6, 2014
Journal of Nutrition Education and Behavior Volume 46, Number 6, 2014 families making < $20,000 reported broadband Internet connections in their homes.10 Another recent report showed that 44% of all age groups and 61% of young Americans (aged 14–24 years) living in poverty used the public library for computer and Internet access.11 In addition to Internet access on computers, modern technology allows for the access and delivery of information via cell phones. In 2009, 32% of American adults reported accessing the Internet with a cell phone or other mobile device.12 Over 70% of American adults indicated that they used their cell phone to text.13 These studies suggest that many low-income families access information via technology, but these studies do not indicate how often low-income families use various technologies or their preferences for receiving information. Effective efforts to reach lowincome families need to be informed by the types of Internet sites (eg, blogs, e-mail, social media) most typically visited by these families as well as the rate at which they are viewed (eg, daily, weekly). For example, data suggest that 47% of US adults and 72% of young adults (aged 18–29 years) use Facebook for social networking.14 Another social media site, Twitter, is used by 13% of adults, with the highest rates of use by younger adults (18%) and African Americans (25%).15 The Pew Research group reported that 74.3% of US adults use e-mail and 25.3% read blogs.16 Yet results from these studies do not consider patterns of use within lowincome families. Despite widespread technology use, there are documented digital discrepancies in the general population. Whites in the US are more likely to be daily Internet users, but this effect is confounded with a higher average income. However, African Americans are the most likely group to use mobile Internet, offsetting the inequity in traditional modes of access.12 African Americans and English-speaking Hispanic adults are more likely (7%) than whites to own a cell phone and are more intense users of their cell phones, making calls and texting more frequently.13 Adults aged > 65 are the least likely to have cell phones,13 and age is negatively related to having a computer in the home and use
of the Internet in any location.17 Demographic differences in access and use of technology within a sample of low-income families are unknown. Two exceptions are noted.18,19 Among Supplemental Nutrition Assistance Program recipients in Indiana, whites and younger adults (aged # 50) were the most likely to have a computer and Internet access in the home.19 In another study, usage of the Internet by Supplemental Nutrition Assistance Program recipients in New Jersey was related to age but not ethnicity.18 However, the Indiana sample was overwhelmingly white (85%) and the New Jersey sample, although more racially balanced than the Indiana sample, was modest in size (n ¼ 93). Replication and extension of these findings are needed in additional locations and among large and diverse samples of low-income families. It is possible that participant characteristics could interact in unique ways to influence use of technologies. For example, the impact of age may vary across ethnicity. No studies were identified exploring such possibilities. A final consideration is warranted if technology is to be a viable option for intervention and information delivery for low-income families. Access and use of technology do not necessarily overlap with interest in receipt of information from intervention and education programs. However, a few studies suggest families may be open to using technology in this way. For example, a Pew survey of US adults found that 46% of respondents agreed that mobile Internet was important for ‘‘getting information on the go.’’12 Hesse and colleagues20 reported that the Internet is already the most used resource for health information for US adults. Before designing technology-based nutrition programs for low-income parents, research is warranted to determine whether these interests in technology extend to parents for learning information on nutrition. In light of this gap, the current study sought to determine whether technology is a viable and preferred avenue of information delivery among low-income parents and caregivers. Specifically, this study examined the following questions: (1) How often are various technologies (eg, cell phones, Facebook) accessed by low-income families? (2) Do low-income families
Swindle et al 485 express interest in technology as a mode of information delivery? (3) How do frequency of use and interest differ by age and ethnicity within a low-income sample? (4) Do demographic factors interact to have a unique impact on frequency of or interest in technology use?
METHODS Study Design The researchers recruited sites for this descriptive, cross-sectional study in a Southern state from early childhood programs targeting low-income families in 2013. Recruitment sought to reflect the geographic region (urban vs rural) and ethnic composition of children typically served by these programs in the state. Urban was defined using the US Census Bureau classification of an urbanized area of $50,000 people.21 Of 20 total sites, 11 were in urbanized areas (population range: 65,934– 195,314) and 9 were in rural areas (population range: 1,745–36,295).22 A total of 806 parents and caregivers were surveyed with 65.5% from urban and 34.5% from rural sites. Data collectors completed standardized training to minimize procedural differences. To collect surveys, trained data collectors positioned themselves during drop-off and pickup times to request completion of a survey from parents and caregivers. Participants were paid a $1 coin after completing the survey. Data collectors offered assistance to every individual and aided (ie, read aloud and recorded responses) any participant who indicated interest in help. Data collectors were trained to monitor for signs of confusion and repeat offers of assistance. The survey was approved and deemed as minimal risk by the University of Arkansas for Medical Sciences Institutional Review Board. Consent was not required.
Setting Early child care programs included in this study were Head Start (n ¼ 5 urban, 4 rural) or state-funded early child care programs (n ¼ 6 urban, 5 rural). Head Start is a governmentfunded early child care program that serves low-income families (#100% of poverty) with children birth to age 5. As a part of this state's efforts to
486 Swindle et al provide quality child care and education to children, children in lowincome families can qualify to attend state-funded child care. This program serves children in families that do not qualify for Head Start but fall at # 200% of poverty. Between Head Start and state-funded child care, 25,950 children affected by poverty are served. The 20 sites selected represent a convenience sample based on directors' willingness to participate and accessibility by data collectors. However, sites were included from every geographic quadrant of the state.
Participants All parents and caregivers entering and exiting the center were asked to complete a survey. Participation was voluntary and anonymous; 806 parents and caregivers elected to complete the survey. Signatures were collected from participants on a separate log sheet to account for expenditure of the $1 coins and minimize duplication. At larger centers ($5 classrooms) that required $ 2 survey collection periods, data collectors were positioned at the same entry and exit points as at previous collections to further minimize duplication. When several parents entered or exited at the same time, data collectors announced a general invitation to participate in the survey. Multiple data collectors with clipboards were available to facilitate simultaneous completion of surveys. Although most parents who were invited participated, some children were picked up by after-school programs, which precluded parental invitation. As such, it was not possible to estimate an accurate rate of refusal to the survey. When parents and caregivers declined, the most common reason was time constraints. The total rate of participation at these centers was 29% based on enrollment statistics. Because Head Start and statefunded classrooms were often in the same building with the same pickup and drop-off locations, distinction between the types of classrooms could not be made reliably. Thus, results were combined for the groups.
Variables Measured The authors developed a 1-page, 20-item survey to assess parents' and
Journal of Nutrition Education and Behavior Volume 46, Number 6, 2014 caregivers' frequency of use and interest in various technologies (ie, dependent variables) and collect demographic information (ie, independent variables). A panel of earlychildhood teachers and parent service coordinators reviewed the survey for readability and ease of completion. In addition, the survey was piloted in 2012 with over 500 respondents and then modified for the 2013 survey to improve clarity, readability, and formatting based on participant and data collector feedback. Pilot survey participants were recruited using the same criteria as the full sample and were similar in terms of race, geography, age, and income. The final version of the survey was at a 3.2 Flesch-Kincaid reading level. The extensive preliminary testing and review of the instrument with both parents and teachers attest to the face and content validity of the instrument. Furthermore, refinements made to the instrument in light of pilot feedback served to reduce participant confusion (ie, error variance) and thus increase the reliable variance in participant responses. Typical assessment of internal consistency (eg, Cronbach alpha) was not appropriate to assess reliability because the items were checklists or single item constructs. Participants were asked, ‘‘How often do you use each of these?’’ Responses were given for Internet, cell phone (for texting), smartphone (for Internet access), Facebook, Twitter, e-mail, and blogs. Participants indicated frequency of use on a 1 (‘‘never’’) to 6 (‘‘daily’’) scale. In addition, participants were asked to ‘‘Indicate all the ways (yes or no) that you would like to receive information on the following topics: (a) what's going on at your child's center (b) help with parenting, and (c) healthy eating.’’ The former 2 topics (eg, center activities and parenting help) served as a reference for interest in nutrition information. For each topic, participants indicated whether they were interested in using each technology (Facebook, text messaging, phone calls, twitter, e-mail, and blogs) for receiving information (‘‘yes’’) or if they were not (‘‘no’’). These questions reflect 3 changes that were made based on feedback from the pilot survey: (1) the response scale for frequency of technology use was expanded to 6 intervals instead
of 4 to reflect participant request for greater refinement; (2) the instructions and response options for indicating interest in methods of information delivery were modified to have parents indicate ‘‘yes’’ or ‘‘no’’ for each topic for each technology rather than asking them to ‘‘mark all that apply,’’ to address confusion in pilot testing; and (3) Internet and smartphone were not included in the list of methods of information delivery through which parents indicated interest in receiving new information. This final change was made because pilot participants consistently indicated confusion about the intent of these methods. Recurring comments from pilot participants were that they were unsure whether the Internet and smartphone options were intended to be self-directed (eg, a Google search) or externally driven (eg, a Web page or phone application housing information).
Statistical Analyses The researchers calculated descriptive statistics for individual items to document rates of use and interest for each technology type (frequencies). An overall interest variable was created for each technology mode by summing across the positive responses for each technology. For example, a ‘‘yes’’ response for interest in receiving Facebook updates on parenting and healthy eating, but not activities at the center, would correspond to a Facebook Interest score of 2. Thereafter, a MANOVA approach was selected to account for the intercorrelations of the dependent variables (ie, frequency of use and overall interest across technologies). Follow-up ANOVAs and Bonferroni post hoc comparisons were conducted when the omnibus effects were at statistically significant levels. Wilks' lambda values for the omnibus analysis were considered an indication of effect size. Because of the number of tests, a conservative P ¼ .01 was adopted for omnibus tests.
RESULTS Descriptives Most respondents were female (76.1%). About a third (34.2%) identified as white, 41.2% identified as African American, 19.5% identified
Journal of Nutrition Education and Behavior Volume 46, Number 6, 2014
Swindle et al 487
Table 1. Frequency of Use and Interest in Delivery of Information, by Technology Type (n ¼ 806)
Frequency: ‘‘How Often Do You Use Each of These?’’
Internet
A Few 1–3 1–2 3–6 Never Times/y Times/mo Times/wk Times/wk Daily 6.1 3.1 5.1 6.9 9 69.9
Interest: ‘‘Indicate All the Ways (Yes or No) That You Would Like to Receive Information on the Following Topics’’ What Is Going on at the Center –
Parenting Help –
Healthy Eating –
Text messaging
4.5
2.5
2.8
3.3
6.3
80.6
64.6
49.4
53.8
Cell phone (for calls)
1.3
1.3
0.9
2.2
3.5
90.9
58.5
47.7
43.1
Smartphone (for Internet)
21.1
1.7
2.2
2.2
4.3
68.4
–
–
–
13.7
3.5
6.9
7.6
11.3
57.0
52.3
47.1
53.2
72.2
4.9
4.4
3.2
2.8
12.5
6.6
4.8
6.4
15.4
4.1
8.1
8.5
10.9
53.0
68.8
62.2
64.5
Blogs
75.6
5.6
4.5
2.7
2.4
9.2
6.8
9.3
11.4
Note: Numbers presented are percentages. For Frequency, rows total 100%. For Interest, participants were encouraged to select as many or as few ‘‘yes’’ answers as they liked. Thus, interest columns or rows do not total to 100%. Rather, percentages represent the number of participants who endorsed the options or interest. Participants were not asked about interest in Internet and smartphone because of ambiguity about initiation. Additional rationale is offered in the text. as Hispanic, and 4.8% identifed as multiracial or ‘‘other.’’ Of the entire sample, 17.6% requested and completed the survey in Spanish. On average, participants were 33.14 years of age (SD, 10.08 years; range. 16–74 years). To examine for generational differences in use and interest in technology, age was grouped into 4 categories for inclusion in the analysis: 16–25 (n ¼ 150), 26–35 (n ¼ 285), 36–45 (n ¼ 100), and $46 years (n ¼ 76). Of those surveyed, 77.9% were parents of the child, 12.8% were grandparents of the child, 7.8 % were other relatives, and 1.5% were partners of a parent. Table 1 presents the proportions of participants responding in each frequency and interest category for each technology type. Most technologies (6 of 8) were used daily by $ 50% of the sample. The most popular daily technology uses were cell phones for calls (90.9%), cell phones for texting (80.6%), the Internet (69.9%), smartphones for Internet (68.4%), and Facebook (57%). More than half of participants used e-mail on a daily basis (53.0%). When weekly usage was considered ($1–2 times/wk), $ 72% of participants used most technologies (6 of 8) on a regular basis. However, most families surveyed did not access Twitter or blogs regularly.
Approximately 75% said they never used those options at all. Table 1 also presents the percentage of respondents indicating each technology type as a preferred information delivery mode on the topics presented. E-mail was the technology of most interest overall, with over 60% indicating interest across topics. Text messaging and Facebook were also of interest for $ 47% of participants across topics, similar to rates of indicated interest for phone calls. Interest in Facebook and text messaging for receiving information on healthy eating were within 11% of the most preferred technology for that topic, e-mail. As seen in Table 1, interest in receiving information on healthy eating varied by technology type, ranging from being of lowest comparative interest via texts and phone calls and of highest comparative interest via Facebook.
Demographic Differences The omnibus 2-way interaction of ethnicity and age was nonsignificant. However, main effects for age (Wilks' l ¼ .87; F ¼ 1.98; P < .001) and ethnicity (Wilks' l ¼ .79; F ¼ 3.39; P < .001) were observed. Table 2 presents mean usage rates by ethnicity and age with significant group differences denoted. The significant
omnibus main effect of age was explored to reveal a significant univariate effect on 5 dependent variables: use of Internet (F ¼ 8.82; P < .001), Facebook (F ¼ 9.26; P < .001), e-mail (F ¼ 4.73; P ¼ .003), texting (F ¼ 4.98; P ¼ .002), and smartphone for Internet use (F ¼ 10.66; P < .001). Where group differences were observed, the oldest age ($45 years) used technology least frequently. However, this group was not significantly less likely to indicate interest in any type of technology for receiving information. Because of the significance of the overall test for ethnicity, the authors examined the univariate main effects. Significant univariate main effects for ethnicity were obtained for use of Internet (F ¼ 11.18; P < .001), Facebook (F ¼ 6.07; P < .001), Twitter (F ¼ 6.20; P < .001), e-mail (F ¼ 11.99; P < .001) and interest in e-mail (F ¼ 3.57; P ¼ .014), and phone call (F ¼ 6.39; P < .001). Hispanics used Internet, Facebook, Twitter, and e-mail significantly less than at least 1 other group and were significantly less interested in the use of e-mail than all other groups (Table 2). African Americans used Twitter significantly more than all other ethnic groups and were more interested in receiving phone calls than were whites.
Journal of Nutrition Education and Behavior Volume 46, Number 6, 2014
488 Swindle et al
Table 2. Mean Frequency of Parent Technology Use and Preference, by Age and Ethnicity 17–25 y (n ¼ 150)
26–35 y (n ¼ 285)
36–45 y (n ¼ 100)
45–55 y (n ¼ 76)
Total (n ¼ 611)
Age
Mean
SD
Mean
SD
Mean
SD
Mean
SD
Mean
SD
P
Internet use
5.5a
1.1
5.4a
1.3
5.5a
1.2
4.6b
2.0
5.3
1.4
< .001
Facebook use
4.9a–c
1.6
5.1ac
1.6
4.5bc
2.0
3.7d
2.2
4.8
1.8
.02
Twitter use
2.2
1.9
2.00
1.8
1.9
1.7
1.5
1.4
2.0
1.8
NS
E-mail use
4.4a–c
1.9
4.9ab
1.7
4.9ab
1.8
4.0c
2.1
4.7
1.8
.002
Blog use
1.6
1.4
1.8
1.6
2.1
1.9
1.3
1.0
1.7
1.6
NS
Cell phone use for texting
5.7a
0.9
5.6a
1.2
5.7a
1.0
4.6b
1.9
5.5
1.3
< .001
Cell phone use for calls
5.9
0.6
5.8
0.8
5.9
0.5
5.6
1.1
5.8
0.8
NS
Smartphone use for Internet
5.2ab
1.7
5.1a–c
1.9
4.5bc
2.3
3.2d
2.4
4.8
2.1
.03
Facebook interest
1.0
1.3
1.2
1.3
0.8
1.2
0.8
1.2
1.0
1.3
NS
Twitter interest
0.1
0.6
0.1
0.5
0.8
0.4
0.9
0.1
0.1
0.5
NS
E-mail interest
1.4
1.4
1.4
1.4
1.4
1.4
0.9
1.3
1.3
1.4
NS
Blog interest
0.2
0.6
0.2
0.6
0.2
0.7
0.1
0.4
0.2
0.6
NS
Texting interest
1.2
1.3
1.2
1.3
1.1
1.3
0.8
1.1
1.1
1.3
NS
Phone call interest
1.0
1.2
0.9
1.2
0.9
1.2
1.0
1.2
0.9
1.2
NS
White (n ¼ 209) Ethnicity Internet use
Mean
SD
African American (n ¼ 254) Mean SD
5.5a
1.3
5.5a
Facebook use
a
5.1
1.7
4.7
Twitter use
1.6a
1.4
2.5b a
a
ab
Hispanic (n ¼ 119)
Multiracial/ Other (n ¼ 29)
Total (n ¼ 611)
Mean
SD
Mean
SD
Mean
SD
1.3
4.7b
1.7
5.5a
1.1
5.3
1.4
.02
1.8
b
4.3
2.0
5.1
1.5
4.8
1.8
.001
2.1
1.6a
1.4
1.6a
1.5
2.0
1.8
.03
b
a
ab
E-mail use
5.0
1.7
4.9
1.7
3.6
2.1
4.7
2.0
4.7
1.8
.02
Blog use
1.8
1.6
1.8
1.6
1.6
1.5
1.7
1.4
1.7
1.6
NS
Cell phone use for texting
5.6ab
1.3
5.7a
1.1
5.2b
1.5
5.3ab
1.5
5.5
1.3
.005
Cell phone use for calls
5.9
0.5
5.8
0.8
5.8
0.8
5.6
1.4
5.9
0.8
NS
Smartphone use for Internet
4.6a
2.2
5.2b
1.8
4.5a
2.2
4.6ab
2.2
4.8
2.1
Facebook interest
1.1
1.3
1.0
1.2
1.0
1.3
1.2
1.3
1.0
1.3
NS.
Twitter interest
0.1
0.4
0.1
0.6
0.1
0.4
0.03
0.2
0.1
0.5
NS
E-mail interest
1.4a
1.4
1.5a
1.4
0.9b
1.3
1.4a
1.4
1.3
1.4
.01
.008
Blog interest
0.2
0.7
0.1
0.6
0.1
0.5
0.1
0.5
0.2
0.6
NS
Texting interest
1.0
1.2
1.3
1.3
1.0
1.2
0.8
1.0
1.1
1.3
NS
Phone call interest
0.7a
1.1
1.1b
1.3
0.9ab
1.2
0.8ab
1.0
0.9
1.2
< .001
NS indicates comparisons that had no significantly different group means. Note: Where significant group differences were found based on Bonferroni post hoc mean comparisons (adjusted P < .05), homogenous groups are indicated by superscripts. The largest significant adjusted P for the set of comparisons is given.
DISCUSSION Rates of technology use by this lowincome sample exceeded previous study estimates12,13,15,16,18,23 and rates observed in the 2012 pilot survey.24 Results suggest that lowincome families find means to access the Internet and that cell phone use by low-income families is similar to
rates of cell phone use by those of higher socioeconomic status.25 Additionally, this study documented how often low-income families use technology and the types of social media outlets they do (and do not) frequent. Text messaging, Facebook, and e-mail reached over 70% of this sample on a weekly basis. Rates of use for these technologies among this low-income
sample are consistent with or higher than rates found in broader samples of the US public.13,14,16,23 Furthermore, the technologies least likely to be used by low-income parents and caregivers in this sample seem to match those least used by the general public. Twitter and blogs, specifically, were used by < 20% of participants regularly, which is similar to reported
Journal of Nutrition Education and Behavior Volume 46, Number 6, 2014 rates of general public use from the Pew Research center.15,16 Previous research on demographic differences in technology use suggested that white adults used the internet19 and Facebook26 more than other ethnic groups; African Americans used Twitter15 and mobile Internet12 more than other ethnic groups; young adults used Facebook14 and Twitter15 more than older age groups; and older adults used cell phones less.27 Consistent with previous research, African Americans in this sample were the most frequent users of Twitter and used mobile Internet (ie, smartphones) more than other ethnic groups. In addition, the oldest age group (> 45 years) used Facebook, cell phones for texting, and mobile Internet significantly less than younger adults. Contrary to previous findings, whites used Internet at rates similar to African American and multiracial participants and did not differ from African Americans in Facebook use. Furthermore, there were no age group differences in use of Twitter and cell phones for calls. In addition to providing information on rates and types of technology used by demographic subgroup, this study assessed parents' and caregivers' interest in receiving information via technology. According to the Health Belief Model,5 the individual's value of a health goal is an important predictor of health behavior. Thus, interest in receiving nutrition information is a marker for potential to improve behaviors. In the present study, interest in receiving parenting, child care, or nutrition information was > 50% for at least 1 topic for 4 technology types. Furthermore, only 2 differences in interest were observed across demographic subgroups despite the observed differences in use. Less use of a technology corresponded to less interest for only 1 ethnic group (Hispanics) on 1 mode of delivery (e-mail). In this sample, groups using technologies less frequently were still interested in using these avenues to receive information. Furthermore, the similarity in rates of interest between the topics of healthy eating, parenting help, and activities at the center suggests that nutrition is an area in which many parents were interested in receiving information.
This study recruited to approximate a racial balance similar to low-income families in the state, but the sampling procedure was not adequate to ensure a representative sample. Replications in other states are needed to provide additional support for these findings and to extend understanding of technology use to additional ethnic groups (eg, Native Americans). In addition, the self-selective nature of this survey may have created a selection bias that affected findings. Although the authors were unable to collect data to estimate how parents and caregivers who refused to participate or were absent from the center were different from participating parents, these differences could impact demographic patterns.
IMPLICATIONS FOR RESEARCH AND PRACTICE Data from the present study suggest that technology may be an important tool in closing the gap between those served by nutrition education programs and those interested in service who are not reached. In a survey of 68 Head Start mothers, the greatest reported barriers to involvement were schedules that conflicted with Head Start activities and having another child at home who required care.28 Through technology, parents and caregivers could engage in an online community on their own schedules. Technology could be used to reach parents with new material (eg, recipes), increase program participation (eg, virtual discussions), and obtain needed information (eg, updates on progress or barriers to goals). Technology-based programs also promote an environmentally protective practice (ie, ‘‘green’’) and require less expenditure on materials for the intervention. With successful technologybased campaigns such as text4baby29 emerging, the current study's findings can inform the selection of information delivery modalities that fit the targeted population's use of and interest in technology types. Findings on ethnic and age differences, for example, can inform the development of interventions for specific populations. Although technology-based programs seem to be a
Swindle et al 489 viable option for many low-income families based on access and use, the current study suggests that there are barriers in certain subsamples. For example, older caregivers in this sample reported using several technologies less frequently. Technology efforts to reach this population may require a combination of participant training and supplementation with other approaches (eg, phone calls). Similarly, programs working with Hispanic families should consider using text messaging rather than Internet-based efforts. Further research is needed to better understand the use of technology to reach low-income parents and caregivers. Respondents' reported level of interest in technologies for receiving information may not translate to interest in participation in full-blown interventions. Efficacy trials with participation estimates are needed to determine whether interest levels in this study are consistent with parent and caregiver involvement in implemented technology-based programs. Most important, the successes with and barriers to using technology to reach low-income parents and caregivers must be documented, as well as the impact of these technologybased interventions on targeted behaviors. The current study attests to the potential feasibility, but not utility, of such programs. It is possible that effectiveness of technology interventions would be greater with low-income families compared with programs delivered through in-person contacts because of the many barriers families face to recruitment and retention in onsite programs. The mobile nature of a technology intervention may be more flexible to the frequent physical moves of low-income families.30 In addition, research is needed to determine whether interest in technology as an information pathway increases with exposure and experience. Surveying low-income parents and caregivers about their preference for technology information delivery before and after a technology-based intervention could provide evidence on this issue. In addition, there are other ways to use technology for intervention (eg, Web sites, smartphone applications) that were not assessed in the current survey. Future studies could provide additional detail on the sites that
490 Swindle et al low-income parents and caregivers visit and how smartphones are used. More study is also warranted to determine the optimal ways to reach families who are the hardest to access (ie, Hispanic and older caregivers).
ACKNOWLEDGMENTS This project was supported by Agriculture and Food Research Initiative Competitive Grant 2011-6800130014 from the US Department of Agriculture, National Institute of Food and Agriculture.
REFERENCES 1. Golley RK, Smithers LG, Mittinty MN, Emmitt P, Northstone K, Lynch JW. Diet quality of UK infants is associated with dietary, adiposity, cardiovascular, and cognitive outcomes measured at 78 years. J Nutr. 2013;143:1611-1617. 2. Slopen N, Fitzmaurice G, Williams DR, Gilman SE. Poverty, food insecurity, and childhood internalizing and externalizing behaviors. J Am Acad Child Adolesc Psychiatry. 2010;49:444-452. 3. Zeisel SH. Epigenetic mechanisms for nutrition determinants of later health outcomes. Am J Clin Nutr. 2009;89: 14885-14935. 4. Lorson BA, Melgar-Quinonez HR, Taylor CA. Correlates of fruit and vegetable intake in children. J Am Diet Assoc. 2009;109:474-478. 5. Becker MH, Maiman LA. Sociobehavioral determinants of compliance with health and medical care recommendations. Med Care. 1975;13:10-24. 6. Baker CN, Arnold D, Meagher S. Enrollment and attendance in a parent training prevention program for conduct problems. Prev Sci. 2011;12:126-138. 7. Winslow EB, Bonds D, Wolchik S, Sandler I, Braver S. Predictors of enrollment and retention in a preventive parenting intervention for divorced families. J Prim Prev. 2009;30:151-172. 8. Heinrichs N, Bertram H, Kuschel A, Hahlweg K. Parent recruitment and retention in a universal prevention program for child behavior and emotional problems: barriers to research and program participation. Prev Sci. 2005;6:275-286.
Journal of Nutrition Education and Behavior Volume 46, Number 6, 2014 9. Huebner CE. Evaluation of a clinicbased parent education program to reduce the risk of infant and toddler maltreatment. Public Health Nurs. 2002;19:377-389. 10. Rainie L. Internet, broadband, and cell phone statistics. http://www.pewinternet .org/Reports/2010/Internet-broadbandand-cell-phone-statistics.aspx. Accessed July 10, 2014. 11. Becker S. First-ever national study: millions of people rely on library computers for employment, health, and education. Lib Times Int. 2010;26:47. 12. Horrigan J. Wireless Internet use. http:// www.pewinternet.org/Reports/2009/12Wireless-Internet-Use.aspx. Accessed July 10, 2014. 13. Lenhart A. Cell phones and American adults. http://www.pewinternet.org/ Reports/2010/Cell-Phones-and-AmericanAdults.aspx. Accessed July 10, 2014. 14. Lenhart A, Purcell K, Smith A, Zickuhr K. Social media and young adults. http://www.pewinternet.org/2010/02/ 03/social-media-and-young-adults/. Accessed July 10, 2014. 15. Smith A. Twitter update 2011. http:// www.pewinternet.org/Reports/2011/ Twitter-Update-2011.aspx. Accessed July 10, 2014. 16. Zickuhr K. Generations 2010. http:// www.pewinternet.org/2010/12/16/gener ations-2010/. Accessed July 10, 2014. 17. Ono H, Zavodny M. Digital inequality: a five country comparison using microdata. Soc Sci Res. 2007;36: 1135-1155. 18. Corda K, Palmer D, Adler A. New Jersey EFNEP and FSNE program participant Internet usage: a study to examine the efficacy of using the Internet to deliver nutrition education. J Nutr Educ Behav. 2008;40:S25. 19. Neuenschwander LM, Abbott A, Mobley AR. Assessment of lowincome adults’ access to technology: implications for nutrition education. J Nutr Educ Behav. 2012;44:60-65. 20. Hesse BW, Nelson DE, Kreps GL, Croyle RT, Arora NK, Rimer BK. Trust and sources of health information: the impact of the Internet and its implications for health care providers. Findings from the first Health Information National Trends survey. Arch Intern Med. 2005;165:2618.
21. Department of Commerce. Federal Register: Urban criteria for the 2010 census 2011. http://www.census.gov/ geo/www/ua/fedregv76n164.pdf. Accessed July 10, 2014. 22. US Census Bureau. Current Population Survey (CPS): a joint effort between the Bureau of Labor Statistics and the Census Bureau. http://www.census. gov/cps/about/cpsdef.html. Accessed July 10, 2014. 23. Lenhart A. Social & mobile Internet use among teens and young adults. http:// pewinternet.org/Reports/2010/SocialMedia-and-Young-Adults.aspx. Accessed July 10, 2014. 24. Swindle T, Whiteside-Mansell L, Ward W, Bokony PA, Pettit D. Technology use and preference by low-income parents of young children: demographic patterns and implications for intervention. Presented to the 2013 biennial meeting of the Society for Research in Child Development; Seattle, WA; April, 2013. 25. Lane W, Manner C. The impact of personality traits on smartphone ownership and use. Int J Business Soc Sci. 2011;17:22-28. 26. Hampton KN, Goulet LS, Purcell, K. Social networking sites and our lives: how people’s trust, personal relationships, and civic and political involvement are connected to their use of social networking sites and other technologies. http://www.pewinternet.org/ 2011/06/16/social-networking-sites-andour-lives/. Accessed July 10, 2014. 27. Lenhart A. Cell phones and American adults. http://www.pewinternet.org/ Reports/2010/Cell-Phones-and-AmericanAdults.aspx. Accessed July 10, 2014. 28. Lamb-Parker F, Piotrkowski CS, Baker AJL, Kessler-Sklar S, Clark B, Peay L. Understanding barriers to parent involvement in Head Start: a research–community partnership. Early Child Res Q. 2001;16:35-51. 29. Evans WD, Wallace JL, Snider J. Pilot evaluation of the text4baby mobile health program. BMC Public Health. 2010;12:1031. 30. Ma T, Gee L, Kushel MB. Associations between housing instability and food insecurity with health care access in low-income children. Ambulatory Pediatrics. 2008;8:50-57.
