Background

Children (defined as those under 14 years of age) make up roughly a quarter of the world’s population and, while they do not directly control family spending, they influence it. Children can influence not only things bought directly for them, but also family purchases such as restaurants to eat at, holidays to take, or cars to buy (Berey and Pollay 1968; Dotson and Hyatt 2005; John 1999; Lee and Beatty 2002). Research estimates that children influence up to a fifth of all household purchases (McNeal 1992). Given this, it is surprising how scant the research is in this area. In part, this is just reflective of the general lack of observational research about the fundamental nature of shopping trips.

Children can influence what parents buy in two ways. The first is passive, where parents buy things they know children want or like (Isler et al. 1987; John 1999). The second is where children ask for something directly, known as ‘pester power’ (Gunter and Furnham 1998). Research into ‘pester power’ has mostly relied on surveys as observational research is difficult and time-consuming. When parents are asked about children’s requests, they usually talk about negative persuasive or emotional strategies, such as nagging, whining, or simply sneaking what they want into the trolley (Gelperowic and Beharrell 1994; Hill and Tilley 2002; Rust 1993b; Wilson and Wood 2004). But, when parents are actually observed, only nineteen per cent of requests resulted in conflict (Holden 1983). The mainstream perception that conflict is common is perhaps attributable to parents being more likely to recall events that have some negative emotion or conflict attached, or wanting to tell researchers ‘war stories’.

Prior research has also found that, in interviews, parents claim that they take longer to do the shopping when they have children physically accompanying them (Pettersson et al. 2004; Wilson and Wood 2004) and that shopping with children makes both parenting and shopping more ‘difficult’ (Craig 2006; Holden 1983).

Surveys are limited when trying to understand low-involvement, habitual behaviour, such as shopping (Young and Hetherington 1996). When people are navigating supermarkets, which have tens of thousands of items, and getting in and out in only a few minutes (Sorensen et al. 2017), it is unlikely that people will be able to remember everything they’ve done, or be able to accurately describe or explain their behaviour (Sharp and Tustin 2003). When researchers use observational techniques, they find that children have more influence than parents give them credit for (Berey and Pollay 1968; Ebster et al. 2009; Lee and Beatty 2002). For instance, Ebster et al. (2006)found that parents underestimated their granting of requests by about half, compared with observations, where 52% of requests were granted, although they fail to report what the incidence of requests was. This means that if we really want to know what happens in stores, and with parents and children, direct observation is best (Bloch et al. 1994; Granbois 1968; Shankar et al. 2011), as it gives more reliable and realistic information (East and Uncles 2008; Lee and Collins 1999; Rust 1993a; Scamell-Katz 2012; Schwarz 1999).

With this in mind, this research examined the incidence of children in supermarket shopping parties and their effect on trip length. We captured the product requests children make and the parental response they receive, and the frequency with which disruptive behaviours are drawn upon by children to exert their influence. This was all done via observational research.

Method and data

We conducted this research in four supermarkets in two Australian capital cities, using entry and exit interviews for the capture of descriptive data on the incidence of children and timing of trips. Every fifth shopper was recruited on entry to the store and interviewed on their exit, with a 77% recruitment rate. We obtained basket size, spend, time in-store, shopping party size and composition, household size and composition, and shopper gender. A total of 1611 exit interviews were conducted. We used payment receipts to validate the total SKUs purchased.

Video footage was used to investigate ‘pester power’. Shoppers with children were asked to wear a recording device and were only informed of the true intent of the study afterwards, in order to not influence their behaviour during data collection. Permission to use the recording was obtained, post collection, with full understanding. No respondent denied the request. We recorded 1839 parent-child interactions, across 89 trips and 30 hours of shopping footage. We use the well-validated Eyberg Child Behaviour Inventory (ECBI) (Collett et al. 2003; Eyberg and Robinson 1983) as a way to categorise the potentially disruptive behaviours seen. The ECBI is used in a clinical setting to understand disruptive behaviours and is completed by caregivers and educators. The list of behaviours is shown in Table 1.

Table 1. Eyberg Child Behaviour Inventory

Results

How often children are present and on which types of trip

We found that the proportion of shoppers accompanied by children in the supermarket is a function of the demographic makeup of the surrounding area. Table 2 shows that the incidence of shoppers with children in-store ranged from 10% to 26%, with an average of 17%. The incidence is captured as a function of the proportion of households in the catchment area with children living in them. On average, the percentage of shoppers in-store with children will be half that of the proportion of households in the catchment area with children. Catchment areas with a higher proportion of households with children have a (slightly) higher likelihood of children being present on a shopping trip: there is a slight Double Jeopardy effect.

Table 2. Proportion of shoppers in-store with children, of children in the surrounding catchment area

The size of the shopping trip, in terms of SKUs purchased, is not related to children being part of the shopping party or not. Children are found to be present across all types of trip sizes and durations. So, adults do not try and exclude their children from bigger and longer shops (see Figure 1). But interestingly, this last conclusion differs between male and female shoppers.

Figure 1: Distribution of SKUs purchased by the presence of children

Women still do the bulk of caring for children: over two-and-a-half times as much as men (Australian Bureau of Statistics 2016b), and the bulk of the supermarket shopping. In this study, seven in 10 shopping groups observed were exclusively female. Not only are men seen shopping less often, but they also have children with them on fewer occasions. In a household with children, women are 1.6 times more likely to take a child shopping with them (than a man in such a household). While women are equally likely to take children with them to the supermarket regardless of how many items they purchase, men are less likely to take children with them on quick trips for fewer items.

Shoppers who have children with them buy the same number of items as those without, but they buy them at a faster rate. This contradicts the findings from the claimed behaviour of parents in prior studies, highlighting the limitations of claimed behaviour research in this area. The size of this effect increases as the basket size increases (see Figure 2). This faster shopping speed also holds for people who have children in the household, but not with them on that trip, but the effect is seen to a lesser degree. Parents are under time pressure, especially when they have their children physically present.

Figure 2: Number of SKUs purchased cf average trip duration 

Requests and ‘Pester Power’

There is a big spread in terms of what categories requests from children are seen in and also what categories adults make suggestions for. This is shown in Table 3.

One in five shopping trips saw no requests being made by children. Of the trips where children made a request, they made four requests on average. Expectedly, this varied by the trip duration. For trips under ten minutes, children made an average of 1.7 requests; 10 to 20 minutes they made 3.5 requests; 20 to 30 minutes, five requests, and trips over 30 minutes, six requests. While the overall number of requests increased, the rate of requests per minute declines gradually as the duration of the trip increases.

Requests are made across the whole range of supermarket categories. The highest single category for the children-initiated request was confectionary at 14%. The category where adults made the most suggestions to children was fruit and vegetables at 16% of observed interactions. Non-food requests from children (12%) varied greatly from hand wash to shoe inserts. For some categories, such as dairy, juice and fizzy drinks, we see requests coming from both parties at a similar incidence rate.

Table 3. Child requests and parent suggestions by category.

Salty snacks and dairy were the two categories with the highest granting of a child’s request at 67% and 52% granting respectively.

Overall, parents refused most requests (72%). A fifth (22%) of requests were granted, and in 6% of cases, parents and children negotiated. Four in 10 parents did not grant any requests at all. Granting of children’s requests is very variable across categories. Parents were most likely to give in to requests for salty snacks and dairy, and fruit and vegetables. It probably comes as no surprise that juice, fizzy drink, non-food items, cereal, and toys were the least granted requests.

Half of all trips saw no behaviour that would be rated on the ECBI as ‘disruptive’. Of the ECBI behaviours observed, whining was by far the most common behaviour (See Table 4). More extreme behaviours of fighting, tantrums and arguing were rare, suggesting that children are generally well behaved when accompanying an adult supermarket shopping. Only two tantrums were recorded, but this may have been a selection effect: parents may be both more likely to leave children prone to tantrums at home, and less likely to agree to take part in research where such behaviour would be captured.

Table 4. Observed behaviours from Eyberg CBI.

We saw a higher rate of requests from children on trips where a behaviour on the ECBI occurs (one request every 4.5 min) compared with trips where no such behaviour occurs (one request every 7.7 min). Potentially disruptive behaviours appear to be fairly evenly distributed through the duration of trips, though slightly more in the final quarter. Twenty-one percent of disruptive events occurred in the first quarter of a trip, 22% in the second quarter, 24% in the third quarter, and 32% in the final quarter.

Implications

On average, 17% of grocery store shoppers are accompanied by a child. This number will vary, reflective of the surrounding catchment area. Overall, while children were only present on 60% of the trips made by households that could take children with them, shoppers do not appear to favour taking their children according to the size of the shopping trip. Stores expecting customers to bring children need to be designed to accommodate them on all trip types, not just larger ones. Things that might help parents pick one store over another, especially on shorter shopping trips that may not be commonly associated with children accompanying parents, are things like shallow shopping carts with a child seat, or enough room to have a child at the self-check, or a small piece of fresh fruit at the entry to keep a child occupied. It is also critically important to be sure that there is nothing that will make the trip more difficult for parents, as negative experiences are more often recalled (and potentially avoided) than an ordinary, fast and easy shopping trip.

Prior research had suggested that shoppers with children would spend longer in the store, because they need to buy more items, and have to manage a child (Sommer et al. 1992; Thomas and Garland 1993). However, we find that when you take into account all trip types, and control for household size, the opposite is true: shoppers with children shop approximately 15% faster than those without children. The implications of this for in-store marketing efforts are clear. Accompanied shoppers are going to be less receptive to complicated messaging, or offers that require reading. Using extremely clear branding, including the use of Distinctive Assets, visible to the most distracted shopper, is critically important where shoppers with children are concerned. Out-of-store marketing efforts are also important to build mental availability for fast recognition in-store.

‘Pester power’ is prevalent across trips (80% of trips see it used) and supermarket categories. However, it is unsuccessful in the vast majority of instances (80% parental refusal on average). Children’s ‘passive dictation’ of product purchasing (Wells 1965), where parents buy products because they know children will eat or like them, appears to be the more effective route. Parents should be the target for building memory structures, including Distinctive Assets, and children the target for usage and product development. Rather than seeing children influencing their parents as a battle to be won, it would be better to think of it as a collaboration – parents and children working together to get the shopping done – as quickly as possible.