The Influence of Children on Family Purchasing in Lithuania and Azerbaijan

In the context of marketing, children are considered to be increasingly important influencers of parents’ purchasing decisions. However, their influence varies depending on the products and cultures, and many particularities of this phenomenon remain under-researched. This is especially true in regard to the countries that are categorized as “emerging economies�?. Some of them still do not have reliable measurements of their cultural dimensions, and this increases difficulties in performing comparative analysis there. On the other hand, these countries offer a broad and important scene for child influence studies. This study concentrates on analysis and comparisons of children’s influence on parental purchase decision-making in Lithuania and Azerbaijan. These countries are different in, at least, two Hofstede’s cultural dimensions that are important in family decision-making: individualism and uncertainty avoidance. Also, the study contributes to the existing research by using a product use-related categorization of product groups. Such an approach reveals significant difference in terms of how children exert their influence on purchasing services versus tangible products for the family use. Interesting differences between the countries in terms of children’s influence on purchasing products for their personal use also opens a new scene for future studies that might consider a similar product categorization approach.


Introduction
Decision making and consumption patterns of one of the most important units of markets, a family, have attracted the attention of marketers and researchers over the years (Granbois, 1979;Filiatrault & Ritchie, 1980;Lackman & Lanasa, 1993;Shoham & Dalakas, 2006;Su & Wang, 2010). While the early studies mainly focused on the adult members of the families, the increasing role of children as direct consumers as well as influencers led to closer investigation of their direct and indirect influence on the family purchase decision-making process. A number of studies on consumer socialization processes of younger generations revealed that children not only represent a rapidly growing segment with a high spending power of its own, but also a powerful segment with substantial influencing power (Mehrotra & Torges, 1977;Acuff, 1997). According to public surveys, 71% of parents in the US say they solicit opinions from their kids regarding purchases (White, 2013).
Children's influence is well acknowledged in a wide range of disciplines including child and family psychology (Gianinno & Crittenden, 2005), communication (Moschis & Moore, 1979, Belch et al., 1985Chan & McNeal, 2006), and consumer behaviour (Robertson et al., 1985;Haynes et al., 1993;Shoham & Dalakas, 2006;Su & Wang, 2010), which analyze children's influence on purchasing decisions within a family. While the influence is observed in various cultures and societies (Holdert & Antonides, 1997;Wimalasiri, 2000;Rose, 2002;Chan & McNeal, 2003), the extent of the influence is dependent on a number of factors including but not limited to product type, family communication, stage of the purchase decision-making process as well as age, gender, and birth-order of a child or children (Caruana & Vassallo, 2003;Flurry, 2007;Shergill et al., 2013;etc.).
In this study, we aim to add to the current body of research on children consumer socialization in emerging economies and to investigate the influence of children on purchasing decision-making within a family in two culturally different environments (Azerbaijan and Lithuania), which share certain similarities and have significant differences, and have not been compared on this aspect yet. We contribute to the existing research by suggesting a different categorization of product groups that are closer related with product use than just specifying their technical similarity.

Theoretical Background and Hypotheses Development
Several theories ranging from consumer socialization to children cognitive development, family communication and decision-making guide the current research of children's influence on parental purchase decision-making. For clarity purposes, in this study we accept Sprey's definition of influence, i. e.: influence is a relative and subjective concept that reflects a perceived rather than actual situation and "may be achieved as a result of exchange process between two or more parties" (Sprey in Flurry & Burns 2005, p. 595).
Consumer Socialization. Children consumer socialization is defined as a long-term process by which young people acquire skills, knowledge, and attitudes relevant to their functioning as consumers, and it takes into account the content, the processes, and the changes in children's learning about the marketplace (Ward, 1974). One of the earliest models of consumer socialization developed by Moschis and Moore (1979) identifies five influencing variables: socialization agents, social structural variables, age/life cycle, learning processes, and learning properties. Transposing the model into children socialization topic assumes that a child's stage of development (Acuff, 1997;Chan, 2005) and social environment including family type and family communication (Moschis, 1985;Darley & Lim, 1986;Caruana & Vassallo, 2003;Bakir et al. 2006), and cultural and social differences (Ward et al., 1986;Rose, 2002) are among the most influential factors to affect child's consumer behaviour including influencing. The model implies that the relationship between the learner (i.e., a child) and the agent (e.g., a parent) is continuous and reciprocal suggesting that these roles are interchangeable (Ekström et al., 1987;Roedder-John, 1999;Rose, 2002;Ekström, 2007) and are affected by both social and cognitive learning.
Cognitive and Social Development. Piaget's theory of intellectual development (Piaget, 1970) and Vygotsky's theory of social development (Selman, 1980) are among the most frequently used theories of cognitive and social development that employ age as a proxy variable for a child's development and use it to describe the age-related differences of children as growing human beings and consumers. In his theory of intellectual development Piaget (1970) identified four major stages of the child's cognitive development: 1) Sensorimotor stage (up to 2 years of age) when learning is through motor and reflex actions; 2) Preoperational stage (2-7 years of age) is identified as the beginning of symbolic thinking; 3) Concrete operational stage (7-11 years of age) is when abstract thinking and rational judgement are developed, and finally, 4) Formal operations stage (11-14 years of age) is described as a stage when abilities of more complex thoughtful analysis and hypothetical deductive reasoning evolve.
The theory of social development (Selman, 1980) compliments Piaget's theory of cognitive (intellectual) development by focusing on the social activities of a child. The theory suggests that every function of a child's development occurs on two levels: social and individual. According to Selman (1980), a child goes through five stages of social development: 1) Egocentric stage (age 3-6), when a child has only one perspective -his/her own; 2) Social information role-taking stage (age 6-8), when a child understands that people have different opinions without still having developed the ability to think from another person's perspective; 3) Self-reflecting stage (age 8-10), when a child is aware of a different perspective of other people and has already developed skills to consider these perspectives without being able, though, to consider other's opinions at the same time as his/ her own; 4) Mutual role-taking stage (age 10-12), at which a child has skills and abilities for simultaneous consideration of multiple points of view, and, finally, 5) Social system role-taking stage (age 12 and older), when a child is able to relate different perspectives to social groups and/or systems he/she belongs to.
The applications of cognitive and social development theories can be found in the majority of studies on children consumer socialization Caruana & Vassallo, 2003;Chan, 2005;Ekström, 2007). These theories are widely used as a methodological tool for sample description at the early stages of research of children's consumer socialization and employ age and gender as proxy variables for a child's development with an argument that these demographic factors are the main causes of the type and extent of children's influence on parental purchase decision-making. For example, in her research of children's influence on parental consumer learning Ekström (2007) explicitly stated that age was a major factor for the children's sample selection. She stated that only those older than thirteen years of age were selected based on the assumption that at this age children have already developed some independent consumer behaviour and frequently act as active agents of retrospective socialization. Past studies reported mixed results regarding children gender in regard to the nature and importance of influencing parental purchase decision-making (Kaur & Singh, 2006;Shoham & Dalakas, 2006;Beneke et al, 2011;Dikcius & Medeksiene, 2008;Martensen & Grønholdt, 2008). For example, girls were found more interested in clothing and more influential in the choice of style and colors, whereas boys exhibited more interest in clothes with licensed characters (Haynes et al., 1993). Contrary to Haynes et al. (1993), Carlson et al. (1990) and Flurry (2007) found boys to exert more influence than girls on toy purchasing. Meanwhile, no significant effects of the child's gender were found by Martensen and Grønholdt (2008).
Family Decision-Making and Communication. The patterns of intra-family communication, the style and effects of parent-child interaction in consumer learning (Moschis, 1985;Carlson & Grossbart, 1988;Caruana & Vassallo, 2003), the gate-keeping role of parents (Atkin 1978;Haynes et al., 1993) together with other social and individual factors are important for the development of a child as an individual and a member of a society, including the development of his/her consumer behaviour. The basic premise of the theory of decision-making is that given a set of possible alternatives, a "rational" person will attempt to make a decision that would lead to the best available outcome. Based on earlier research, Moschis and Moore (1979) stated that one of the important parental goals is to transfer the basics of "rational" consumption behaviour to their children. The theory assumes that in a decision-making process the decisionmaker: 1) defines and recognizes the problem, 2) searches for the information, 3) evaluates the alternatives, 4) makes a final decision, and 5) performs an action.
Studies of family decision-making employ a simplified three-or four-step decisionmaking process (Szybillo & Sosanie, 1977;Moschis, 1985;Filiatrault & Ritchie, 1980;Holdert & Antonides, 1997). However, the nature of family decision-making process is different from individual decision-making (Filiatrault & Ritchie, 1980;Lackman & Lanasa, 1993). A family is a group or a decision-making unit, where the decisions are not individual but joint, and, depending on the situation, members of this group perform various roles, such as a gate-keeper/initiator, influencer, decision-maker, buyer, and direct consumer (Filiatrault & Ritchie, 1980), which vary depending on family communication style, decision-making stage, product category, or life-cycle pattern (Lackman & Lanasa, 1993) as well as cultural specifics (Moschis, 1985;Caruana & Vassallo, 2003;Bakir et al., 2006 ). With this complex nature of family decision-making, Davis (1976) highlights three main issues: 1) group members may not agree on goals, 2) families as groups differ from other types of groups, and 3) there is a wide range of strategies in family decision-making that are applied according to the particular situation. In the research of children's influence, these issues are reflected in the strategies children develop to make requests and in the ways parents react to children's requests.
The early studies on family decision-making focused largely on the influence of the husband-wife dyad overlooking the role of children in family decision-making process (Davis, 1976;Granbois, 1979). Meanwhile, the increasing research on children's consumer socialization revealed that children exert considerable influence on family decision-making processes by participating in different stages of decision-making process (Szybillo & Sosanie, 1977;Belch et al., 1985;Moschis & Mitchell, 1986;Foxman et al., 1989). The effects of children's influence on family decision-making are explored from the perspectives of the family role structure, which is affected by such factors as cultural and/or societal changes, child-rearing styles, communication methods, etc. (Lalwani & Mehta, 2000;Rose, 2002;Shoham & Dalakas, 2006).
Resource Theory of Exchange. The resource theory of social exchange emphasizes the ability of the social interaction participants to satisfy their physical and/or psychological needs (Blood & Wolfe in Flurry, 2007). The main types of resources involved in exchange are emotions (e.g., parental love or affection), money (e.g., monetary contribution to the purchase), service (e.g., housework), information (e.g., better knowledge on the subject), status (e.g., birth order), and goods (e.g., exchangeable items) (Foa in Flurry 2007). Applied to the children's influencing ability, the theory suggests that with the increase of resources children have more chances to participate in and influence family purchase decision-making. This assumption is confirmed by the results of early studies on children's influence which show that income contributions, parental affection, or more sophisticated communication skills comprise a short list of resources used by children to influence parental purchasing behaviour (Berey & Pollay, 1968;Churchill & Moschis, 1979;Flurry, 2007). To influence the decisions of parents, a child may use his/her expertise on certain product categories, reward parents by good behaviour or punish by misbehaviour, exercise his/her perceived right of selection of the product, or express a strong desire to be associated with a reference group.
Culture and country. These findings are rather consistent among various cultural contexts (Wimalasiri, 2000;Beneke et al., 2011;Jeevananda & Kumar, 2012;Kaur & Singh, 2006;Ramzy et al., 2012;Su & Wang, 2010;Su, 2011). However, the early crossnational study of children influence and parental yielding by Ward et al. (1987) found that the extent and nature of parental yielding is affected by culture. These findings are also supported by a more recent study by Shoham and Dalakas, 2006, who found that the strength of the children influence varies significantly from country to country. When considering cultural influence, researchers often relate Hofstede's cultural dimensions (2001) and the interpersonal influences, largely dependent on cultural difference in child upbringing or in family purchase situations (Shoham & Dalakas, 2006;Su & Wang, 2010;Su, 2011;Ramzy et al., 2012;Shergill et al., 2013).
The multidimensional cultural influences typically result in a few main influences within the family purchasing process. First of all, more feminine (according to Hofstede's dimensions) cultures are expected to have stronger family ties, and therefore a child might have better chances to actively participate in purchasing processes. Another important cultural dimension is individualism, which makes people accept the influence of others less. Considering purchasing processes in a family, this suggests lower levels of a child participation and influence. One more cultural dimension that deserves special attention in a family purchasing is uncertainty avoidance. Since children often express spontaneous needs, emotions and behaviours, these would be more accepted in societies that tolerate less careful planning in order to avoid uncertainties. This implies that children from the culture that is less individualistic and has lower uncertainty avoidance index more actively participate in family purchasing processes. As a result, dimensions of individualism and uncertainty avoidance are the main in comparison with children influence in family purchasing; when it is needed for a study -power distance dimension is also included (Shergill et al., 2013).
The literature review indicates that there is rather comprehensive knowledge accumulated with regard to a number of key aspects of a child influence in family purchasing decision-making. However, a comparison of the phenomenon in two cultures that differ in terms of individualism and uncertainty avoidance within the context of emerging economies might bring additional insights. Additional value of this research can be achieved with more specific categorization of products and services that would better correspond to their types from the standpoint of consumers.
With this in mind, the following hypotheses were developed.
Hypothesis H 1 aims to verify whether children have stronger influence on purchases of products for their own use than products for the whole family in two emerging economies: H 1 : Children's perceived influence on parents' purchase decision-making is stronger for child-consumed than for family-consumed products.
Former studies (Kaur & Singh, 2006;Shergill et al., 2013) just partially or indirectly compared cases of purchasing services versus tangible products for a family. We believe this is rather important, since many services (e.g., recreational trips) have to be chosen with a very strong consideration of a child's specific needs (Darley & Lim, 1986). Many family-oriented tangible products, in contrary, may be used by all family members in a similar way. This yields the following hypothesis: H 2 : Children's perceived influence on parents' purchase decision-making is stronger for family-consumed services than family-consumed tangible products.
Children socialization has many aspects, including those that are inseparable from the public use of products. Being publicly visible, these products might be of special importance to children for their social lives. Since the study is performed in the countries that belong to the emerging economies, here these products may be even more important due to their signaling about personal social status (Ahmed & d' Astous, 2004). This is formalized in the hypothesis: H 3 : Children's perceived influence on parents' purchase decision-making is stronger for public products for a child than for any other child-consumed products.
The current study is performed in two countries that represent rather different cultural contexts for family decision-making. The differences of cultures are especially expressed in different scores of Hofstede's cultural dimensions of individualismcollectivism and uncertainty avoidance. We develop a hypothesis for testing whether perceived child influence is different in the two cultures that are different on these two dimensions. Following this idea, we would like to examine differences, if any, at the various levels of product groupings: H 4 : Children's perceived influence on parents' purchase decision-making is stronger in a more collectivistic culture with lower uncertainty avoidance than in a culture that is more individualistic and stronger oriented towards uncertainty avoidance.

Research methodology
The survey was performed in two countries -Lithuania and Azerbaijan. The choice of the countries is based on their similarities and differences that well correspond to the objectives of this paper.
The two countries belong to the group of emerging economies; both Lithuania and Azerbaijan share the historical period of being part of the former Soviet Union, which has influenced personal and social identities (Tereskinas, 2009). On the other hand, in terms of Hofstede's dimensions, Lithuania is described as above average individualism (60) and relatively higher uncertainty avoidance (65) (Borker, 2012), while Azerbaijan is characterised by low individualism (26) and low uncertainty avoidance (23) (Hatcher, 2008) Generally, analysis of children's influence on the purchase decision-making can be based on the standpoint of a child (Singh & Aggarwal, 2012), or aim to disclose a child's influence from the standpoint of other family members, typically -mothers (Ülger & Ülger, 2012). However, the latter approach seems to be prevailing, in part due to the complexity of collecting data directly from children (Morrow & Richards, 1996). In this survey, we also follow this route.
The respondents were selected via judgmental sampling and included mothers who had a child (or children) within 4 to 18 years of age. If a respondent had several children of the age within this range, she was asked to answer only about one child of her choice. Therefore we measured mothers' general attitudes towards the strength of children influence on family purchasing decisions.
Analysis that is presented in this paper uses just a fraction of data collected during a much broader survey. The part of the questionnaire that is used in the current study has been developed on the basis of studies that emphasize children's perceived influence differences in buying various products (Aslan, & Karalar, 2011).
In this analysis we used 14 questions about the influence of a child on parents' decisions to buy specific products that have been selected during the pilot study. The products represented two large categories: products for the use of the whole family (family-consumed products) and products for the individual use of a child (childconsumed products). Each category was divided into subcategories. Family-consumed products included two subcategories: tangible products (computers, furniture, cars and TV sets) and services (restaurant services, holiday vacations). Child-consumed products included four subcategories: knowledge-consumed products (books, school supplies), toys (toys and computer games), food (ice cream and cornflakes), and publicly visible products (clothes and bicycles). Respondents were asked to indicate influence of a child in the case of purchasing each of the above mentioned product on a 10 point scale (1 -no influence; 10 -very strong influence).
The questionnaire also included questions about the typical demographic characteristics of the respondent (age, employment and education) and the child (gender and age). The questionnaire was presented in the Lithuanian language for the respondents from Lithuania and was translated into Russian for the respondents from Azerbaijan (the Russian language is still commonly understood in larger cities of this country).
The analysis is based on 200 questionnaires (100 per each country). Initial analysis showed satisfactory Cronbach's alpha level for all product groups, except for "food for a child": The gender representation of children in the overall sample was roughly even, the sample was made up of 51% of boys and 49% of girls (in Lithuania -50% and 50%, in Azerbaijan -52% and 48%). Age-wise, the sample was broken into two groups: 4-12 years old (59%) and 13-18 years old (41%). With regard to the mothers' age distribution, mothers aged 36-55 accounted for 59% of the sample, and those aged 26-35 for 41%.

The overall influence of a child on parents' decisions to purchase products
It was expected that children will have stronger influence on parents' decision to purchase products for their own use than products for the whole family. Paired sample t-test showed significant difference (t=15.497, p=0.000). Children had stronger influence in regard to products for themselves (M=6.95) than for products for the family use (M=3.88) ( Table 2). Therefore hypothesis H 1 has been confirmed.
The second hypothesis tested the differences of the perceived influence for tangible products versus services for the use of the whole family. The results pointed to the stronger influence with regard to services (M=4.43) than to tangible products (M=3.54) (t= 4.654, p=0.000) ( Table 2), confirming H 2 .
With regard to child-consumed products, four product groups have been analyzed: public products, knowledge-related products, toys, and food. Repeated measures ANOVA with the Greenhouse-Geisser correction disclosed that children's influence on parents' decision among these product groups was statistically significantly different (F=8,010, p=0.000). Post hoc tests using the Bonferroni correction revealed that children were perceived to have lower influence on parents' decision to buy public products than to buy food, toys or knowledge-related products (M public =6.54 vs. M food =7.14, M toys =7.47, M knowledge =7.59, p<0.05). However, there was no statistically significant difference among food, toys, and knowledge-related products (Table 2). Therefore no support was found for hypothesis H 3 . In fact, the opposite result was obtained, i.e., children's influence in the case of purchasing publicly visible products was lower than in the three other compared product groups. This finding requires additional discussion and needs new interpretation.

Comparison of children influence in the two countries
Differences of the cultural contexts in the two countries suggest that there should be differences in a child's influence in Azerbaijan and Lithuania. Based on the two Hofstede's dimensions (individualism and uncertainty avoidance), we hypothesized that a child's influence generally is stronger in Azerbaijan (H 4 ). This hypothesis was first tested for all the surveyed products together, and then -for the largest analyzed product groups, i.e., child-and family-consumed products.
Based on all these results, hypothesis H 4 (i.e., children's perceived influence on parents' purchase decision-making is stronger in a more collectivistic culture with lower uncertainty avoidance than in a culture that is more individualistic and stronger oriented towards uncertainty avoidance) has been rejected. However, analysis by smaller product groups showed rather different results. First of all, results for various child-consumed products varied between the countries. It appeared that Lithuanian children had stronger influence with regard to purchasing food products (M=7.31) than Azerbaijani children (M=6.03, t=3.520, p=0.001) ( Table 3). In all other product groups, as it was expected, Azerbaijani children had stronger influence than Lithuanian children. This explains why analysis by the aggregate product group (all child-consumed products) did not bring the expected result: one group is distorting the generalized result due to its opposite difference.
In the cases of other three product groups, the influence of Azerbaijani children was noticeably stronger. Significant differences were found for public products (M PG Lith =6.17, M PG Azerb =7.12, t= -2.322, p=0.021), for toys (M T Lith =6.97, M T Azerb =8.10, t= -2.556, p=0.012), and for knowledge-related products (M Kn Lith = 6.97, M Kn Azerb =8.44, t= -3.357, p=0.001) ( Table 3). These findings suggest that aggregate product groups might hide differences that occur in analysis of smaller product groups, and this might be the reason for differences of findings in former studies. It might be reasonable to suggest avoiding generalizations that are based on large product groups; instead, analysis that is based on more specific product items might generate unexpected findings.

Discussion and conclusions
The aim of this study was to deepen knowledge of perceived children's influence in family purchasing in different cultures. For this, Azerbaijan and Lithuania have been chosen as the countries of comparison. The two countries are comparable when it comes to the Hofstede power distance dimension, but have very different levels of individualism and uncertainty avoidance -the cultural dimensions that are typically considered as the most important in regard to child influence research (Shergill et al., 2013).
The results of the current study suggest that children exert stronger influence when families buy products for their personal use than when products are purchased for the use of the whole family. In this sense, we have confirmed numerous previous studies.
More interesting results have been observed in the analysis of smaller product groups.
We assumed that large product categories do not necessarily allow disclosing peculiarities and differences of children influences in purchasing. At a minimum, a researcher may consider separate measurements in regard to tangible products versus services, since we predict that child influence in these two instances is significantly different. Therefore we have developed H 2 , based on the consideration about the different participation of a child in the usage of these two categories. Children definitely have rather specific needs and priorities in the analysed services (restaurant services, and holiday vacations). In the case of restaurants, they might want or prefer certain types of food and environment; in the case of vacations -the spectrum of their specific needs might be much broader, ranging from travelling and accommodation conditions to special leisure activities and time planning of the whole event (because the trip has to fit the child's school schedules, etc.). Parents are willing to address all these specificities when planning and purchasing services for the family.
Even more interesting results have been obtained in the analysis of product groups directly used by children. We have expected that children could be rather concerned about the products they use in public. In order to balance up the stereotype priorities of genders, we assumed that one of the analysed products of this category might be more important for girls (clothes), another -for boys (bicycle). However, the assumption about the relatively higher influence of children in purchasing publicly visible (and perhaps socially important) products was wrong. This can be explained by at least one of the following reasons: -children do not see analysed publicly visible products as being more important than products from other groups (they also have aspects of visibility and may influence the status of a child among his/her peers); -being concerned about children socialization, parents take responsibilities to use their own influence in buying these product on the basis of their own understanding and priorities, thus partially ignoring the opinion of a child.
Both these reasons need additional analyses and require specific studies. Therefore this aspect could be considered as a possible avenue for future research.
Analysis on the basis of smaller product groups that are intended for a child's personal use brought about one more rather unexpected insight. In one of the groups (food items), children from a more individualistic culture that also has higher uncertainty avoidance score (Lithuania) had stronger influence than their peers from Azerbaijan, which was opposite to the theoretical assumptions. This suggests that there is something specific about food category or at least about the specific products (ice cream and cornflakes) that have been used in this analysis. The two products have been chosen with an assumption that both of them are consumed individually, with no or little sharing with others within a family. It could be expected that in this case parents are willing to more rely on the opinions of children. However, the data does not support this premise: the influence of children in purchasing food is the lowest among all the four product groups (the average for the whole sample equals to 7.14). More than this -it does not explain the 'opposite' difference in this aspect between the analysed countries (why the influence is stronger in Lithuania). It seems that the issue requires deeper analysis in terms of eating habits or overall food importance in the two cultures. Also, some influence might occur because of the market factors: availability and range of the products, their price, disposable income of families, etc. Again, this suggests an interesting direction for the future studies.
The study also yields one more general conclusion on the methodology of similar studies. Analysis showed different (or opposite) results in various rather small product groups. We suppose that the use of large aggregated groups hides these specifics and in some instances does not allow seeing valid results. Therefore we see the reason to continue research covering rather small groups that are closely related with the way products are consumed by children.
However, this study has several limitations. The main of them is related with rather small samples in both countries that cannot be seen as representative ones. They are sufficient for the presented findings, but did not allow deeper analysis. Also, a number of additional aspects could possibly be disclosed if the analysis included more product groups. Unfortunately, the study could not control market supply-side factors (products, prices) and family income (or disposable income).
One of the limitations of the study is related with the use of Hofstede's dimensions. First of all, their reliable measures for Azerbaijan are not available yet; on the other hand -many other approaches that help explain cultural similarities and/or differences may be explored (Ramzy et al., 2012). This perhaps could lead to deeper culture-related insights that this study is lacking.