1
Predicting Who Will be Most Susceptible to Peer Influence Regarding Substance Use:
Individual, Familial, and Peer Risk Factors
Joseph Allen, Joanna Chango, Megan Schad, Dave Szwedo
University of Virginia
Paper Presented to the Biennial Meeting of the Society for Research in Adolescence
Philadelphia, PA, March, 2010
I. Appreciation to Whitney & Brad for organizing this symposium. SLIDE#:
II. SLIDE#:Peer Influence in Adolescence
A. As we think about peer influences in adolescence we start with 4 basic premises:
1. 1. Peer influence processes in adolescence are inevitable (and we’d argue they are just as inevitable in adulthood as well).
2. Two, These processes are not at all restricted to low functioning teens (indeed we’ve found some of the most popular teens are some of those most likely to be attending to and following peer norms),
3. Third, Peer influences can be either positive or negative in valence.
a. Some teens get influenced to do drugs, but others compete with their peers in terms of how well they perform on sports teams and what college they can get in to.
4. And finally, in spite of their inevitability, we’ve had some trouble as a field in finding peer influences that are as substantial as we might expect. Much that we’d once thought was influence simply reflects peers selecting other peers who happen to be like them, and thus they look similar.
B. So, we’ve got a bit of a conundrum:
1. We have a process that seems incredibly powerful…that we all feel like we’ve seen and experienced…and yet it’s hard to detect empirically.
2. Today, we wanted to propose a way of explaining this conundrum.
C. SLIDE#:Key Premise
1. Perhaps influence processes really the same for all teens, or even for the same teen under different conditions?
a. In other words, perhaps peer influences are very powerful for some teens, and under some conditions, but relatively modest at other times. Thus we have a powerful force, but one we only see occasionally.
2. __ And that leads to our core question for our paper today:
What determines who is most likely to be influenced by peers?
a. In other words, when are these influence processes strong, and when are they likely to be not so strong?
D. And we’re going to suggest that the answer to this question is going to require viewing peer influence processes in terms of a larger developmental context.
E. SLIDE#:Theoretical Framework Fortunately, peer influence processes can be understood very well from a developmental perspective. __
F. We view them as an extension of the processes by which the adolescent is seeking to negotiate his or her autonomy in important relationships, and that’s a process that as a field we’ve already done a good deal of research in examining.
1. __ Autonomy development is a task that we see as beginning with parents, but then rapidly becoming equally, if not more important within the peer world.
2. And in fact Larry Steinberg and others have described a transfer process, in which teens, as they first gain autonomy from parents, are willing to live with less autonomy with peers…in part as they simply shift their source of influence from parents to peers in various domains.
__ So from that perspective, autonomy development will be critical to understanding peer influence processes in adolescence.
__ we need to view these processes as existing within a larger developmental context.
And that this context can actually help us understand which teens will be most affected by peer influence processes.
3. SLIDE#:Hypotheses Which leads to the following set of hypotheses.
4. We proposed that peer influence would be linked to several contextual factors in teens’ lives, including:
a. __ their family experiences establishing, or failing to establish autonomy while maintaining a sense of connection
b. __ their own individual developing skills at REFUSING to go along with peer influences when they don’t want to; and
c. __ and the status of their close friends within the larger peer group.
G. And that’s precisely what we set about testing.
III. Methods
A. SLIDE#Sample : We examined these questions in a sample of 184 adolescents, originally drawn from the public school system of Charlottesville, Virginia, who were interviewed and observed in interactions along with their parents and best friends.
1. The sample was evenly divided between males and females,
2. Adolescents were initially assessed at age 13, and have been re-assessed annually ever since.
3. But we’re going to focus today on the period of PEAK peer influence, which both our intuition, and our analyses suggest is between ages 15 and 16.
4. The sample was a normal community sample, and was representative of the population of the surrounding community in both socioeconomic and racial/ethnic terms.
5. We’ve managed to maintain extremely low attrition over time (we’ve had one participant become deceased; one disabled, and 1 participant dropout and say don’t contact us further, and at any given data point we typically have better than 90% participation.
IV. SLIDE#:Defining Influence
A. __ So, we’re going to operationalized peer influence behaviorally…not in terms of how much people say a teen is influenced by peers, but in terms of how much teens actually change their behavior and become more like their peers over time.
B. __ And we’ll focus on adolescent substance use as our key variable, both __ because it’s problematic in its own right, and because it’s probably the single behavior where parents have the greatest concern about peer influence.
C. SLIDE#: Measures substance use
1. We assessed substance use using the Alcohol and Drug Use Questionnaire from the widely regarded monitoring the future survey, and we look at the combined frequency of teens’ self reported alcohol and marijuana use in the past 30 days.
2. We assess this both for the teen, and by bringing in the teen’s closest friend and having that friend report about their own use.
D. So, let’s look at how we do this. SLIDE#:Assessing Peer Influence
1. __ What we’re in essence going to do is look at how a teen changes in their own substance use over the course of a one-year period, from age 15 to age 16.
2. __ Then we’re going to examine the extent to which that change can be predicted by their close friend’s substance use at baseline.
3. __ Or, is a teen’s use changing in the direction of their friends’ use over time?
4. __ And I should note that this is all predicting relative change in the context of a general pattern of increasing substance over this period in the sample as a whole.
E. So let’s look at some hypothetical examples to get the idea of this:
1. SLIDE#:
2. In this first slide, we see teen change in substance use on the y axis, and close friend baseline level of use on the x axis, and in this __ hypothetical example, we see a flat line as the graph shows, there’s no relation between the two.
a. __ If we’d found this, it would indicate that changes in teen use were unrelated to what their close friend was doing.
3. SLIDE#:
4. Alternatively, the next slide shows what it would look like if we had real evidence of peer influence…__ here we’d see that as the friend’s level of use was higher at baseline, target teens’ would be tending to change at a higher rate over time __ (that is they’re age 16 level of abuse would be higher, even after accounting for where they started at age 15.).
SLIDE#:Actual Peer Influence
And looking at this next slide, based on our actual data, that’s pretty much exactly what we find.
a. And we’ll show all the graphs today using this same scaling and using standard scores for both measures, so you can literally get a sense of how effect sizes compare visually.
5. SLIDE#: Predicting Target Teen Substance Use
a. We can see this same finding depicted differently on this next slide.
b. __ Here we see, first that levels of teen use ina 30 day period actually displays strikingly high stability from age 15 to 16…high enough that it would raise the question of whether we really can understand any change that occurs.
c. __ But then we see that if we know about peer’s substance use…we actually predict a good deal of the future teen use…almost 9% of the variance, even after accounting for age 15 use.
6. So that’s a good sized prediction (and as I mentioned, we picked this age to look at because it’s the age at which peer influence appears its largest), but it’s still not all that huge.
7. SLIDE#:Key Question The key question then, was would SOME teen peer dyads account for more of this apparent influence than others? And which youths would that be?
8. SLIDE#:Hypotheses So, in line with our idea of looking developmentally at this process, let’s start by looking at autonomy and relatedness processes within the family.
9. SLIDE#:Measures: Recantations
a. We start by looking at how the teen handles autonomy processes when interacting with their mother at age 13, as an early marker.
10. We bring parents and teens together and ask them to talk about an ongoing disagreement that we’ve identified by interviewing them separately. They talk for 8 minutes and we code this interaction using the Autonomy & Relatedness coding system
a. What we focus on specifically here is the extent to which the teen backs down, or recants their position.
b. We define this as a change of position that does NOT appear to reflect being convinced by a parent’s argument…for example when a teen changes their position before their mother has offered them any reason to do so.
c. And we see this, at age 13, as an indication of a fundamental problem establishing autonomy within the family.
11. So, we’re going to look at how these recantations interact with peers’ baseline use to predict change in teen use over time.
12. SLIDE#:Predicting Substance Use (Recantations as a moderator) Looking at this next slide makes this more clear.
a. We see first, that there’s no main effect for backing down easily with mothers in arguments…these don’t lead to more or less substance use
b. __ BUT, there IS an interaction, as we would predict.
c. And it’s easiest to see on this next slide SLIDE#:Recantations Graph
d. __ what we see first is that for teens who DON’T recant in arguments with their mother, we see very little relation between peer use and teen change over time…
i. very close to the hypothetical flat line we saw above, suggesting that
ii. these teens don’t appear to be being very influenced.
e. __ However for teens who are higher in recantations at 13 (that is 2 years before we begin assessing substance use), we see a much steeper line
f. They are much more likely to be influenced by their peers’ substance use.
g. Notably, recantations themselves don’t predict teens’ substance use…rather, its how recantations interact with peer use…
i. That’s exactly what we’d expect…these teens who struggle with autonomy in the family at 13 look like they’re more susceptible to peer influence…but that influence is going to depend on who they end up with as peers.
13. SLIDE#:Measures Maternal support
a. But we can look at this autonomy picture from a different perspective…our work and others’ work has suggested that autonomy is easiest to establish in the context of a supportive relationship…we don’t mind disagreeing if we feel a relationship is solid. So in that vein, we turn next to look at maternal support.
b. We assess this with a slightly different task, a supportive behavior interaction task, in which the teen is asked to talk with a parent about a situation where the teen feels that he or she could use some advice or support.
14. Again we find an interaction as we would predict, and its displayed on the next chart.
15. SLIDE#: Maternal support graph:
a. First, as before, there’s no main effect of maternal support on teen substance use.
b. __ But, similarly to what we see before, teens with a high degree of maternal support appear relatively unaffected over time by their best friend’s substance use
c. __ but this is not the case for teens lacking in maternal support. They change their substance use in the direction of their peer’s use over time.
d. Notably, the effects for maternal support and teen recantations are both independent…that is if we put them into a model at the same time, both contribute to explaining teen substance use, and together they account for about 6% of the variance in teen use at follow-up, even after accounting for baseline teen use.
16. SLIDE#:Hypotheses Next, we wanted to take a look, not at the family, but at a marker of the adolescents’ own individual skill development, so we took a look at teens’ peer refusal skills. SLIDE#:Peer Refusal Skills.
a. We use an analogue measure to get at this, in which teens are presented with a hypothetical problem situation:
i. For example, imagine you’re in a department store with a friend who says, hey, stick this shirt under your coat and we can each get one free…what would you do?
ii. We then code the teen’s response for its competence, not only in terms of staying out of jail, but also in terms of managing the relationship with his friend.
b. And that’s the basic procedure, and we went through it with 5 different situations.