Social control and smoking: examining the moderating effects of different dimensions of relationship quality.
Keywords: social control, relationship quality, smoking, health behavior, couples
Smoking is an important risk factor for a variety of life-threatening diseases, such as cardiovascular disease and cancer (Mokdad, Marks, Stroup, & Gerberding, 2004). Despite this well-known fact, 27% of the adult population in Switzerland are smokers (Keller, Radtke, Krebs, & Hornung, 2011). Like every change of health behavior, quitting smoking is a complex and difficult endeavor.
There is growing evidence that a person's partner influences his or her health behaviors in general and smoking behavior in particular (e.g., Homish & Leonard, 2005; Rohrbaugh, Shoham, Skoyen, Jensen, & Mehl, 2012). One such influence is social control. Social control can be defined as attempts of influencing and regulating another person's health behavior even if this person is not willing to change (Lewis & Rook, 1999). The most common differentiation of social control strategies is between positive and negative control (e.g., Butterfield & Lewis, 2002). Positive social control refers to strategies that are generally perceived as nice or positive behaviors, such as making compliments whenever the target person engages in the new health behavior or discussing this behavior with the target person. Negative control strategies in contrast rather refer to behaviors that are generally understood as "bad" or unfriendly and emerge in form of nagging or threatening to withdraw affection if the target person engages in the behavior that is to be changed (Butterfield & Lewis, 2002).
This study sets out to examine whether received positive and negative social control from one's partner is beneficial for a person's smoking behavior and whether and how this depends on different aspects of their relationship quality.
Social Control and Health Behaviors
Social control seems to be quite prevalent in all kinds of social relationships and the partner is usually the most prominent person to execute social control (Tucker, 2002; Umberson, 1992). Results on the effects of social control and health behavior, however, are not unambiguous. Whereas some studies reported positive effects (e.g., Lewis & Rook, 1999), others found negative effects of social control on health behaviors (e.g., Helgeson, Novak, Lepore, & Eton, 2004). In the literature on social control and health behaviors, most studies applied a general measure of health behaviors and relatively few studies focused explicitly on the associations between smoking and social control. Existing studies on smoking behavior demonstrate rather positive effects of social control on smoking (Rook, Thuras, & Lewis, 1990; Umberson, 1992; Westmaas, Wild, & Ferrence, 2002) so far.
With regard to positive and negative social control strategies, positive social control seems to be more beneficial in terms of health behavior change of the target person than negative social control (Lewis & Butterfield, 2007; Stephens, Rook, Franks, Khan, & Iida, 2010; Tucker & Anders, 2001). Indicators of these more beneficial effects are not only the health behavior per se, but different behavioral reactions of the target person after receiving social control, such as not hiding the unhealthy behavior from one's partner (Tucker, 2002). Tucker and Anders (2001) for example found that individuals who reported being positively socially controlled indeed changed their health behavior in the partner-desired direction whereas negative control was related to more hiding of the unhealthy behavior. To the best of our knowledge, however, there is no study explicitly examining potentially different associations between positive and negative social control on smoking or hiding smoking.
The Contextual Model: Relationship Quality as a Moderator
Due to the mixed results of social control on general health behaviors, Okun, Huff, August, and Rook (2007) suggested a contextual model which specifies relationship quality as a moderator of the association between social control and health behavior. This model suggests that social control is only effective in promoting health-behavior change in individuals reporting a high relationship quality with the controlling partner. Recent evidence confirms this model's assumptions with regard to the targeted health behavior (Knoll, Burkert, Scholz, Roigas, & Gralla, 2012; Tucker, 2002), however, again this does not seem to be the case in all studies addressing this topic (cf. Okun et al., 2007). Moreover, higher relationship satisfaction also seems to be beneficial in terms of the association between social control and hiding the unhealthy behavior as this was less reported in people with higher relationship satisfaction (Okun et al., 2007; Tucker, 2002).
So far, studies considering relationship quality as a potential moderator either focused on relationship satisfaction (Knoll et al., 2012; Tucker, 2002) or relationship quality as a one-dimensional construct (Okun et al., 2007). There is an ongoing debate on the conceptualization and assessment of relationship quality (Gottman & Notarius, 2002; Hassebrauck & Fehr, 2002) resulting in a vast amount of concepts and measures. In order to be able to gain a more differentiated understanding of the role of relationship quality in the dyadic context of social control and smoking cessation, the focus of the present study was on the multidimensional conceptualization of relationship quality as dyadic adjustment (Spanier, 1976). Dyadic adjustment is a relational construct consisting of subjective evaluations of the relationship with regard to different dimensions such as consensus, cohesion and satisfaction (Spanier & Lewis, 1980). Consensus refers to an agreement with the partner with regard to central domains, such as values, friends and family. Cohesion addresses the companionship or team spirit of the couple. Finally, satisfaction refers to the evaluation whether or not things are going well in the relationship. To the best of our knowledge no study so far examined the contextual model (Okun et al., 2007) in the context of smoking behavior with a differentiation between positive and negative control, as well as a multidimensional assessment of relationship quality.
Aims of the Present Study
The aims of the present study were twofold. First, we were interested in examining the effects of positive and negative social control of smokers received from a nonsmoking partner on the numbers of cigarettes smoked and on hiding smoking. We hypothesized that positive social control should display more beneficial effects on smoking and hiding smoking than negative social control. Second, we aimed at providing a detailed analysis of the moderating function of different dimensions of relationship quality (i.e., consensus, cohesion and satisfaction) on the association between positive and negative social control and smoking and hiding smoking. For the dimension relationship satisfaction, we hypothesize that higher relationship satisfaction goes along with a more positive effect of positive and negative social control on smoking (i.e., resulting in fewer cigarettes smoked) and on hiding smoking (i.e., resulting in less hiding smoking). Because of the lack of existing findings with regard to cohesion and consensus in the contextual model of social control, we applied an exploratory approach for the moderating function of these two dimensions.
Sample and Procedure
This correlational longitudinal online study comprised two points of measurement: a baseline assessment (T1) and a follow-up that took place 4 weeks later (T2). The study was advertised on different smoking-related webpages, via mailing lists of staff and students at a Swiss University and participants of the Swiss Tobacco Monitoring System (Keller et al., 2011) as well as via flyers that were distributed among smoking individuals. All participants provided informed consent before entering the online-questionnaire. Moreover, participants generated a personal code in order to ensure anonymity. This code was used to match the first and the second online questionnaire data. Four weeks after the baseline, participants were invited by email to fill in the second online questionnaire. If after one week participants had not followed the invitation, a second invitation email was sent. After completing both online questionnaires participants received a voucher for books, or other items of an online book store worth 20 Swiss Francs.
Inclusion criteria were being a regular smoker (i.e., smoking daily or several times a week), being married or in a committed (heterosexual) relationship for at least 1 year and being at least 18 years old. Moreover, the partner of participants, who was the source of the social control in this study, was required to be nonsmoking. All individuals were treated according to the ethical guidelines of the American Psychological Association (American Psychological Association, 2002).
Overall, 144 individuals (n = 72 women, 50%) took part in the baseline assessment. The mean age of the sample was M = 31.78, SD = 10.04 (range = 20-67 years). The majority of participants were living with their partner without being married (n = 100, 69.4%) and mean duration of relationship was 8.33 years (SD = 9.04). Most participants reported having attended 12 years of school (n = 115, 79.9%), indicating a rather highly educated sample. The average number of smoked cigarettes per day was 10.99 (SD = 8.20). Of the 144 smokers, n = 105 reported to smoke daily and n = 39 reported to smoke several times a week. (1)
Of the original 144 participants at baseline, 118 (82%) completed the 4 weeks follow-up. In terms of social control, relationship quality, number of cigarettes smoked per day, social desirability, gender, marital status and years of schooling no significant differences emerged between dropouts and continuers. There was a significant difference between dropouts and continuers for age, t([57.sup.2]) = -2.66, p = .01 (M = 28.15, SD = 6.93 for dropouts, M = 32.68, SD = 10.51 for continuers), and for duration of relationship, t([37.sup.2]) = - 2.5, p = .02 (M = 4.58, SD = 7.86 for dropouts, M = 9.11, SD = 9.10 for continuers).
All item examples of the scales described below are translated from German. Means are pooled means from 5 imputed data sets. Because pooling of standard deviations is not available, ranges of standard deviations (SDs) across the 5 imputed data sets are reported. Cronbach's alpha stems from the original (i.e., unimputed) data set as imputations were done on scale level only.
Social control was assessed at baseline with items adapted from Butterfield and Lewis (2002). All items followed the stem: "My partner tried to influence my smoking behavior by ... ". Positive social control (M = 2.03, SD = 1.20-1.27, Cronbach's alpha = .92) comprised seven items, such as "... making suggestions." Negative social control (M = 1.90, SD = 0.971.01, Cronbach's alpha = .78) comprised five items, such as " ... withdrawing affection." Response format was 1 "never" to 7 "at least once a day." A principle component analysis with all social control items using an Eigenvalue >1 criterion and varimax rotation resulted in a clear two-factor structure (first factor explained 54.5% variance before and 38% variance after rotation, second factor explained 9.70% variance before and 26.2% after rotation). The lowest factor loading of the positive control items was .59, the lowest factor loading of the negative control items was .74. No double loadings emerged. This loading pattern indicates that despite the translation into another language and potential cultural differences between Switzerland and the U.S.A., the perception of negative and positive social control strategies could be captured unambiguously with the items used. For the interested readers, all items are displayed in original English and translated German version in the Appendix.
Relationship quality was assessed at baseline by the German short version of the Dyadic Adjustment Scale (DAS-12; German version: Dinkel & Balck, 2006). The German scale comprises three subscales: consensus (M = 4.68, SD = 0.55-0.58, Cronbach's alpha = .66), cohesion (M = 4.27, SD = 0.91-0.93, Cronbach's alpha = .83), and satisfaction (M = 5.02, SD = 0.64-0.66, Cronbach's alpha = .75), which were assessed with four items each. To answer the items measuring consensus participants were asked to indicate the amount of agreement with regard to important domains, such as friends, and making important decisions. Response format was from 1 "always disagree" to 6 "always agree." The items measuring cohesion comprised questions such as "How often do you laugh together?" Response format was from 1 "never" to 6 "more often than once a day." Satisfaction with the relationship was assessed with 4 items, such as "How often do you think that things are going well between your partner and you?" Response format was from 1 "never" to 6 "always."
Number of cigarettes smoked was assessed at baseline (M = 10.92, SD = 8.18-8.27) as well as at Time 2 (M = 10.93, SD = 8.87-9.67) by using a single item from the Fagerstrom-Test (Heatherton, Kozlowski, Frecker, & Fagerstrom, 1991): "How many cigarettes per day do you smoke?" with an open answering format. Single-item measures are always less reliable than multiple-item measures. In case of number of cigarettes smoked, however, a multiple item measure would be rather artificial and thus not useful. With regard to the validity of this global assessment of number of cigarettes smoked, critical points are the existence of a digit bias (i.e., participants' report of cigarettes smoked cluster around rounded values, such as around 20 cigarettes = 1 package per day). And daily variability in smoking patterns cannot be assessed validly with global retrospective reports (Shiffman, 2009). The present study aimed at investigating between-person associations with number of cigarettes smoked. As demonstrated by Shiffman (2009), a global measure as used here validly assesses light, moderate and heavy smoking despite the digit bias. Thus, for the purpose of the present study, this global assessment of number of cigarettes smoked is a valid and useful measure.
Hiding smoking behavior was assessed at baseline (M = 1.75, SD = 1.18-1.21), as well as at T2 (M = 1.70, SD = 1.08-1.19) by a single item from Tucker and Anders (2001): "How often did you hide your smoking from your partner?" Response format was 1 "never" to 7 "always."
Social desirability was assessed at baseline by the SDS-17 (Stoeber, 2001) and served as a control variable. It was assessed at T1 only (M = 10.30, SD = 2.63-2.77, Cronbach's alpha = .58). Answering format was dichotomous (1 no, 2 yes). Higher values indicated higher social desirability.
Duration of relationship was assessed at baseline by asking "How long have you and your partner been together for?" and also served as control variable. Responses could be given in years and months and were converted in years for the analyses (M = 8.27, SD = 8.94-9.12).
Because attrition seemed to be systematically biased, multiple imputation (MI) was applied to account for missing data (Graham, 2009) using SPSS 19. We included all central variables as well as the missing-related variables (age and relationship duration) into the imputation model. MI accounts for missing-data uncertainty by generating multiple values for respective points of missing data, creating multiple data-sets, and introducing between-imputation variance. Each of these data-sets was analyzed separately. Results were integrated in a last step using SPSS 19 to obtain overall estimates and standard errors. Using MI, however, no standardized regression weights, pooled standard deviations or R2s can be obtained. Thus, the range of these coefficients across the imputed data sets is presented. For the present study, five datasets were generated and analyzed.
Our main analyses were hierarchical regression analyses in order to test the moderation hypotheses. For the prediction of change in number of cigarettes smoked and change in hiding smoking at T2 the baseline measures of the respective outcome together with the control variables (social desirability and relationship duration) were entered together with the target subscale of relationship quality and either positive or negative social control in the first step. In a second step, the interaction terms of positive/negative social control and the respective subscale of relationship quality were entered. For the examination of possible interaction effects between positive or negative social control and either consensus, cohesion, or satisfaction all these variables were grand-mean centered to avoid problems with multicollinearity (Cohen, Cohen, West, & Aiken, 2003). Interaction terms were generated by multiplying the mean-centered variables. To display the interaction effects, low and high values of the variables were generated by using the lower and upper possible values of predictor and moderator (Cohen et al., 2003). To test whether the regression slopes were significant, we ran simple slope analyses (Preacher, Curran, & Bauer, 2006). Because the information needed for simple slopes analyses is not available for multiple imputation data sets, we used the information of the first of the five imputed data sets.
In Table 1, all correlations between variables in the study are displayed. As can be seen, relationship duration and social desirability were not related to numbers of cigarettes smoked or hidden smoking. Thus, these variables were not included in the main analyses. A positive association between number of cigarettes and gender at both points of measurement indicate that men smoked more cigarettes than women. Moreover, there was a slight tendency for age to be positively related to numbers of cigarettes. Thus, gender and age were included in the main analyses as control variables.
On a bivariate level, negative control was positively related to number of cigarettes and both positive and negative social control were positively related to hiding one's smoking behavior.
Prediction of Change in Numbers of Cigarettes Smoked
The results of the regression analyses on the interplay of positive and negative social control with the three dimensions of relationship quality for the prediction of change in numbers of cigarettes smoked are displayed in Table 2. Overall, there was a rather strong effect of the baseline of numbers of cigarettes smoked. In addition to this baseline effect, only the interaction between positive social control and consensus was significant. This interaction (displayed in Figure 1) indicates that a lower consensus goes along with overall more cigarettes smoked and no association between social control and change in number of cigarettes smoked, simple slope analysis: -0.61 (0.50), t = - 1.22, p = .23. In contrast, there was a negative association between positive social control and change in number of cigarettes smoked for those individuals who reported a high consensus with their partner, simple slope analysis: -6.99 (2.95), t = -2.37, p = .02.
When baseline control of numbers of cigarettes smoked was excluded from the models with positive social control, gender displayed a significant positive effect (indicating that men reported more cigarettes smoked at follow-up than women) in addition to the ones displayed in Table 2. Likewise, when baseline control of numbers of cigarettes smoked was excluded from the models with negative social control, gender displayed a significant positive effect (indicating that men reported more cigarettes smoked at follow-up than women) and negative social control displayed positive marginally significant effects in addition to the ones displayed in Table 2.
Prediction of Change in Hiding Smoking
The results of the regression analyses on the interplay of positive and negative social control with the three dimensions of relationship quality for the prediction of change in hiding smoking are displayed in Table 3. Again, there was a rather strong effect of the baseline of hiding smoking behavior. In addition to this baseline effect, there was again the interaction between positive social control and consensus that became significant on the 10%-level. This interaction (displayed in Figure 2) indicates that a lower consensus went along with overall less hiding of smoking and with a slight positive association between positive control and hiding, simple slope analysis: 0.34 (0.11), t = 3.18, p = .01. In contrast there was a stronger positive association between positive social control and change in hiding smoking behavior for those individuals who reported a high consensus with their partner, simple slope analysis: 1.59 (0.58), t = 2.77, p = .01. Very similar interaction effects emerged for the interaction between positive social control and cohesion, simple slope for low cohesion: 0.21 (0.09), t = 2.38, p = .02; simple slope for high cohesion: 1.22 (0.46), t = 2.64, p = .01, as well as for negative social control and consensus, simple slope for low consensus: 1.45 (0.35), t = 4.14, p = .01; simple slope for high consensus: 0.38 (0.15), t = 2.57, p = .01.
When baseline control of hiding smoking was excluded from the models with positive social control, positive social control displayed a significant positive effect on hiding smoking at T2 in all three models. Moreover, in the model with consensus as component of relationship quality, consensus also became a significant positive predictor, when baseline control was not included in the model. The very same applied to negative social control as well as to consensus in combination with negative social control.
This study set out to fill an important gap in the literature on social control and health behavior in general and in the literature on the contextual model of social control (Okun et al., 2007) in particular. It was examined whether different dimensions of relationship quality served different moderating functions on the association between positive and negative social control and smoking behavior and hiding smoking.
The main results of this study were that consensus as one dimension of relationship quality seemed to play a moderating function with regard to the association between number of cigarettes smoked and positive control and with regard to the association between hiding smoking and both positive and negative social control. Individuals reporting higher consensus and more positive social control were more likely to reduce their number of cigarettes smoked. In contrast, for individuals with lower consensus the number of cigarettes smoked was independent from received positive social control. Thus, this result is in line with studies focusing on different behaviors than smoking and assessing relationship quality as a one-dimensional construct (Okun et al., 2007) or assessing relationship satisfaction (Knoll et al., 2012; Tucker, 2002). Consensus as conceptualized in this study (Spanier, 1976) stands for an agreement with the partner on important domains in life. Thus, a potential underlying mechanism of our results could be that smokers high in consensus with their nonsmoking partner in general agree on health being an important value and therefore social control attempts of the nonsmoking partner are perceived as appropriate and are associated with better health behavior.
With regard to hiding smoking, a rather unexpected result emerged in that individuals with higher consensus in their relationship reported more hiding of their unhealthy smoking behavior when being both positively or negatively controlled than individuals low in consensus. This is rather in contrast to other studies that reported less hiding as a result of (direct) social control in a high-quality or highly satisfying relationship (e.g., Okun et al., 2007; Tucker, 2002). With regard to a measure of indirect social control, however, Tucker (2002) found similar effects. Indirect social control assesses felt obligations toward others in terms of staying healthy on a more general level or in terms of changing the unhealthy behavior on a more specific level. Although the measures of positive and negative social control used in the present study were measures of direct social control, meaning that the controlling partner rather openly engaged in strategies in order to make the other change their smoking behavior, the mechanisms might be the same as for indirect control: Individuals with a high consensus in their partnership should be in agreement with their partner that health is an important value and that they should stop smoking in order not to endanger their health. Likely, this goes along with felt obligations and responsibilities. When smoking, the obligation to protect one's health is rather jeopardized and hiding smoking might result when craving is too strong.
An alternative explanation might be that smoking could lead to a certain disagreement in the relationship and thus could potentially endanger the high consensus with the partner. In this line of reasoning, hiding smoking behavior might also serve the function of a relationship maintenance strategy (Stafford, 2011) as potential conflicts are avoided. Future studies should thus apply a research design that allows for testing potential reciprocal effects of hiding an unhealthy behavior and consensus.
Although we could in part replicate the moderating effects of relationship quality/relationship satisfaction on social control and health behavior with the subdimension consensus, the most striking unexpected result is the failure to find any moderating effect of the subdimension satisfaction in our study. Potentially, in the global assessments of relationship quality as applied in other studies (Okun et al., 2007) consensus might be the driving factor and not so much relationship satisfaction. Moreover, as this was the first study to test the contextual model in the domain of smoking behavior, it could be that with regard to smoking it is not satisfaction but indeed rather consensus that is the relevant dimension of relationship quality with regard to the association between social control and behavior. Another explanation for this unexpected null-finding could be a methodological one: There was a rather high mean value of relationship satisfaction in our sample, pointing to a ceiling effect which makes it difficult to find any moderating effect. Future studies are clearly needed to replicate these results in a more heterogeneous sample with regard to reported relationship satisfaction.
The third subdimension cohesion resulted only in one significant interaction effect: Higher reports of cohesion and of positive social control were rather strongly positively associated with hiding smoking behavior, whereas for lower levels of cohesion the association between positive control and hiding was weaker, albeit still positive. Cohesion as measured in this study comprises the companionship (Rook, 1990) or team spirit of the couple (Spanier, 1976). Likely, when cohesion is perceived as high, positive social control attempts of the nonsmoking partner might address this team spirit. In turn, this might trigger the smoker's wish not to disappoint the partner and as a consequence hiding smoking might result. In contrast, negative social control might rather promote reactance in some but not all smokers, which might explain the nonsignificant interaction effect of cohesion and negative social control.
Overall, the different results for the three dimensions of relationship quality that is, consensus, cohesion, and satisfaction across control and smoking as well as hiding smoking provide evidence for the usefulness of these distinctions.
Results of this study further demonstrate that the distinction between positive and negative social control strategies seems to be useful. It is important to note, however, that the conceptualization applied in this and other studies (Butterfield & Lewis, 2002) is rather a lay person perspective of nice (i.e., positive) and bad (i.e., negative) behaviors. This is a sensible way to differentiate these two kinds of social control as long as there is an agreement on what is positive and negative. Some strategies, such as "involving other people," might be ambiguous, however. Thus, future research might want to further this distinction by for example, introducing positive and negative social control strategies that refer to basic learning principles, such as reinforcement and punishment in applying or in taking away respective stimuli. This would strengthen the theory-based approach and potentially avoid equivocal findings.
There are several limitations that need to be kept in mind when interpreting the results. First, all measures regarding smoking behavior were self-report. To account for potential bias, we included social desirability. The latter however, potentially because of its low reliability, was unrelated to the reports of numbers of cigarettes smoked and hiding smoking. Moreover, global assessments of self-reported smoking are valid measures with regard to between-person associations only (Shiffman, 2009), which were aimed at in the present study. Validity issues for this measure emerge for within-person relations. Furthermore, online studies have the advantage of higher anonymity compared to common paper and pencil studies and as a consequence are less prone to trigger socially desired responses (Joinson, 1999). Nonetheless, future studies should operate with objective smoking indicators, such as salivary cotinine levels or expired air carbon monoxide (West, Hajek, Stead, & Stapleton, 2005). Moreover, although the response format of hiding smoking has been demonstrated to be useful (e.g., Tucker & Anders, 2001), the "never" to "always" categories leave room for individual interpretation. Future studies should thus apply a response format that refers to real frequencies, such as "never" to "several times a day," in order to clarify responses. Second, this study solely focused on couple constellations of nonsmoking and smoking partners. Future studies are needed that examine whether the results can be replicated in couples with both partners smoking. Third, it could make an important difference for the effectiveness of social control attempts whether or not the smoking partner really intends to stop smoking or not. Thus, the motivational status of the target persons should be taken into account in further research. Fourth, this study's focus was on the control receiver's perspective. Future studies might want to adopt a dyadic approach in order to capture the full story of control dynamics and outcomes in couples in the context of smoking cessation. And finally, as this study applied a longitudinal-correlational design, assumptions of causality are not appropriate.
To conclude, this study could demonstrate that positive and negative social control interacts differently with different dimensions of relationship quality with regard to smoking and hiding smoking.
Appendix Positive and Negative Social Control Items in Original English and Translated German Wording Original English wording Translated German wording. Items from (Butterfield & Lewis, are adapted to the smoking 2002). Items describe context. All items had the stem social control situations "Mein Partner/Meine Partnerin and are behavior versuchte mich dazu zu bewegen, unspecific. weniger zu rauchen, indem er/sie ..." ["My partner tried to influence my smoking behavior by ..."] Make suggestions ... Vorschlage machte. P Praise or compliment the ... mich lobte und mir Komplimente P target machte. Offer to make the change ... mir anbot, die Veranderung mit P with the target mir gemeinsam zu machen. Try to reason with the ... versucht hat, mich mit P target / be logical logischen Argumenten zu uberzeugen. Express positive emotions ... positive Gefuhle zum Ausdruck P brachte. State how important it is ... betonte, wie wichtig es fur P to you ihn/sie ist. Persuade ... versuchte, mich zu fiberreden. P Try to make the target feel ... versucht hat, mir ein N guilty schlechtes Gewissen zu machen. Withdraw, become silent or ... sich zuruckzog und schweigsam N clam up wurde. Express negative emotions ... negative Gefuhle zum Ausdruck N brachte. Drop hints ... Anspielungen machte. N Withdraw affection ... mir seine/ihre Zuneingung N entzogen hat. Note. P = positive social control; N = negative social control.
American Psychological Association. (2002). Ethical principles of psychologists and code of conduct. American Psychologist, 57, 1060-1073. doi: 10.1037/0003-066X.57.12.1060
Butterfield, R. M., & Lewis, M. A. (2002). Health-related social influence: A social ecological perspective on tactic use. Journal of Social and Personal Relationships, 19, 505-526. doi:10.1177/0265407502019004050
Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2003). Applied multiple regression/correlation analysis for the behavioral sciences (3rd ed.). Mahwah, NJ: Erlbaum.
Dinkel, A., & Balck, F. (2006). Psychometrische Analyse der deutschen Dyadic Adjustment Scale. [Psychometric analysis of the German version of the Dyadic Adjustment Scale]. Zeitschrift Fur Psychologie, 214, 1-9. doi:10.1026/0044-3409.214.1.1
Gottman, J. M., & Notarius, C. I. (2002). Marital research in the 20th century and a research agenda for the 21st century. Family Process, 41, 159-197. doi: 10.1111/j. 1545-5300.2002.41203.x
Graham, J. W. (2009). Missing data analysis: Making it work in the real world. Annual Review of Psychology, 60, 549-576. doi: 10.1146/annurev.psych .58.110405.085530
Hassebrauck, M., & Fehr, B. (2002). Dimensions of relationship quality. Personal Relationships, 9, 253-270. doi: 10.1111/1475-6811.00017
Heatherton, T. F., Kozlowski, L. T., Frecker, R. C., & Fagerstrom, K. O. (1991). The Fagerstrom Test for Nicotine Dependence: A revision of the Fagerstrom Tolerance Questionnaire. British Journal of Addiction, 86, 1119-1127. doi:10.1111/j.13600443.1991.tb01879.x
Helgeson, V. S., Novak, S. A., Lepore, S. J., & Eton, D. T. (2004). Spouse social control efforts: Relations to health behavior and well-being among men with prostate cancer. Journal of Social and Personal Relationships, 21, 53-68. doi:10.1177/0265407504039840
Homish, G. G., & Leonard, K. E. (2005). Spousal influence on smoking behaviors in a U.S. community sample of newly married couples. Social Science & Medicine, 61, 2557-2567. doi:10.1016/j.socscimed.2005.05.005
Joinson, A. (1999). Social desirability, anonymity, and Internet-based questionnaires. Behavior Research Methods Instruments & Computers, 31, 433-438. doi: 10.3758/BF03200723
Keller, R., Radtke, T., Krebs, H., & Hornung, R. (2011). Der Tabakkonsum der Schweizer Wohnbevolkerung in den Jahren 2001 bis 2010. Tabakmonitoring --Schweizerische Umfrage zum Tabakkonsum. [Tobacco consumption of the Swiss population during 2001 and 2010. Swiss Tobacco Monitoring System]. Zurich, Germany: Psychologisches Institut der Universitat Zurich, Sozial--und Gesundheitspsychologie.
Knoll, N., Burkert, S., Scholz, U., Roigas, J., & Gralla, O. (2012). The dual-effects model of social control revisited: Relationship satisfaction as a morderator. Anxiety, Stress, & Coping, 25, 291-307. doi: 10.1080/10615806.2011.584188
Lewis, M. A., & Butterfield, R. M. (2007). Social control in marital relationships: Effect of one's partner on health behaviors. Journal of Applied Social Psychology, 37, 298-319. doi:10.1111/j.0021-9029.2007.00161.x
Lewis, M. A., & Rook, K. S. (1999). Social control in personal relationships: Impact on health behaviors and psychological distress. Health Psychology, 18, 63-71. doi:10.1037/0278-618.104.22.168
Mokdad, A. H., Marks, J. S., Stroup, D. F., & Gerberding, J. L. (2004). Actual causes of death in the United States, 2000. JAMA: Journal of the American Medical Association, 291, 1238-1245. doi: 10.1001/jama.291.10.1238
Okun, M. A., Huff, B. P., August, K. J., & Rook, K. S. (2007). Testing hypotheses distilled from four models of the effects of health-related social control. Basic and Applied Social Psychology, 29, 185-193. doi: 10.1080/01973530701332245
Preacher, K. J., Curran, P. J., & Bauer, D. J. (2006). Computational tools for probing interactions in multiple linear regression, multilevel modeling, and latent curve analysis. Journal of Educational and Behavioral Statistics, 31, 437-448. doi: 10.3102/10769986031004437
Rohrbaugh, M. J., Shoham, V., Skoyen, J. A., Jensen, M., & Mehl, M. R. (2012). We-Talk, communal coping, and cessation success in a couple-focused intervention for health-compromised smokers. Family Process, 51, 107-121. doi:10.1111/j.15455300.2012.01388.x
Rook, K. S. (1990). Social relationships as a source of companionship: Implications for older adults' psychological well-being. In B. R. Sarason, I. G. Sarason, & G. R. Pierce (Eds.), Social support: An interactional view (pp. 219-250). Oxford, England: Wiley.
Rook, K. S., Thuras, P. D., & Lewis, M. A. (1990). Social-control, health risk-taking, and psychological distress among the elderly. Psychology and Aging, 5, 327-334. doi:10.1037/0882-7974.5.3 .327
Shiffman, S. (2009). How many cigarettes did you smoke? Assessing cigarette consumption by global report, time-line follow-back, and ecological momentary assessment. Health Psychology, 28, 519526. doi:10.1037/a0015197
Spanier, G. B. (1976). Measuring dyadic adjustment: New scales for assessing quality of marriage and similar dyads. Journal of Marriage and the Family, 38, 15-28. doi: 10.2307/350547
Spanier, G. B., & Lewis, R. A. (1980). Marital quality: A review of the 70s. Journal of Marriage and the Family, 42, 825- 839. doi: 10.2307/351827
Stafford, L. (2011). Measuring relationship maintenance behaviors: Critique and development of the revised relationship maintenance behavior scale. Journal of Social and Personal Relationships, 28, 278-303. doi: 10.1177/0265407510378125
Stephens, M. A. P., Rook, K. S., Franks, M. M., Khan, C., & Iida, M. (2010). Spouses Use of Social Control To Improve Diabetic Patients' Dietary Adherence. Families, Systems, & Health, 28, 199-208. doi:10.1037/a0020513
Stoeber, J. (2001). The Social Desirability Scale-17 (SDS- 17): Convergent validity, discriminant validity, and relationship with age. European Journal of Psychological Assessment, 17, 222-232. doi:10.1027//1015-5722.214.171.124
Tucker, J. S. (2002). Health-related social control within older adults' relationships. The Journals of Gerontology Series B-Psychological Sciences and Social Sciences, 57B, P387-P395. doi:10.1093/geronb/57.5.P387
Tucker, J. S., & Anders, S. L. (2001). Social control of health behaviors in marriage. Journal of Applied Social Psychology, 31, 467-485. doi:10.1111/j.1559-1816.2001.tb02051.x
Umberson, D. (1992). Gender, marital-status and the social-control of health behavior. Social Science & Medicine, 34, 907-917. doi:10.1016/02779536(92)90259-S
West, R., Hajek, P., Stead, L., & Stapleton, J. (2005). Outcome criteria in smoking cessation trials: Proposal for a common standard. Addiction, 100, 299-303. doi: 10.1111/j. 1360-0443.2004.00995.x
Westmaas, J. L., Wild, T. C., & Ferrence, R. (2002). Effects of gender in social control of smoking cessation. Health Psychology, 21, 368-376. doi: 10.1037/0278-6126.96.36.1998
Received September 14, 2012
Revision received April 12, 2013
Accepted April 15, 2013
Urte Scholz, PhD
University of Konstanz, Konstanz, Germany
Philippe Goldammer, MSc
University of Zurich, Zurich, Switzerland
Rainer Hornung, PhD
University of Zurich, Zurich, Switzerland
Corina Berli, MSc
University of Bern, Bern, Switzerland
Janina Luscher, MSc
University of Bern, Bern, Switzerland
Nina Knoll, PhD
Freie Universitaet Berlin, Berlin, Germany
(1) Daily and weekly smokers did not differ with regard to mean levels of negative (t = -.042) and positive (t = -.307) social control.
(2) Degrees of freedom were adjusted because of nonequal variances.
This article was published Online First August 19, 2013. Urte Scholz, PhD, Department of Psychology, Developmental & Health Psychology, University of Konstanz, Konstanz, Germany; Corina Berli, MSc, University of Bern, Bern, Switzerland; Philippe Goldammer, MSc, University of Zurich, Zurich, Switzerland; Janina Luscher, MSc, University of Bern, Bern, Switzerland; Ralner Hornung, PhD, University of Zurich, Zurich, Switzerland; Nina Knoll, PhD, Freie Universitaet Berlin, Berlin, Germany.
The second (Corina Berli) and fourth (Janina Luscher) authors are funded by the Swiss National Science Foundation [PP00PI_133632/1].
Correspondence concerning this article should be addressed to Urte Scholz, PhD, Department of Psychology, Developmental & Health Psychology, University of Konstanz, P.O. Box 5560/14, 78457 Konstanz, Germany. E-mail: email@example.com
Table 1 Correlations of Main Study Variables and Potential Control Variables NCS Hiding Hiding T2 T1 T2 PC NC CON NCS TI .85 ** -.06 -.12 .12 .15# .08 NCS T2 -.12 -.18# .13 .22 * -.01 Hiding T1 .64 ** .30 ** .34 ** .13 Hiding T2 .29 ** .29 ** .19 * PC .70 ** .12 NC .05 CON COH SAT RD SDES Gender (1 = women; 2 = man) COH SAT RD SDES Gender Age NCS TI .12 .12 .13 .04 .34 ** .16# NCS T2 .15# .10 .11 .01 .31 ** .13 Hiding T1 -.07 .11 -.04 .14 .15# -.01 Hiding T2 .09 .03 -.10 .15 .11 -.07 PC .20 * .27 ** -.07 .08 .29 * -.10 NC .07 .17 * -.02 .03 .21 * -.07 CON .21 * .50 ** .09 .08 -.02 .01 COH .42 ** .01 .03 .09 -.02 SAT -.01 .12 .16# -.06 RD .27 ** .01 .85 ** SDES .01 .15# Gender (1 = women; .06 2 = man) Note. NCS = number of cigarettes smoked; Hiding = hiding smoking behavior; PC = positive control; NC = negative control; CON = consensus; COH = cohesion; SAT = satisfaction; RD = relationship duration: SDES = social desirability; Gender: 1 = women, 2 = men. # p<.10. * p<.05. ** p<.01. Table 2 Coefficients of Multiple Regression Analysis for the Prediction of Change in Numbers of Cigarettes Smoked b SE b b SE b NCS 1 .94 ** .09 .95 ** .09 Gender -.01 1.17 .25 1.20 Age .01 .05 -.01 .05 PC .48 .37 .08 .44 NC CON -1.64 1.11 COH .48 .52 SAT PC*CON -1.85 * .82 PC*COH .12 .39 PC*SAT NC*CON NC*COH NC*SAT b SE b b SE b NCS 1 .96 ** .09 .95 ** .09 Gender .27 1.24 -.06 1.07 Age -.01 .05 .01 .05 PC .31 .43 NC .92 .65 CON -1.56 1.22 COH SAT -.29 .87 PC*CON PC*COH PC*SAT -.26 .66 NC*CON -1.52 1.12 NC*COH NC*SAT b SE b b SE b NCS 1 .95 ** .09 .95 ** .09 Gender .01 1.10 .08 1.14 Age .01 .05 .01 .05 PC NC .78 .61 .90 .71 CON COH .45 .48 SAT -.30 .92 PC*CON PC*COH PC*SAT NC*CON NC*COH .48 .55 NC*SAT -.02 .95 Note. Findings are based on 5 imputed data sets. Only final steps of the hierarchical regression analyses are displayed. NCS1 = number of cigarettes smoked at T1, gender: 1 = women, 2 = men; PC = positive social control; NC = negative social control; CON = consensus; COH = cohesion; SAT = satisfaction. PC and CON: 1. step: [R.sup.2] = .66-.79, F(5, 138) = 54.05-102.25, p = .001; 2. step: [DELTA][R.sup.2] = .011-.035, F(1, 137) = 7.32-22.14, p = .001-.01; PC and COH: 1. step: [R.sup.2] = .64-.79, F(5, 138) = 50.02-101.24, p = .001; 2. step: [DELTA][R.sup.2] = .001, F(1, 137) = .013-.69, p = .41-.91; PC and SAT: 1. step: [R.sup.2] = .65-.79, F(5, 138) = 50.27-100.98, p = .001; 2. step: [DELTA][R.sup.2] = .001, F(1, 137) = .01-1.77, p = .19-.97; NC and CON: 1. step: [R.sup.2] = .69-.79, F(5, 138) = 60.31-102.38, p = .001; 2. step: [DELTA][R.sup.2] = .01-.02, F(l, 137) = 1.06-9.93, p = .002-.30; NC and COH: 1. step: [R.sup.2] = .67-.79, F(5, 138) = 54.89-101.47, p = .001; 2. step: [DELTA][R.sup.2] = .001-.01, F(1, 137) = .11-2.20, p = .03-.74; NC and SAT: 1. step: [R.sup.2] = .67-.79, F(5, 138) = 55.95-101.16, p = .001; 2. step: [DELTA][R.sup.2] = .001, F(1, 137) = .04-1.89, p = .17-.84. * p < .05. ** p < .01. Table 3 Coefficients of Multiple Regression Analysis for the Prediction of Change in Hiding Smoking b SE b b SE b HIDEI .55 ** .08 .62 ** .08 Gender .04 .18 -.05 .19 Age -.01 .01 -.01 .01 PC .07 .07 .01 .08 NC CON .23 .15 COH .16 * .08 SAT PC*CON .21# .12 PC*COH .15 * .07 PC*SAT NC*CON NC*COH NC*SAT b SE b b SE b HIDEI .57 ** .08 .55 ** .08 Gender .01 .19 .02 .17 Age -.01 .01 -.01 .01 PC .10 .08 NC .10 .10 CON .24# .14 COH SAT -.12 .15 PC*CON PC*COH PC*SAT .02 .12 NC*CON .32# .16 NC*COH NC*SAT b SE b b SE b HIDEI .58 ** .10 .58 ** .09 Gender -.03 .19 .01 .19 Age -.01 .01 -.01 .01 PC NC .10 .10 .10 .11 CON COH .16# .08 SAT -.10 .14 PC*CON PC*COH PC*SAT NC*CON NC*COH -.09 .09 NC*SAT -.01 .14 Note. Findings are based on 5 imputed data sets. Only final steps of the hierarchical regression analyses are displayed. HIDE] = = hiding smoking at T1, Gender: 1 = women, 2 = men; PC = positive social control; NC negative social control; CON = consensus; COH = cohesion; SAT = satisfaction. PC and CON: 1. step: [R.sup.2] = .41-.50, F(5, 138) 2 = 19.35-27.00, p = .001; 2. step: [DELTA][R.sup.2] = .02-.03, F(1, 137) = 1.15-9.66, p = .002-.29; PC and COH: 1. step: [R.sup.2] = = .42-.49, F(5, 138) = 20.00-26.00, p = .001; 2. step: [DELTA][R.sup.2] = .01-.04 F(1, 137) = 2.94-9.75, p = .002-.09; PC and SAT: 1. 2 step: [R.sup.2] = .42-.47, F(5, 138) = 19.54-24.67, p = .001; 2. step: [DELTA][R.sup.2] = .001, F(1, 137) = .01-1.78, p = .18-.93; NC and CON: 1. step: [R.sup.2] = .42-.49, F(5, 138) = 19.64-25.96, p = .001; 2. step: [DELTA][R.sup.2] .01-.05, F(1, 137) = 2.92-13.53, p = .001-.09; NC and COH: 1. step: [R.sup.2] = .42-.48, F(5, 138) = 20.33-25.29, p = .001; 2. step: [DELTA][R.sup.2] = = .001-.01, F(1, 137) = .37-2.54, p = .11-.55; NC and SAT: 1. step: [R.sup.2] = .42-.46, F(5, 138) = 19.65-23.54, 2 p = .001; 2. step: [DELTA][R.sup.2] = .001, F(l, 137) = .08-.11, p = .36-.78. # p<.10. * p<.05. ** p<.01.
|Printer friendly Cite/link Email Feedback|
|Author:||Scholz, Urte; Goldammer, Philippe; Hornung, Rainer; Berli, Corina; Luscher, Janina; Knoll, Nina|
|Publication:||Families, Systems & Health|
|Date:||Dec 1, 2013|
|Previous Article:||Provider practices in the Primary Care Behavioral Health (PCBH) model: an initial examination in the Veterans Health Administration and United States...|
|Next Article:||Similarity, agreement, and assumed similarity in proxy end-of-life decision making.|