افزایش اطلاعات و تکنیک های تعیین هدف برای افزایش انگیزش تطبیقی و کاهش میل به نوشیدن الکل

Abstract Objective The aim of the study was to determine whether experimental manipulation of sense of control would change moderate drinkers' (N = 106) task-specific motivational structure and explicit and implicit determinants of their urge to drink alcohol. Method The effects of various levels of information-enhancement and goal-setting on participants' performance on experimental tasks were assessed. Participants were randomly assigned to a high-sense-of-control, low-sense-of-control, or no-intervention group. Dependent measures were indices derived from a task-specific version of the Personal Concerns Inventory and the Shapiro Control Inventory, Alcohol Urge Questionnaire, and alcohol Stroop test.

1. Introduction Human beings are goal strivers. They try to achieve things that they want and to get rid of things that they do not want. They decide how and when to pursue particular goals or to give up doing so. People's chances of success in achieving their goals depend on the pattern of their goal strivings; this pattern is called motivational structure. Motivational structure varies from one person to another; it is the more-or-less stable way in which each person pursues his or her goals. However, motivational structure is not entirely rigid because people's current concerns and their success with or failure at goal pursuits can modify the way in which they strive for goals in the future. To measure motivational structure, Cox and Klinger (2011a) and Klinger and Cox (2011b) developed the Motivational Structure Questionnaire (MSQ) and the Personal Concerns Inventory (PCI). A person's motivational structure can be adaptive or maladaptive. Research using the MSQ and PCI (e.g., Cox and Klinger, 2002, Cox et al., 2002, Fadardi and Cox, 2008 and Klinger and Cox, 2011b) has indicated that compared to people with an adaptive motivational structure, people with a maladaptive motivational structure have (a) fewer positive goals, (b) less hope for achieving their goals, (c) less anticipated happiness from achieving their goals and less anticipated sorrow from not achieving them, (d) longer expected distances from reaching goals, (e) less feeling of commitment to their goals, and (f) less perceived personal control over achieving their goals. As Klinger and Cox (2011a) suggest, people's sense of control should be an important component of their motivational structure. Having a sense of control is essential for a human's functioning. People need to have a sense of control over their lives because it provides them with opportunities to better recognize and organize their resources for achieving their goals (Mirowsky, 1995 and Mirowsky, 1997). Perceived control refers to people's perceptions that events in their lives are controlled by their own choices and actions rather than external factors, such as chance, luck, or fate, or by the authority and actions of other people. Some people feel little control over difficult situations or over preventing bad things from happening; they might also believe that good things that happen to them are due only to luck (Mirowsky & Ross, 1990). People's perceptions of choice, chance, and authority (i.e., the factors that determine their sense of control) vary from individual to individual, and from situation to situation. Perceptions of control are related to beliefs, emotions, and behaviors, and how people respond to both aversive and positive events. A sense of control is acquired by having actual control over desirable and undesirable outcomes (Shapiro, 1994). Feeling a lack of control reduces individuals' efforts and impairs their ability to succeed in achieving their goals. Part of the problem might arise because these people do not have a plan, or cannot think of one, about how to achieve their goals or complete their tasks (Shapiro, 1994). Perceived lack of control could then adversely affect the quality of their lives and have health-damaging consequences (Seligman, 1990). When people frequently confront uncontrollable events or experience failure repeatedly, they might develop feelings of helplessness and depression (Abramson et al., 1978 and Seligman, 1994). In fact, there is a general consensus (e.g., Bandura, 1990, Shapiro and Bates, 1990 and Taylor and Brown, 1988) that perceived control can often be as important as having actual control. If individuals believe that they have some degree of control, they are more likely to take action, even if there is no guarantee they will succeed (Lachman & Weaver, 1998). To summarize, sense of control is an important psychological construct. Perceived sense of control not only influences individuals' inner worlds and their relationship with themselves; it also determines the quality of their social relationships and their physical and mental health. The construct motivational structure is important for understanding goal-directed behavior. Previous studies (e.g., Cox and Klinger, 2002 and Fadardi and Cox, 2008) have shown that an adaptive motivational structure is associated with greater expected chances of achieving one's goals than a maladaptive motivational structure. It has also been shown that people's adaptive motivational structure is inversely related to their alcohol consumption ( Cox and Klinger, 2002 and Shamloo and Cox, 2010). In addition, Shamloo and Cox (2010) found that (a) sense of control was an important component of motivational structure, and (b) having more sense of control was associated with having stronger adaptive motivation and habitually consuming less alcohol. Their results also indicated that motivational structure mediated the effects of sense of control on participants' alcohol consumption. However, Shamloo and Cox's (2010) study did not provide definitive evidence that these relationships were causal, although causality is an important question to address empirically. Demonstrating cause-and-effect relationships would have both theoretical and practical significance. It might, for instance, pave the way for developing interventions for improving people's motivational structure and sense of control. The current study used novel techniques to experimentally manipulate the effects of sense of control on participants' motivational structure and, in turn, on implicit and explicit indicants of their desire to drink alcohol. It was hypothesized that (a) experimental induction of a high sense of control would increase participants' perceptions of control; (b) experimental induction of a low sense of control would reduce participants' perceptions of control; (c) the increase in sense of control would be associated with an increase in adaptive motivation and a reduction in the urge to drink alcohol and alcohol-related attentional bias; (d) post-experimentally, on adaptive motivation the groups would be ordered: High-Sense-of-Control Group > No-Intervention Group > Low-Sense-of-Control Group; and (d) post-experimentally participants' urge to drink alcohol and their alcohol attentional bias would be ordered: Low-Sense-of-Control Group > No-Intervention Group > High-Sense-of-Control Group.

3. Results Participants (N = 106, 48.1% males) were randomly assigned to the No-Intervention Group (N = 35, 54.3% males), Low Sense of Control Group (N = 36, 44.4% males), or High Sense of Control Group (N = 35, 45.7% males). One-way ANOVAs showed that there were no significant differences among the groups on age [F(2, 103) = .83, p = .44] or years of education [F(2, 103) = .30, p = .74]. A Kruskal–Wallis nonparametric test showed that the groups also did not differ on gender, X2 (2) = .80, p = .67. 3.1. Preliminary analyses 3.1.1. Factor analysis and summary index of TSPCI To facilitate data analysis and for ease of interpretation, the PCI indices were subjected to principal components analysis (PCA). The appropriateness of the PCI results for PCA was checked against Preacher and MacCallum's (2002) guidelines. Bartlett's test of sphericity was significant [X2(55) = 455.90, p < .0001], and the value of KMO was .73. Hence, a PCA with a two-factor solution was conducted on the mean indices for the rating scales. Factor 1 and Factor 2 explained 34.71% and 15.08% of the variance, respectively; together the two factors accounted for 49.76% of the variance. The results of a factor analysis are specific to the particular sample on which the data were collected. For example, if a researcher administered the PCI to a given sample and then did so a second time (e.g., conducted a post-test) and factor analyzed both sets of data, the results of the two factor analyses would likely not be the same (unless in the unlikely case that all respondents gave identical answers on the two occasions). Moreover, factor scores for a group of participants are relative to each other; each participant's factor score depends on all other participants' scores. For this reason, a particular respondent might have answered the PCI exactly the same on two occasions, yet that participant's factor scores would not remain the same on the two administrations of the test if other participants responded differently on the two occasions. Thus, it would not be legitimate to perform a factor analysis on the two sets of data and then to compare factor scores across the two sets. Accordingly, a TSPCI summary score was developed that made comparisons across two administrations of the test possible. It was based on the TSPCI rating scales that loaded on Factor 1 from the pre-test (see Table 2). The rating scales with negative loadings were summed, and the sum was then subtracted from the sum of the rating scales with positive loadings. The formula was: adaptive motivation = (like the task + control + know what to do + chances if try my best + happiness if succeed + commitment + sorrow from failure) − (dislike the task)] / 8]. This formula was used to calculate TSPCI summary scores for both the pre- and the post-test. Table 2. Factor loadings for the TSPCI indices before and after the experimental manipulation. TSPCI rating scales Anagrams and concept identification Pre-test Post-test Factor 1 Factor 2 Factor 1 Factor 2 Liking the tasks .77 # .82 # Disliking the tasks − .69 # − .82 # Control over success .63 − .40 .80 # Know what to do .60 # .73 # How likely if try best .80 # .83 # How likely if lucky # .40 # .56 Joy if succeed .49 .33 .66 .31 Unhappiness (conflict) # .36 − .41 .69 Sorrow from failure .30 − .49 .55 − .54 Commitment .56 .52 .68 # Goal distance # .57 − .57 .47 Note. TSPCI = Task-Specific Personal Concern Inventory. # = loadings < .30. Table options 3.1.2. Performance on concept-identification cards and anagrams One-way ANOVAs showed a significant main effect for Groups on both number of correctly answered concept-identification cards [F(2, 103) = 38.61, p < .0001] and number of correctly answered anagrams [F(2, 103) = 33.06, p < .0001]. Post-hoc Tukey HSD tests showed that the High-Sense-of-Control Group (M = 4.08, SD = .86) correctly answered significantly more concept-identification cards than either the No-Intervention Group (M = 2.54, SD = .98; p < .0001) or the Low-Sense-of-Control Group (M = 2.23, SD = .99; p < .0001). 3.1.3. Changes in sense of control Three one-way ANOVAs showed that there was no difference among the groups on pre-test Positive Sense of Control [F(2, 103) = .48, p = .62], Negative Sense of Control [F(2, 103) = .28, p = .76], or Overall Sense of Control [F(2, 103) = .55, p = .78]. To further determine whether the groups differed from one another on post-test sense of control, a MANCOVA was conducted. Levene's test of equality of variances was significant, indicating the adequacy of the MANCOVA model. The groups differed on the combined dependent variables [F(4, 198) = 23.70, p < .0001, η2 = .33, Wilk's Lambda = .46; Power = .94]. When the dependent variables were tested separately, there were also main effects for Group on all three sense of control scales: post-test Positive Sense of Control [F(2, 100) = 57.08, p < .0001, η2 = .53], Negative Sense of Control [F(2, 100) = 34.04, p < .0001, η2 = .41], and Overall Sense of Control [F(2, 100) = 56.49, p < .0001, η2 = .53]. Post hoc pairwise comparisons showed that there was a significant difference among the groups on the Sense of Control scales. On Negative Sense of Control, the High-Sense-of-Control Group was lower than the No-Intervention Group, and both of these groups were lower than the Low-Sense-of-Control Group (ps < .001). On Positive and Overall Sense of Control, the High-Sense-of-Control Group was higher than the No-Intervention Group, and both of these groups were higher than the Low-Sense-of-Control Group (ps < .004). Additional t-tests for paired-samples were conducted to assess changes from the pre-test to the post-test on the Sense of Control scales, separately for each of the three groups. The No-Intervention Group did not change from pre- to post-test on any of the scales (p > .05). However, the Low-Sense-of-Control Group showed an increase in Negative Sense of Control from pre-test to post-test [t(35) = − 4.53, p < .0001, d = − .66], but it showed a reduction in Positive Sense of Control [t(35) = 7.76, p < .0001, d = .68] and Overall [t(35) = 7, p < .0001, d = .87]. On the other hand, the High-Sense-of-Control Group showed an increase in Positive Sense of Control [t(34) = − 4.20, p < .0001, d = − .58] and Overall [t(34) = 4.35, p < .0001, d = .59], but a reduction in Negative Sense of Control [t(34) = − 4.67, p < .0001, d = − .71]. 3.1.4. Changes in adaptive motivation A one-way ANOVA showed that the groups did not differ on pre-test adaptive motivation [F(2, 103) = .53, p = .59]. To further test between-group differences on adaptive motivation on the pre-test, a univariate analysis of covariance (ANCOVA) using GLM was conducted. There was no effect for Group on pre-test adaptive motivation (p = .59). This indicates that the three groups did not differ from one another on adaptive motivation prior to the experimental induction. To test whether the groups differed from each other on post-test adaptive motivation, an ANCOVA was performed. The results showed that there was a significant main effect for Group [F(2, 99) = 38.96, p < .0001, η2 = .38] after controlling for the pre-test adaptive motivation [F(1, 99) = 35.35, p < .005, η2 = .26] and the covariates—i.e., familiarity with (a) anagrams [F(1, 99) = .23, p = .63, η2 = .002] and (b) concept-identification cards [F(1, 99) = .36, p = .55, η2 = .004]. Pairwise comparisons revealed significant differences among the groups (ps < .000 for all comparisons); the groups were ordered from highest to lowest on post-test adaptive motivation as follows: High-Sense-of-Control Group > No-Intervention Group > Low-Sense-of-Control Group. Although the groups differed from one another on the summary index of adaptive motivation, it was also worthwhile to identify the particular TSPCI indices on which they differed. First, a series of one-way ANOVAs on the 11 pre-test TSPCI indices showed no difference among the groups. To determine whether the groups differed on the indices at the post-test, a MANCOVA was again conducted. Levene's test of equality of variances was not significant, indicating the adequacy of the MANCOVA model. The groups differed on the combined dependent variables [F(4, 198) = 23.70, p < .0001, η2 = .33, Wilk's Lambda = .46; Power = .94]. When the dependent variables were tested separately, there were also main effects for Group on six of the TSPCI indices: Control [F(2, 92) = 17.88, p < .0001, η2 = .49]; What to do [F(2, 92) = 8.18, p < .0001, η2 = .63]; Try my best [F(2, 92) = 14.28, p < .0001, η2 = .61]; Joy [F(2, 92) = 13.60, p < .0001, η2 = .49]; Sorrow [F(2, 92) = 34.04, p < .0001, η2 = .59]; and Commitment [F(2, 92) = 10.80, p < .0001, η2 = .70]. Moreover, on these six indices, the groups were ordered from highest to lowest as follows: High-Sense-of-Control Group > No-Intervention Group > Low-Sense-of-Control Group. 3.2. Relationship between sense of control and motivational structure A linear regression analysis was conducted to determine whether on the post-test, participants' sense of control was associated with adaptive motivation. Before entering the variables into the regression analysis, a scatter plot was drawn to depict how the independent variables were related to the dependent variable. The fit line indicated that the data were normally distributed. To identify simple relationships among gender, age, adaptive motivation, and Overall Sense of Control, bivariate Pearson correlations were computed (see Table 3). As Table 3 shows, adaptive motivation was positively correlated with Overall Sense of Control. Table 3. Intercorrelations among gender, age, post-test adaptive motivation, positive sense of control, negative sense of control, and overall sense of control. Variables Gender Age Adaptive motivation Overall SoC Positive SoC Age .02 Adaptive motivation − .11 .04 Overall SoC − .10 .11 .59⁎⁎ Positive SoC − .080 .061 .57⁎⁎ .94⁎⁎ Negative Soc .068 .036 − .45⁎⁎ − .89⁎⁎ − .76⁎⁎ Note. Positive SoC = Positive Sense of Control; Negative SoC = Negative Sense of Control; Overall SoC = Overall Sense of Control. ⁎⁎ p < .01. Table options Next, hierarchical regression analyses were conducted to determine the relative contributions of age, gender, and Positive, Negative, and Overall Sense of Control to adaptive motivational structure. The results are shown in Table 4. The hierarchical regression analysis showed that for all models in Step 1, gender and age were not significant predictors of adaptive motivation [F(2, 103) = .47, p = .79]. Adding each of the predictor variables in the regression models (i.e., the second step) led to significant increments in the variance accounted for in the model. That is, after the effects of gender and age had been controlled in Step 1, post-test Overall Sense of Control explained 31% of the variance in adaptive motivation [F(1, 102) = 15.34, t = 6. 73, p < .0001]; post-test Positive Sense of Control explained 32% of the variance in adaptive motivation [F(1, 102) = 15.34, t = 7.01, p < .0001]; and post-test Negative Sense of Control explained 20% of the variance in adaptive motivation [F(1, 102) = 8.67, t = − 5.04, p < .0001]. In summary, the results showed that Sense of Control predicted adaptive motivation independently of age and gender. Table 4. Results of a hierarchical regression analysis in which adaptive motivation was predicted from gender, age, and overall sense of control. Model Steps Variables B SE B ΔR2 β t Overall Soc 1 Gender − .13 .19 − .067 − .68 Age .003 .024 .005 .013 .13 2 Overall SoC .05 .00 .31 .55⁎ 6.73 Positive Soc 1 Gender − .13 .19 − .067 − .68 Age .003 .024 .005 .013 .13 2 Positive SoC .061 .009 .32 .57⁎⁎ 7.01 Negative Soc 1 Gender − .13 .19 − .067 − .68 Age .003 .024 .005 .013 .13 2 Negative SoC − .069 .014 .20 − .44⁎⁎ − 5.04 ⁎ p < .05. ⁎⁎ p < .001. Table options 3.2.1. Alcohol-related measures 3.2.1.1. Urge to drink A one-way ANOVA showed no difference among the groups on urge to drink at the pre-test [F(2, 98) = .68, p = .51]. To determine whether the groups differed from each other on post-test urge to drink, an ANCOVA was conducted. The results showed that after controlling for pre-test urge to drink [F(1, 97) = 100.01, p < .0001, η2 = .51], there was a significant main effect for Group [F(2, 97) = 19.17, p < .0001, η2 = .28]. Pairwise comparisons revealed that the Low-Sense-of-Control Group was higher than both the No-Intervention Group and the High-Sense-of-Control Group (ps < .0001). Paired-sample t-tests were also conducted to test whether the groups' urge to drink changed from the pre- to the post-test. The results were as follows: (a) the Low-Sense-of-Control Group increased [t(32) = − 4.13, p < .0001, d = − .36]; (b) the High-Sense-of-Control Group decreased [t(34) = 3.05, p = .004, d = .27]; and (c) the No-Intervention Group did not change [t(32) = .27, p = .79, d = .02]. 3.2.1.2. Alcohol Stroop test A one-way analysis of variance revealed that there were no differences among the three groups in the number of errors made across each category of words (i.e., alcohol and neutral) (p > .05). Further analysis of the errors was unnecessary because the number of errors was negligible. To identify whether there were differences among the groups' reaction times (RTs) on the alcohol-related and neutral stimuli, a one-way ANOVA was performed. There was a main effect for Group on the alcohol-related stimuli [F(2, 98) = 7.91, p = .001], with the Low-Sense-of-Control Group taking longer to respond to these stimuli than either the No-Intervention Group (p < .011) or the High-Sense-of-Control Group (p < .001). There was no main effect for Group on the neutral stimuli [F(2, 98) = 2.18, p = .12]. Alcohol-interference scores were calculated by subtracting each participant's mean RT to the neutral stimuli from his or her mean RT to the alcohol-related stimuli. To test whether the groups differed on the alcohol-interference scores, a one-way ANOVA was performed. The results showed a main effect for Group [F(2, 98) = 6.23, p < .003]. Post hoc Tukey HSD tests showed that the Low Sense of Control Group had larger alcohol-interference scores than both the No Intervention Group (p < .013) and the High Sense of Control Group (p < .005); however, the latter two groups did not differ from each other (p = .17).