INTRODUCTION 1 3 4 5 6 7 8 9 10 11 METHOD 12 14 et al 8 8 r Z 15 Z r RESULTS 3 16 17 18 23 24 24 33 26 Table 1 Table 1 Quantitative studies into the determinants of using ecstasy and measures used. No. Sample details n Age % ♀ Relevant variables I R [24] UK, before July 2002 Polydrug users 364 19 44% Negative mood function scale 3 5      Social function scale 5 5      Negative effects 4 5      Other functions (9 functions) 1 5      Extent of peer use 1 4      Partner/best friend use 1 2      Intensity of use 1 – [25] UK, before March 1998 Alcohol and drug users 100 19 45% Mood function scale 3 5      Social/contextual function scale 5 5      Negative effects/events scale 3 5      Extent of peer use 1 5      Intention 1 7      Intensity of use 1 –      Frequency of use 1 – [26a] UK, March 1992 Students 186 19–25 58% Attitudes 6 7      Subjective norms 1 7      Perceived behavioural control 6 7      Behavioural beliefs (17) 2 19      Normative beliefs (5) 2 19      Control beliefs (8) 2 19      Intention to use ecstasy 4 7 * UK, mid-1996 Club members t1: 203 t2: 123 23 41% Attitudes 8 7      Normative influences 13 7      Perceived behavioural control 1 7      Self-efficacy 1 7      Behavioural beliefs 13 7      Control beliefs 6 7      Intention 3 7      Behaviour (longitudinal) 4 – [27] Netherlands, 2000, 2001 Party attendants 844 22 33% Negative outcome expectancies 11 2      Enhancement outcome expectancies 3 2      Euphoria outcome expectancies 4 2      Sex outcome expectancies 4 2      Dancing outcome expectancies 3 2      Insight outcome expectancies 4 2      Communication outcome expectancies 4 2      Whether ecstasy was currently used 1 2 [28] * UK t1: October 1994–1995 t2: May1995–1996 Students t1: 461 t2: 136 19–22 19–22 55% 65% Attitude 2 7      Injunctive norms 6 7      Perceived behavioural control 8 7      Descriptive norms 2 6      Moral norm 1 7      Intention 1 8      Behaviour (longitudinal) 1 8 [29] UK, before 2003 College students 657 19 55% Frequency of past use 1 7      Intentions to use 1 9      Normative influence (friends’ use) 1 6      Beliefs about ecstasy use (7 beliefs) 1 5 [30] * UK t1: 84 t2: 32 20 74% Attitude 10 8      Subjective norm 2 8      Perceived behavioural control over obtaining ecstasy 3 8      Perceived behavioural control over taking ecstasy 4 8      Intention 5 8      Habit 2 8      Specific attitudinal beliefs (13 beliefs) 1 5      Behaviour (longitudinal) 1 2 [31] the Netherlands, 2001–2002 490 22 34% Energy motives 4 5      Euphoria motives 3 5      Self-insight motives 2 5      Sociability/flirtatiousness motives 8 5      Sexiness motives 4 5      Coping motives 3 5      Conformism motives 4 5      Perceived positive effects 24 2      Perceived negative effects 11 2      Perceived friends’ use 1 5      Frequency of ecstasy use 1 5 [32] UK 200 21 66% Attitude 5 ?      Subjective norm 5 5      Perceived behavioural control over obtaining ecstasy 3 7      Perceived behavioural control over taking ecstasy 11 ?      Intention 6 7      Habit 2 7 [33] USA Club rave attendees 70 20 47% Risk associated with using ecstasy once or twice 1 4      Risk associated with using ecstasy regularly 1 4      Harmful short-term physical effects 1 4      Harmful long-term physical effects 1 4      Harmful short-term psychological effects 1 4      Harmful long-term psychological effects 1 4      Positive physical effects 1 4      Positive psychological effects 1 4      Ecstasy use within the past 12 months 1 2 No. = number in reference list, I = number of items used to measure variable, R = number of scale points on response scale of each item. * Longitudinal design. 34 28 30 32 35 36 37 38 39 40 24 25 27 31 33 41 24 25 27 31 33 12 35 In both the TPB and outcome expectancy models, higher-level constructs, such as attitudes, are based on lower-level beliefs. Most studies tested models involving this hierarchical cognitive structure. For the purposes of this review, higher-level constructs are referred to as ‘compound constructs’ and lower-order cognitions as ‘expectancies’ (e.g. beliefs about probable consequences of ecstasy use). Details of the particular theoretical models explored in the original studies are not provided here. R 2 28 R 2 28 30 32 R 2 24 27 31 25 r 28 30 Table 2 12 35 Table 2 Table 2 Effect sizes of predictors of ecstasy use and intention to use.   Association with behaviour Association with intention     Variable type Compound or expectancy k + k + Attitude Attitude [C] 5 0.53 (L) 5 0.63 (L) Positive expectancies Social function [C] 4 0.09 2 0.30 (M)  Produces a positive mood state 3 0.12 (S) 2 0.28 (M)  Mood function [C] 4 0.20 (S) 1 0.41 (L)  Helps lose weight 2 0.09 1 0.08  Enhances sex 3 0.11 (S)    Helps stay awake 3 0.08    Helps keep fit 1 0.14 (S) 1 0.22 (S)  Enhances insight/openness 2 0.18 (S)    Helps to relax/helps coping 2 0.08    Produces excitement   1 0.31 (M)  Produces intoxication 1 0.19 (S)    Eases after-effects 1 0.15 (S)    Improves other drugs' effects 1 0.11 (S)    Is conforming to peers 1 0.05    Helps work 1 0.05   Negative expectancies Short-term negative effects [C] 4 −0.47 (L)    Produces mood swings 1 −0.34 (M) 2 −0.31 (M)  Leads on to more frequent use 1 −0.30 (M) 2 −0.39 (L)  Produces mental side-effects 1 −0.24 (M) 2 −0.40 (L)  Leads on to ‘worse’ drugs 1 −0.23 (S) 2 −0.31 (M)  Produces physical side-effects 1 −0.18 (S) 2 −0.39 (L)  Makes one unhealthy 1 −0.18 (S) 2 −0.30 (M)  Produces depression 1 −0.18 (S) 2 −0.28 (M)  Produces lethargy 1 −0.05 2 −0.20 (S)  Produces addiction 1 −0.32 (M) 1 −0.12 (S)  Leads to death 1 −0.30 (M) 1 −0.32 (M)  Produces paranoia 1 −0.27 (M) 1 −0.26 (M)  Makes one feel run down 1 −0.17 (S) 1 −0.25 (M) Subjective injunctive norm Subjective norms (approval) [C] 5 0.39 (L) 5 0.52 (L)  Best friends' approval 1 0.36 (M) 1 0.25 (M)  Partners approval 1 0.15 (S) 1 0.30 (M)  Parents' approval 1 0.10 (S) 1 0.01  Health experts' approval 1 0.09 1 0.00  Other ecstasy users' approval 1 0.07 1 0.08 Subjective descriptive norm Subjective norms (descriptive) [C] 1 0.52 (L) 1 0.63 (L)  Perceived peer use 2 0.39 (L) 1 0.43 (L)  Perceived use in close relatives 2 0.21 (S)    Perceived best friend/partner use 1 0.37 (L)   Perceived behavioural control PBC [C] 3 0.33 (M) 3 0.52 (L)  PBC (over obtaining ecstasy) 2 0.20 (M) 2 0.25 (M)  PBC (over taking ecstasy) 2 0.01 2 −0.03  Being with friends who use 1 0.45 (L) 2 0.62 (L)  Going out dancing 1 0.41 (L) 2 0.67 (L)  Being offered ecstasy 1 0.40 (L) 2 0.59 (L)  Ecstasy being available 1 0.40 (L) 2 0.57 (L)  Cheap ecstasy 1 0.19 (M) 2 0.54 (L)  Having alcohol 1 −0.02 2 0.10 (S)  not 1 −0.06 1 0.07  Needing to lose weight 1 −0.05 1 0.00  Needing to exercise 1 −0.01 1 0.05 Miscellaneous Habit [C] 2 0.45 (L) 2 0.46 (L)  Moral norm [C] 2 −0.28 (M) 2 −0.31 (M)  Denial of negative consequences 1 0.17 (S) 1 0.18 (S)  Anticipated regret 1 −0.11 (S) 1 −0.22 (S) k r 42 Table 2 24 25 28 30 32 26 31 27 t r t 2 t 2 29 33 d 42 r d r r r r r r r r r r r r r 29 29 33 t t P Table 3 Table 3 29 33 Variable Study Significance Association Effect size Perceived use by friends 29 < 0.001 Positive M Anticipated regret (‘use would induce guilt’) 29 < 0.001 Negative M Is hard to resist 29 < 0.001 Positive M Moral norm (‘ecstasy use is immoral’) 29 < 0.05 Negative S Perceived availability of ecstasy 29 < 0.05 Positive S Is bad for one's physical health 29 None – – Is bad for one's mental health 29 None – – Subjective norm (approval) 29 None – – Harmful long-term physical effects 33 < 0.01 Negative M Risk associated with using regularly 33 < 0.01 Negative M Risk association with using once or twice 33 < 0.05 Negative M Harmful long-term psychological effects 33 < 0.05 Negative M Harmful short-term physical effects 33 None – – Harmful short-term psychological effects 33 None – – Positive physical effects 33 None – – Positive psychological effects 33 None – – 42 These additional results confirm the relevance of descriptive norm, negative expectancies (particularly long-term effects) and perceived control, and add anticipated regret, with at least a medium effect size, to the list. DISCUSSION Synthesis of the included studies shows the main predictors of intention to use and actual ecstasy use to be attitude (specifically positive outcomes regarding mood control and social facilitation and negative outcomes regarding escalating use and physical and mental side-effects); subjective and descriptive norms regarding one's friends, partner and peers; perceived control regarding obtaining ecstasy and control in relation to being with friends who use, going out dancing, being offered ecstasy and ecstasy being available; and habit, moral norm and anticipated regret. In addition to lending support to both the TPB and expectancy models, these findings show that some expectancies underlying attitude are irrelevant (e.g. ecstasy enhances sex), as are norms relating to some social referents (e.g. parents), and perceived control regarding some ecstasy-related behaviours (e.g. taking ecstasy). 43 6 44 45 46 47 48 49 35 50 8 51 39 et al 8 52 53 Regarding the theories that have been studied, in order to gain a more comprehensive understanding of motives for ecstasy use future research should combine the two theoretical perspectives studied thus far, so that relative overlap can be determined. It would be interesting to see whether, and to what degree, particular expectancies account for the predictive utility of the TPB attitude measures. Also, the list of expectancies that has been studied so far may omit a number of consequences (such as ‘suicide Tuesday’, a term for a period following use when low serotonin levels can induce depressive feelings). 12 54 In conclusion, this review suggests that there is sufficient evidence to guide intervention development so that evidence-based practice is established. These interventions could then be evaluated to test the utility of particular theoretical frameworks. The priorities for interventions should be negative expectancies, perceived behavioural control and anticipated regret. Tailored interventions can offer refusal skills training and strategies to avoid risky situations to participants not intending to use ecstasy. Appendix I Search terms used in PsycINFO (equivalent terms used in corresponding fields in MedLine and ERIC) [query in words, as corresponding to ‘concepts’ column, in brackets]. No. Concept Operationalization Fields 1 Language (English) or (Dutch) Language 2 Publication type (journal*) or (peer-reviewed-journal) Publication type 3 Publication date > 1980 Publication year 4 Ecstasy (clubdrug*) or (club near drug*) or (dance near drug*) or (dancedrug*) or (party near drug*) or (partydrug*) or (xtc) or (mdma) or (methylenedioxymethamphetamine) or (‘3,4-methylenedioxymethamphetamine’) or (ecstasy) Title, abstract, keywords 5 Theoretical (theor*) or (attitud*) or (motivat* near functio*) or (mode*) or (norm*) or (perceived near control) or (pbc) or ('social cognitive’) or (self adj efficacy) or (stages near change) or (perceived adj (harm or risk or functions)) or (functional) or (outcome adj (expectancies or expectations)) or (sct) or (tpb) or (patter*) or (psychosoc*) or (health adj belief adj model) or (hbm) Title, abstract, keywords 6 Determinants (determin*) or (facto*) or (variabl*) or (parameter*) or (reason*) or (caus*) or (motiv*) or (incentive*) or (correlat*) or (antecedent*) or (character*) Title, abstract, keywords 7 Initiation (start*) or (commenc*) or (originat*) or (onset) or (initiat*) or (instigat*) or ((use) not (user)) or (using) or (usage) or (establish*) Title, abstract, keywords 8 Maintenance (maint*) or (sustain*) or (continu*) or (uphold*) or (persist*) or (further*) or (prolong*) Title, abstract, keywords 9 Cessation (end*) or (stop*) or (discontinu*) or (terminat*) or (ceas*) or (cessat*) or (abstain*) or (abstin*) or (quit*) or (remiss*) or (resolut*) or (recover*) Title, abstract, keywords 10 Harm reduction (harm or risk or damage or casualt*) and (reduc* or manag* or limit* or minimi*) Title, abstract, keywords 11 Excluded ((treatment not (‘not in treatment’ or ‘non-treatment’ or ‘non- treatment’ or ‘no treatment’)) or rat or rats or mouse or mice or animal or monkey* or pigeon* or spectro* or cardio* or seroton* or dopamin* or neurotransm* or receptor* or psychiatr* or psychopath* or cell* or diagnos*) Anywhere 12 Inclusion #1 and #2 and #3 [Language and Publication Type and Publication Date] – 13 Behaviour #7 or #8 or #9 or #10 [Initiation or Maintenance or Cessation or Harm reduction] – 14 Empirical* #6 near #13 [Determinants near Behaviour] – 15 Final query* #12 and #4 near (#5 or #14) not #11 [Inclusion and Ecstasy near (Theoretical or Empirical) not Excluded] – When executed, the query consisted of one command; therefore the use of the ‘near’-operator was valid here. Appendix II Search procedure, number of resulting hits, and results of each step. Step Activity Number of resulting publications 1 Input of query at 20 August 2007 in PsycINFO (162), MedLine (194) and ERIC (11) 367 2 Removal of duplicate records (83) 284 3 Removal of records about publications that (entries were removed in this order):  3.1 studied biological variables (e.g. sequelae of ecstasy use; 75) 209 3.2 did not study ecstasy use or a related behaviour (such as trying out ecstasy, ceasing use, changing use patterns, or applying harm reduction practices; 32) 177 3.3 studied variables that cannot be changed using a health promotion intervention (e.g. sex, ethnicity or religion; 68) 109 3.4 studied ecstasy use as an independent variable in a multivariate or longitudinal analysis (22) 87 3.5 did not employ quantitative methods (e.g. qualitative studies; 30) 57 3.6 did not study behaviour or cognitions (25) 32 3.7 * † 18 3.8 were not published in a peer-reviewed journal (3) 15 * 12 55 † 56