2.1 Participants
Nine hundred and eight-nine students with 10th grade were recruited from the Chenzhou High School in China. At baseline (T1), 132 students refused to join the survey. Thus, the initial sample at T1 was N = 857. The mean age was 15.14 (±0.55) years, with females comprising 51.6% (N=442) of the sample. Eligible participants at one year after (T2) were students who had participated at T1. Two hundred and eighty-two students declined the second survey and 32 students transferred to other high school. Thus, 543 participants (63.36%) completed the survey at both T1 and at T2, with female students comprising 74.6% (N=405) of the sample. The age distribution of the sample was as follows: 14 years (N=49, 9%), 15 years (N=371, 68.3%) and 16 years (N=123, 22.7%).
2.2 Measures
Trait social anhedonia
The Chinese adolescent version of the Anticipatory and Consummatory Interpersonal Pleasure Scale (ACIPS-A) is a trait anhedonia measure that assesses individual differences in one’s capacity to enjoy interpersonal interactions [17]. The measure is rated on a 4-point Likert scale 0 (totally false for me) to 4 (totally true for me), with lower scores indicating a greater likelihood of social anhedonia. The Cronbach's alpha coefficients in the current sample were 0.89 and 0.92, respectively, at baseline and at follow-up.
State anhedonia
The Chinese version of the Snaith-Hamilton Pleasure Scale is a state anhedonia questionnaire in response to pleasant experiences that span sensory stimuli, social activities, and hobbies in the last few days[25]. Responses to each item is rated on a 4-point Likert scale 1 (strongly Agree) to 4 (strongly Disagree), with a higher total score indicates higher levels of anhedonia. The SHAPS has been validated in adolescents [19]. The SHAPS had an α of 0.90 and 0.93, respectively, at baseline and at follow-up in the present study.
Depressive syndromes
The Chinese version of the Child Mood and Feelings Questionnaire (MFQ-C) is used to measure self-reported experiences of depression [26]. The MFQ-C contains 33 items with five factors: low self-evaluation, suicide ideation, and slow thinking/activity, anhedonia and appetite/sleeping. Anhedonia and suicidal ideation items were removed (see below). The measure is rated on a 3-point Likert scale 0 (not true) to 2 (true), with lower scores indicating a greater depressive symptom. The MFQ-25 had excellent internal consistency in the present study, Cronbach’s alpha = 0.95 and 0.96, respectively, at baseline and at follow-up.
Depressive anhedonia
Depressive anhedonia was rated using the MFQ anhedonia subscale (MFQ-Anhedonia) [26]. Four items of the MFQ-Anhedonia were assessed in the present study: Loss of happiness (felt miserable or unhappy); Loss of enjoyment (I didn’t enjoy anything at all); Loss of interest in friends (I didn’t want to see my friends); Loss of school interest (I didn’t have any fun in school). Other authors have suggested that recent changes in anhedonia are more accurately captured by the Beck Depression Inventory (BDI) anhedonia subscale in adult samples [13, 21]. The Cronbach’s alpha for this four-item scale was 0.72 at baseline and 0.77 at baseline in the present study.
Suicidal ideation (SI)
Suicidal ideation was rated using the MFQ suicide subscale (MFQ-SI) by computing a total score from the following items: I thought that life wasn’t worth living; I thought about death or dying; I thought my family would be better off without me; I thought about killing myself, with each item rated 0 ‘not true’, 1 ‘sometimes’ or 2 ‘true’ (total range 0–8). A previous study has found the MFQ-SI to have high internal consistency and reasonable stability over time [27]. The Cronbach’s alpha of the scale was 0.84 at baseline and 0.86 at follow-up in the present study.
Anxiety syndromes
The Chinese version of the Screen for Child Anxiety Related Emotional Disorders (SCARED) is used to screen for signs of anxiety disorders in children [28]. The SCARED is a 41-item inventory with five factors: generalized anxiety, separation anxiety, social phobia, school phobia and somatic anxiety. The measure is rated on a 3-point Likert scale 0 (not true) to 2 (very true), with lower scores indicating a greater a higher risk of anxiety disorders. In the current study, the Cronbach’s alpha was 0.88 at baseline and 0.89 at follow-up.
Schizotypal personality
The Chinese version of the Schizotypal Personality Questionnaire-Brief (SPQ-B) is used to measure schizotypal personality disorder [29]. The SPQ-B is a 22-itemself report inventory with three subscales: positive symptoms (SPQ-P: Cognitive-Perceptual), negative symptoms (SPQ-N: Interpersonal deficits) and disorganization (SPQ-D: disorganized), rated either “yes” or “no.” The SPQ-B had good validity and reliability in adolescents [30]. In the current study, the three symptom subscales had excellent internal consistency (baseline: α = 0.88, 0.87, 0.81; follow-up: α = 0.87; 0.89, 0.82).
Stressful life events
The Chinese Version of the Adolescent Self-Rating Life Events Check List (ASLEC) was used to evaluate stressful life events experienced during the past 12 months [31]. Twenty-seven stressful life events addressing the following six domains: interpersonal stressful events, academic stressful events, punishment stressful events, bereavement, adaptationstressful events and unscaled items. The ALSEC was rated using a 6-point Likert scale 1 (not at all) to 6 (very severe), with higher total scores indicating a greater number of stressful life events experienced. The Cronbach’s α of the ASLEC was 0.91 at baseline and 0.89 at follow-up in the present study.
2.3 Procedure
The research was approved by the ethics committee of the author’s university. Participants took 45 minutes to complete the paper-and-pencil survey during a regularly scheduled class period. For all participants under the age of 18, parents and guardians were asked to provide written informed consent. Researchers informed students that their responses would be kept confidential and would not be shared with their parents and school staff. Students were not individually compensated for participation but were given small incentives (e.g., sweets).
2.4 Analyses
First, to examine the potential influence of missing data, those with complete data at both time points (N=543; 66.5%) and those with missing data (N=314; 33.5%) were compared. These comparisons revealed no significant differences among important variables (all ps>.05). After data cleaning, absolute stability of symptoms was tested using paired student's t-tests. Relative stabilities (test-retest) of the anhedonia measures were tested using intraclass correlation coefficients. Cohen's d provided an estimate of the effect size of the difference. Next, Pearson’s correlations were used to examine associations between the three types of anhedonia and other self-report symptoms at baseline. In order to control the family-wise error rate when making multiple comparisons, Bonferroni-correction p-values were used (adj p-value of .001). Gender differences in anhedonia at both time points were included in the repeated measures analysis of variance. General linear regression models were used to assess anhedonia and psychopathology symptoms predicting each other while adjusting for baseline scores. Two reciprocal associations were examined: (a) between anhedonia at baseline and psychopathology symptoms at termination, (b) between the changes in anhedonia from baseline to termination and psychopathology symptoms at termination.