Cognitive Therapy and Research

, Volume 33, Issue 5, pp 449–461

Exploring the Efficacy of Cognitive Bibliotherapy and a Potential Mechanism of Change in the Treatment of Depressive Symptoms Among the Chinese: A Randomized Controlled Trial


    • Department of Clinical Psychology, College of MedicineFu Jen Catholic University
  • Wan-Lan Chen
    • Department of Human DevelopmentTzu-Chi University
  • Yi-Hwei Li
    • Department of Public HealthTzu-Chi University
  • Chiao Han Wang
    • Department of Clinical Psychology, College of MedicineFu Jen Catholic University
  • Tze Jing Mok
    • Department of Clinical Psychology, College of MedicineFu Jen Catholic University
  • Hwei Shan Huang
    • Department of Clinical Psychology, College of MedicineFu Jen Catholic University
Original Article

DOI: 10.1007/s10608-008-9228-4

Cite this article as:
Liu, E.T., Chen, W., Li, Y. et al. Cogn Ther Res (2009) 33: 449. doi:10.1007/s10608-008-9228-4


The present study investigated the efficacy of cognitive bibliotherapy in the treatment of depressive symptoms among Chinese individuals in Taiwan. Adults with depressive symptoms (N = 52, M age = 26.4) were randomly assigned to the treatment condition or the delayed treatment control condition. Participants were assessed at pretreatment, posttreatment, and 3-month follow-up. Results indicated that participants’ overall depression level lowered at posttreatment. Analyses were performed on the intention-to-treat basis. Multiple imputation inference procedure (Rubin in Multiple imputation for nonresponse in surveys, John Wiley & Sons, Inc., New York 1987) was adopted to estimate missing values and to draw inferences based on the imputed data. Results of the analyses indicated that the cognitive-affective symptoms of depression, rather than the somatic symptoms of depression, evidenced significant reduction as a result of cognitive bibliotherapy. Further reductions in cognitive-affective symptoms were observed at 3-month follow-up. Lastly, learned resourcefulness was found to be a mechanism through which bibliotherapy reduced depressive symptoms. The present study provides preliminary evidence that cognitive bibliotherapy may be a promising treatment option for Chinese individuals with depressive symptoms. In the meantime, participants’ qualitative feedback may provide important direction for cross-cultural adaptation of cognitive bibliotherapy. Applied implications and cultural issues are discussed.


ChineseBibliotherapyDepressionLearned resourcefulnessMediator


Depression has been projected to be the second leading cause of disability by the year 2020 (Greenberg et al. 2003). With its high prevalence (5–30%) (Myers et al. 1984; Wells 1985), recurrence rate (80–90%) (Ernst et al. 1992), chronicity rate (11–25%) (Angst and Preisig 1995; Angst et al. 1996), and association with suicidality (Chioqueta and Stiles 2003; Kessler et al. 1999; Olgiati et al. 2006), depression has been found to be a debilitating health problem that causes enormous social and economic burden worldwide (Chien et al. 2007; Greenberg et al. 2003; Hu 2004; Thomas and Morris 2003; WHO 1999). In Taiwan, the annual cost for the treatment of depression is approximately US$ 116.6 million (1.2% of total national expenses), and the cost continues to increase in recent years (Chan et al. 2006).

Despite the increasing costs for the treatment of depression in Taiwan, research suggests that individuals with depression are undertreated, and that the treatment rate in Taiwan is only approximately 2.3% (Chan et al. 2006). Largely due to the stigma associated with mental disorders (Chien et al. 2007; Gaw 1993; Lin 2002), difficulties in self-disclosing to out-group members (Leong et al. 1995; Yeh 2002), and the belief that any psychological problem can be overcome by discipline, resolution, and perseverance (Boey 1999; Liou 2004), Chinese individuals are reluctant to seek professional help. In fact, many would only do so several years after the onset of depression, when the symptoms have become much more severe and unbearable (Chien et al. 2007; Horng 1994; Huang 1981).

Research suggests that in the Chinese culture, there is a strong tendency toward self-help in dealing with problems such as depression (Cheung et al. 1984; Loo et al. 1989). When asked to rate their help-seeking preferences, Chinese individuals strongly endorsed the choice of self-help (Sinclair 2000). The abundance of popular self-help psychology books available in bookstores in Taiwan also seems to support the self-help orientation. However, despite the popularity of self-help books in Taiwan, none of the self-help materials on the shelf has been subject to empirical scrutiny. It remains unknown whether these self-help books are in fact helpful.

Deriving from the Greek words for “book” (biblio) and “therapy” (therapeia), “bibliotherapy” has long been used as a therapeutic method (Campbell and Smith 2003; Pardeck 1998), and is thought to serve several functions such as providing information and guidance, generating insight, and offering solutions to problems (Campbell and Smith 2003). Using this approach, individuals would receive a standardized self-help book and work it through independently (Cuijpers 1997). The accumulating evidence suggests that bibliotherapy is an effective treatment modality (Cuijpers 1997; Gould and Clum 1993; Lovell and Richards 2000; Shechtman 2006; Williams 2003), for individuals ranging in age from adolescents (Ackerson et al. 1998) to older adults (Floyd et al. 2004; Gregory et al. 2004; McKendree-Smith et al. 2003; Scogin et al. 1990, 1989). There is also support for parent-implemented bibliotherapy for children (Lyneham and Rapee 2006; Rapee et al. 2006). Further, research findings appear to concur that bibliotherapy seems most useful for individuals who are without suicidality, with mild to moderate levels of depression (Campbell and Smith 2003; Cuijpers 1997; Gregory et al. 2004; McKendree-Smith et al. 2003).

Although quite a few randomized clinical trials have explored the efficacy of bibliotherapy in the treatment of adult depression, no study to date has been based in Asia. The lack of empirical efforts may be associated with the unavailability of a theory-driven, well-structured self-help tool. Based on a cognitive-behavioral perspective, the Chinese version of Greenberger and Padesky’s (1995) MindoverMood: ChangeHowYouFeelbyChangingtheWayYouThink seems to distinguish itself from other self-help books in the bookstores in Taiwan. The efficacy of cognitive-behavioral therapy (CBT) in the treatment of depression has been well documented (e.g., Blockting et al. 2005; Dobson 1989; Haaga et al. 1991). Further, Williams (2003) has noted that the features of CBT (e.g., well-defined theoretical structure, emphasis on psychoeducation, clear focus on current problems) make it an appropriate theoretical model for self-help treatment. However, given the well-documented cross-cultural differences in how individuals construe the self and others (Markus and Kitayama 1991, 1994; Okazaki 1997, 2000), direct transplantation of a Western treatment modality to Asia may raise concerns.

Hodges and Oei (2007) reviewed the literature and reported a high level of conceptual compatibility between CBT and the common values embraced by the Chinese culture (e.g., individualism-hierarchy, orderly autonomy, discipline-assertion, human-heartedness). They maintained that CBT, with some modification, could be effective for treating Chinese individuals. Despite the conceptual support, it remains unknown whether Asians in Asia would actually benefit from a CBT-based self-help book developed in the West.

In addition to examining the efficacy of bibliotherapy in treating depression, recent efforts have begun to attend to the mechanisms through which bibliotherapy may enact change. In outcome research, a mechanism of change refers to “[the] characteristics of the individual that are changed by the treatment and that, in turn, produce change in the outcome of interest” (Whisman 1993, p. 248). According to Kraemer et al. (2002), a mediator is something that “occurs during treatment” (p. 881). Learned resourcefulness, as measured by the Self-Control Schedule (SCS; Rosenbaum 1980), has been operationalized as “an acquired repertoire of behaviors and skills (mostly cognitive) by which a person self-regulates internal responses (such as emotions, pain, and cognitions) that interfere with the smooth execution of a target behavior” (Rosenbaum and Jaffe 1983, p. 216). Essentially, learned resourcefulness refers to an individual’s capacity to self-manage his/her behavior despite the experience of negative feelings. It enables the individual to feel more in control, and at the same time, to perceive the environment as being less demanding. Individuals with learned resourcefulness are able to cope with adverse situations and negative feelings through both formal and informal instructions (Rosenbaum and Jaffe 1983). In the SCS, the self-regulatory skills are categorized into four content categories: (a) use of cognitions and positive self-instructions to control emotional and physiological responses; (b) application of problem-solving strategies; (c) ability to delay immediate gratification; and (d) perceived self-efficacy.

Studies have found a negative association between learned resourcefulness and depression (Burns et al. 1994; Flett et al. 1991; Lewinsohn and Alexander 1990; MacLachlan 1985; Rokke et al. 2000). The construct has also been studied in Asia (Boonpongmanee et al. 2002; Huang et al. 2005; Nakano 1995; Rong 2000). In Taiwan, Rong (2000) reported that learned resourcefulness substantially attenuated the negative effect of stress on adaptive functioning in community-dwelling older adults. Huang et al. (2005) found that adolescents who demonstrated greater learned resourcefulness had fewer depressive symptoms; further, learned resourcefulness was found to mediate the effects of perceived poor health and peer relationship problems on depressive symptoms. It has been suggested that individuals undergoing bibliotherapy are likely to develop more skills and a sense of resourcefulness, thereby feeling well-equipped to work through the cognitive and behavioral tasks in the self-help book and consequently, experience a reduction of depressive symptoms (Rohen 1999). Nevertheless, the number of studies that examined the mediating role of learned resourcefulness is limited, and the results are inconclusive (Rohen 1999, 2002). On the basis of the current literature, it remains unclear whether change in learned resourcefulness may be a mechanism through which bibliotherapy reduces depressive symptoms.

In the assessment of depression, recent studies have begun to separately examine the subscales of the Beck Depression Inventory (BDI, Beck et al. 1961, e.g., Green et al. 2001; Furlong 2002; Ritterband and Spielberger 2001) and the Beck Depression Inventory-II (BDI-II, Beck et al. 1996, e.g., Chang 2007; Johnson 2006; Wenzel et al. 2005). It has been argued that individuals with the same total score likely exhibit different depressive symptom profiles, as such, merely interpreting the total score of BDI-II may lead to a loss of clinically relevant information about specific aspects of depression (Bedi et al. 2001). By examining the specific dimensions of depression, researchers and practitioners can apply appropriate interventions, and also evaluate the changes in clients and the effect of the interventions with “greater precision and specificity” (p. 315; Bedi et al. 2001). Based on the results of factor analytical studies on the BDI-II, symptoms of depression may best be captured by 2 dimensions—a cognitive-affective dimension and a somatic dimension (Beck et al. 1996; Dozois et al. 1998; Steer and Clark 1997; Storch et al. 2004; Whisman et al. 2000). The Chinese version of the BDI-II (Chen 2000) has also been found to comprise these 2 dimensions (Lu et al. 2002; Chang 2005). The specific factor loadings varied across different studies, which may result from the factor analytic extraction procedures used and the demographic characteristics of the specific samples (Chang 2005).

In sum, on the basis of the current literature on bibliotherapy, there is accumulating evidence that bibliotherapy may be a good treatment option for individuals with depressive symptoms. However, the extent to which a CBT-based treatment protocol that was developed in the West may be appropriate for individuals in the East remains unknown. Therefore, the present study sought to investigate the efficacy of cognitive bibliotherapy, as presented in the Chinese version of MindoverMood, in the treatment of depressive symptoms in Taiwan. Cognitive-affective and somatic symptoms of depression were examined separately. Further, the potential mediating role of learned resourcefulness was investigated.



The sample consisted of 52 participants (73% female; 27% male), ranging in age from 18 to 58 (M = 26.4; SD = 8.7). Of the participants, 43 (83%) were never married, and 9 (17%) were married. In terms of current employment status, 32 (61%) were college students, 16 (31%) were employed, and 4 (8%) were unemployed. The majority (n = 36, 69%) of the participants had never sought professional help for depression. Among the participants who had sought professional help for depression (n = 16, 31%), 7 (44%) received both medication and psychotherapy, 5 (28%) received psychotherapy alone, 3 (17%) received medication alone, and 1 (6%) received both medication and telephone counseling.


Potential participants were recruited via the following methods: posted flyers at university counseling centers, cafeterias and dormitories; posted flyers at family physicians’ offices; pamphlets distributed after classes; and posted information on internet message boards. Individuals who contacted the research office were given a preliminary screening using the following inclusion criteria: (a) over 18 years old; (b) self-reported feelings of depression; (c) not currently in psychotherapy; and (d) not currently taking psychiatric medication for depression.

A total of 153 individuals contacted the research office and expressed an interest in participation, 88 initially showed interest but did not enter the study. Among them, 58 could not be contacted, 11 could not commit to the lengthy participation, 9 refused due to a lack of interest, 5 were in psychotherapy, 3 were taking psychiatric medication for depression, and 2 reported that they were problem-free and did not need help.

As it turned out, 65 potential participants were further screened using the BDI-II. To be eligible, participants had to obtain a BDI-II score that was above 10 and below 47, further, they could not endorse the suicidal ideation item on the BDI-II. The lower BDI-II cutoff point, based on previous studies (e.g., Jamison and Scogin 1995; Rohen 2002), was used to prevent very mild cases from entering the study. As for the upper cutoff point, previous studies did not provide an agreed-upon criterion: While some studies (e.g., Ackerson et al. 1998) did not set an upper cutoff point, some (e.g., Floyd et al. 2004) required that their participants meet a DiagnosticandStatisticalManualofMentalDisorders (DSM-IV; APA 1994) diagnosis of major depressive disorder or Dysthymia, and some (e.g., Stice et al. 2006) excluded participants whose BDI scores exceeded 30. In the present study, we excluded individuals whose BDI-II scores exceeded 47, which, according to the psychometric data on the Chinese version of BDI-II (Lu et al. 2002), represents the midpoint of the severe depression range (31–63). As it turned out, 12 participants were excluded: Four had BDI scores that exceeded the upper cutoff point, 6 had BDI scores that were below the lower cutoff point, and 2 exhibited suicidal ideation. Those who were excluded were given a copy of MindoverMood (Chinese version) and appropriate community referrals.

Thus, 53 individuals who consented to participation and who met the inclusion criteria were randomly assigned to either the treatment group (bibliotherapy group) or the control group (delayed-treatment control group). However, the data of 1 participant in the control group had to be removed from analyses due to multiple missing responses at Time 1. During the study, 12 participants (6 from the treatment group; 6 from the control group) dropped out for various reasons (5 could not be contacted; 3 decided to seek psychotherapy; 2 reported not having enough time to participate; 1 decided to take medication for depression; and 1 had to discontinue due to a major surgery). About 40 participants (21 in the treatment group; 19 in the delayed-treatment control group) completed the treatment.

Participants in the treatment group were given a copy of MindoverMood (Chinese version) and were asked to read and complete chapters 1–10 (chapters that are relevant to depression) within 4 weeks. Participants received weekly check-in emails during the bibliotherapy phase, and were instructed that they could contact the research office should any concerns or questions arise. Participants completed the BDI-II and SCS at pretreatment, posttreatment (1 month later), and 3-month follow-up. A treatment compliance/comprehension scale was also administered at posttreatment. The participants assigned to the delayed-treatment control group were informed about their placement in the group, and were asked to wait 4 weeks before beginning treatment. During the waiting period, participants received weekly check-in e-mails, and were instructed that they could contact the researchers for any concerns or questions.

At the end of the waiting period, and prior to beginning bibliotherapy, participants were administered the BDI-II and SCS, after which they were given a copy of MindoverMood (Chinese version), and were asked to complete chapters 1–10 within 4 weeks. The rest of the procedure is identical to that for the treatment group specified above. Upon completion of the study, all participants were compensated TWD $500 (approximately US $15) for their participation.

The option of check-in e-mail was preferred by participants for privacy reasons. Some replied with brief updates or questions, others did not reply. Only one participant in the study did not have access to e-mail, and she was given a brief weekly check-in call by a research assistant. The phone conversation lasted no more than 10 min each time.


Beck Depression Inventory: BDI-II (Chinese Version)

The BDI-II is a 21-item, 4-point Likert-type self-report scale that assesses depressive symptoms that have occurred in the past 2 weeks. Ratings range from notatall to extremely. Each item corresponds to a symptom of depression, and is summed to yield a single scale score. Satisfactory psychometric properties have been reported. Specifically, internal consistency coefficients and test–retest reliability coefficients were found to be .92 and .93 (Beck et al. 1996). With regard to construct validity, the convergent validity of the BDI-II was established through correlations with the amended Beck Depression Inventory (BDI-1A; Beck and Steer 1993) and the Hamilton Psychiatric Rating Scale for Depression (HRSD; Hamilton 1960); the correlations were reported to be .93 and .71, respectively (Beck et al. 1996).

The Chinese version of BDI-II was translated by Chen (2000). The scale was found to be consisted of a cognitive-affective factor with 16 items (#1–14, 17, and 19) and a somatic factor with 5 items (#15, 16, 18, 20, and 21) (Lu et al. 2002). The internal consistency coefficient (Cronbach’s α) was reported to be .94, the split-half reliability was .91, and the convergent validity has been established through its correlation with the Chinese Health Scale (r = .69) (Lu et al. 2002). In the present sample, the internal consistency coefficient (Cronbach’s α) for BDI-II was found to be .84. In the Chinese version of BDI-II, the normal range is from 0 to 16; mild depression ranges from 17 to 22; moderate depression ranges from 23 to 30; and severe depression ranges from 31 to 63 (Lu et al. 2002).

The Self-Control Schedule

Consisting of 36 items, the SCS is a 6-point Likert-type scale (from veryuncharacteristicofme to verycharacteristicofme). It encompasses various topics (e.g., what one does when bothered by a negative thought, when a decision needs to be made or work needs to be done). Points associated with each item are summed across items to form the total score for the scale; higher scores indicate greater self-control (Rosenbaum 1980). Research suggests that the SCS is generally sound in psychometric properties (Boonpongmanee et al. 2002; Nakano 1995; Richards 1985; Rosenbaum 1980; Rude 1989; Zauszniewski 1997). Specifically, internal consistency coefficients have been reported to range from .78 to .93 (Rosenbaum 1980; Zauszniewski 1995). In our sample, the internal consistency coefficient (Cronbach’s α) for SCS was found to be .78. Test–retest reliability over a 4-week-period has been found to be .86 (Rosenbaum 1980), and that over an 11-month- period has been found to be .77 (Leon and Rosenthal 1984). The construct validity of the SCS has been demonstrated by correlating the total score with the following scales: CSE-D (Lewinsohn and Alexander 1990); Rotter’s (1966) I-E scale (Rosenbaum 1980); Jones’ (1968) Irrational Beliefs Test (Rosenbaum 1980); and coping response inventory (Nakano 1995). The SCS was translated into Chinese following the 4-step-procedure recommended by Behling and Law (2000).

Treatment Compliance/Comprehension

Treatment compliance was measured by participants’ self-report of the number of chapters and thought records completed at posttreatment. In addition, they were given an 8-item multiple-choice test that assessed participants’ comprehension of the major points discussed in the book. The short test was devised by the principal investigator. Sample questions include: “Which of the following is not one of the ‘five aspects of one’s life experiences’ discussed in Chapter 1?”; “Which of the following is a negative thought that people frequently hold when they are depressed?”

Feedback Questionnaire

Upon completion of the study, participants were asked to fill out a feedback questionnaire, which included open-ended questions about their general experiences in bibliotherapy, and the perceived strengths and limitations of the self-help book they read.


Preliminary Analyses

Prior to evaluating the hypotheses, potential confounding variables were examined. Chi-square tests and independent-samples t-tests were performed to compare the treatment group and the control group. Results of the analyses indicated that participants in the 2 groups did not differ in age, gender, martial status, employment status, prior experience with professional help-seeking, or Time 1 measures of BDI-II and SCS. Further, no intercorrelations were found among BDI-II, SCS, and the demographic variables. Thus, the demographic variables were not controlled in subsequent analyses. Note that Time 1 refers to the initial starting point, when the treatment group has not received bibliotherapy, and the delayed-treatment control group has not begun the waiting period; Time 2 refers to the posttreatment for the treatment group, and the post-waiting, pretreatment for the control group. The participants’ BDI-II scores ranged from 10 to 46; the mean score of BDI-II at Time 1 for all participants was 25.63 (SD = 8.53).

A comparison between the 12 participants who dropped out during the study and the 40 participants who completed the participation was conducted. Results indicated no significant differences between the dropouts and the completers in any of the demographic variables or Time 1 measures of BDI-II and SCS.

Main Outcome: Intention-to-Treat Approach

Fifty-two participants were randomly assigned to either the treatment group (n = 27) or the control group (n = 25). Analyses were performed on the intention-to-treat basis. Multiple imputation (MI) inference procedure (Rubin 1987) was adopted to estimate missing values and to draw inferences based on the imputed data. We conducted the procedure following Rubin’s suggestions, which involved 3 phases: (1) The regression method in SAS MI procedure (version 9.1) was used to fill in missing values 5 times, and to generate 5 complete data sets. (Rubin recommended the number 5 because additional imputations usually would not improve the precision of estimation in a meaningful way.) In the present study, the missing pattern in the data was found to be monotone. Given such, the missing values at Time 2 were imputed from treatment condition and Time 1 values; (2) The 5 complete data sets were analyzed using standard statistical methods, including paired t-tests for within-group comparisons, 2-sample t-tests for between-group comparisons, and linear regression analyses for testing mediation; (3) The results from the 5 complete data sets were then combined to produce inferential results by SAS MIANALYZE procedure. The average of the 5 separate estimates of effects was used as an overall estimate; the variance of the overall estimate was derived from both between- and within-imputation variances.

The results of intention-to-treat analyses for all participants (n = 52) and the analyses for the treatment completers (n = 40) are all summarized in Table 1.
Table 1

Bibliotherapy effects on BDI-II, cognitive-affective dimension, somatic dimension, and learned resourcefulness

All participantsa

Treatment group (n = 27)

Control group (n = 25)

Effect size (95% CI)


Time 1

Time 2


Time 1

Time 2



27.7 ± 9.1

18.2 ± 9.5

9.6 ± 8.4***

23.4 ± 7.6

20.9 ± 8.6

2.5 ± 10.0

7.0 (1.7 to 12.4)


Cognitive-affective dimension

21.6 ± 7.3

13.4 ± 7.9

8.2 ± 7.1**

17.4 ± 6.4

15.8 ± 7.7

1.6 ± 8.9

6.6 (1.8 to 11.3)


Somatic dimension

6.1 ± 2.6

4.8 ± 2.5

1.3 ± 2.1**

6.0 ± 2.2

5.1 ± 2.0

.9 ± 2.2*

.5 (−.8 to 1.7)


Learned resourcefulness

−3.1 ± 22.5

7.8 ± 23.1

−10.8 ± 15.4***

2.7 ± 16.8

3.6 ± 17.5

−.9 ± 9.7

−9.9 (−17.6 to −2.3)


Participants with complete data

Treatment group (n = 21)

Control group (n = 19)

Effect size (95% CI)


Time 1

Time 2


Time 1

Time 2



28.9 ± 9.7

18.8 ± 10.0

10.1 ± 8.1***

24.9 ± 7.5

20.9 ± 8.5

4.0 ± 9.8

6.1 (.3 to 11.9)


Cognitive-affective dimension

22.4 ± 7.7

13.8 ± 8.3

8.5 ± 6.9***

18.6 ± 6.3

15.8 ± 7.5

2.7 ± 8.7

5.8 (.8–10.8)


Somatic dimension

6.5 ± 2.7

5.0 ± 2.6

1.5 ± 2.1**

6.3 ± 2.4

5.1 ± 2.2

1.3 ± 2.1*

.3 (−1.1 to 1.6)


Learned resourcefulness

−1.5 ± 21.5

9.0 ± 23.4

−10.5 ± 16.3**

5.1 ± 16.8

4.7 ± 18.1

.4 ± 9.3

−10.9 (−19.5 to −2.3)


Data are means ± SD. Effect sizes and listed p values are for comparisons of between-group (treatment versus control) changes (Time 1 to Time 2)

aMissing values at Time 2 were estimated by multiple imputation

p < .05, ** p < .01, *** p < .001, for within-group changes

As shown in Table 1, effect sizes (ES) were defined as the differences of between-group (treatment versus control) change from Time 1 to Time 2 in the outcome measures. Cohen’s d (standardized effect size)—the difference between mean changes divided by the pooled standard deviations for those means—was used as an indicator of the magnitude of treatment effect. According to Cohen (1992), d = .2 is indicative of a small effect, .5 a medium effect, and .8 a large effect.

As shown in Table 1, significant differences between groups were found for the overall depression level (p < .05), cognitive-affective symptoms (p < .001), and SCS (p < .05). However, no significant result was found for somatic symptoms, n.s. As for within-group comparisons in the treatment group, significant changes were found in the overall depression level (p < .001), cognitive-affective symptoms (p < .01), somatic symptoms (p < .01), and learned resourcefulness (p < .001) from Time 1 to Time 2. In the control group, significant changes were found in somatic symptoms (p < .05) only.

Effect size statistics showed a medium to large difference in the overall depression level between the participants of the two groups (ES = 7.0; d = .76). Bibliotherapy had a strong effect on the cognitive-affective symptoms of depression (ES = 6.6; d = .82), but not on the somatic symptoms of depression (ES = .4; d = .23). In addition, a medium to large effect size was found for learned resourcefulness (ES = −9.9; d = −.77).

Results of the analyses for the treatment completers (n = 40) were similar to those of the intention-to-treat analysis displayed above (Table 1).

Clinical Significance

Clinically significant change was assessed using the index suggested by Jacobson and Truax (1991). In the Chinese version of BDI-II, a cutoff score of 17 marks the beginning of mild depression range (Lu et al. 2002). Thus, we examined the proportion of individuals who returned to a normative level of functioning (BDI-II < 17). Changes in depressive symptoms from Time 1 to Time 2 were assessed using McNemar’s test. In the treatment group, 12 (48%) out of 25 subjects who scored BDI-II ≥ 17 at Time 1 achieved clinically significant change following the treatment, and 0 out of 2 subjects whose BDI-II scores were below 17 at Time 1 scored above the cutoff point following the treatment. Results indicated that participants in the treatment group experienced clinically significant change (McNemar, p < .001). In the control group, 8 (40%) out of 20 subjects who scored BDI-II ≥ 17 at Time 1 achieved clinically significant improvement at Time 2, and 3 (60%) out of 5 subjects who scored BDI < 17 at Time 1 became clinically depressed at Time 2. Clinically significant change was not observed in the control group (McNemar, p = .23).

Mediation of Treatment Effects

As shown previously, bibliotherapy was found to be effective for reducing the cognitive-affective, but not the somatic, symptoms of depression. In the following analyses, change in learned resourcefulness during the treatment was examined as a potential mediator in the relationship between bibliotherapy and cognitive-affective symptoms. Baron and Kenny’s (1986) approach was adopted to test the mediation. The criteria for mediation are as follows: (1) Treatment condition is related to the outcome variables; (2) treatment condition is related to the proposed mediator; (3) the proposed mediator is related to the outcome variables; and finally, (4) when the relation between treatment condition and the mediator is controlled for, the relation between treatment condition and the outcome variables is substantially reduced. If the relation between treatment condition and outcome variables is reduced to zero and completely eliminated, then the mediation is considered to be full. On the other hand, if the relation between treatment condition and outcome variables is reduced in absolute size but is still different from zero, then the mediation is considered to be partial (Baron and Kenny 1986).

Results of the previous analyses indicated that the first two criteria were met. As for the third criterion, change in learned resourcefulness was found to be associated with change in cognitive-affective symptoms (r = −.49, p < .001). As for the fourth criterion, a regression model was constructed, with change in learned resourcefulness and treatment condition as predictors of change in cognitive-affective symptoms. Results indicated that when the relation between treatment condition and change in learned resourcefulness was controlled for, the effect of treatment condition on cognitive-affective symptoms was reduced (from ES = 6.6 to ES = 4.04), and no longer significant (p = .08). When treatment condition was controlled for, change in learned resourcefulness was significantly associated with change in cognitive-affective symptoms (p < .01); each point of increase in learned resourcefulness was associated with .25 point decrease in cognitive-affective scores. It appears that the criteria for mediation were met. Learned resourcefulness was found to partially mediate the effect of bibliotherapy in the reduction of cognitive-affective symptoms of depression. Results of the linear regression analyses are summarized in Table 2.
Table 2

Linear regression analyses predicting change in cognitive-affective dimension


Regression coefficient

Standard error


Treatment condition




Change in learned resourcefulness








R2 = .29; results were based on multiple imputation inferences

Maintenance of Treatment Effects

To assess the maintenance of treatment effects, analyses were conducted on the data of treatment completers. The treatment outcome data of the treatment group (n = 21) were combined with those of the delayed-treatment control group (n = 19). Paired sample t-tests that compared BDI-II scores at posttreatment and those at 3-month follow-up were conducted. Results indicated that a further reduction in participants’ overall depression levels took place during the follow-up period, t(39) = 2.25, p < .05. Further, results of paired sample t-tests that compared scores in the cognitive-affective dimension at posttreatment and those at 3-month follow-up indicated that a further reduction in participants’ cognitive-affective symptoms took place during the follow-up period, t(39) = 2.09, p < .05.

Treatment Compliance and Comprehension

Evaluation of treatment compliance data suggested that participants were engaged in bibliotherapy. The average number of chapters completed was 7.83 (SD = 2.99); the average number of thought records completed was 1.73 (SD = .78). In terms of treatment comprehension, the mean score of participants’ quiz score was 89.06 (SD = 10.66). However, results of correlational analyses indicated that neither treatment compliance nor comprehension was significantly correlated with outcome.

Qualitative Feedback Regarding Bibliotherapy and the Specific Self-Help Book

Participants were asked to evaluate their experiences in bibliotherapy and to report the strengths and limitations of the self-help book they read. Most participants pointed out both positive and negative aspects of their bibliotherapy experiences. They reported the following strengths: convenience, no need to talk to others (who may not understand), enjoyment in reading, provision of well-organized and theory-driven information, being able to understand and change negative thoughts using multiple methods, and prevention of a depressive episode. The limitations they reported included: not being able to concentrate on reading when one is feeling low, lack of professional supervision, not being able to complete the exercises by oneself, and a book will never compare to the support given by one’s loved ones.

Regarding the content of the self-help book, MindoverMood (Chinese version), most indicated that the book was easy to read. Some reported that although it took much time and patience to practice, they benefited the most from filling out the thought record. Some stated that they appreciated the cognitive perspective, and the opportunity to objectively analyze their situations by asking themselves, “What’s going on with me? What thoughts do I have now?” They indicated that they were more able to change their negative views in various situations. In the meantime, however, quite a few participants indicated that it would have been much easier for them to identify with the characters, if the characters’ names, cultural backgrounds and conflicts were more similar to their own. In addition, some reported that they would appreciate more interesting examples, a more encouraging tone (e.g., “Come on, let’s give it a try!”), and colorful page layouts with pictures on the side. Further, some found it difficult to understand some jargons, or to complete the worksheets without help from a therapist.


There is considerable evidence base for the use of cognitive bibliotherapy in the treatment of depressive symptoms. However, the absence of Asians as research participants in randomized clinical trials makes it difficult to ascertain if cognitive bibliotherapy may be an appropriate treatment choice for Asians. The present study offers preliminary support for the use of cognitive bibliotherapy in the Chinese population: participants’ depression level lowered at posttreatment; more specifically, cognitive-affective symptoms, rather than somatic symptoms, evidenced significant reduction as a result of bibliotherapy. Further reductions in cognitive-affective symptoms were observed at 3-month follow-up. Moreover, we found evidence suggesting mediation by learned resourcefulness. Lastly, participants’ qualitative feedback may provide important direction for cross-cultural adaptation of cognitive bibliotherapy.

It has been observed that depressed Chinese individuals tend to report somatic symptoms, and to deny or minimize cognitive and affective symptoms of depression (Katon and Kleinman 1982). Recently, however, Chinese individuals have been found to readily acknowledge cognitive-affective symptoms of depression (Chang 2007). Such finding echoes Hwang et al. (2006)’s report that Chinese individuals do experience emotional and cognitive symptoms of depression. In the present study, participants readily reported both cognitive-affective and somatic symptoms. It was found that bibliotherapy lowered participants’ cognitive-affective, but not the somatic, symptoms of depression. Given that MindoverMood focused primarily on the teaching of cognitive skills, it seems reasonable that cognitive bibliotherapy was found to be particularly useful for treating the cognitive-affective symptoms of depression. In the meantime, the lack of effects for the somatic symptoms may have to do with the fact that MindoverMood does not provide behavioral interventions that directly target somatic complaints (e.g., relaxation training techniques). Future studies that evaluate both cognitive bibliotherapy and behavioral bibliotherapy will help clarify their specific effects on the different dimensions of depressive symptoms.

During the waiting period, the control group’s overall depression score did not change. Yet, upon further examination, it was found that while the cognitive-affective symptoms of depression remained the same, the somatic symptoms lowered slightly. It may be plausible that the somatic symptoms may demonstrate a slight reduction during watchfulwaiting, yet cognitive-affective symptoms of depression would not change unless specific interventions were implemented. However, it is merely a speculation that requires further investigation. Also, it has been speculated that participants maintain bibliotherapy treatment gains through rereading sections of the book (Floyd et al. 2006). The present finding that participants’ depression symptoms continued to lower from posttreatment to the follow-up may be related to their rereading parts of the book or mentally rehearsing some of the cognitive strategies.

Although both statistical and clinical significance were observed in the present study, several culturally relevant issues should be noted. Certain cultural variables may have played a role in the responsiveness of participants to the Western-based protocol. An examination of participants’ qualitative feedback reveals that quite a few participants had difficulty identifying with the main characters of the book, for these characters’ names, cultural backgrounds, experiences, and personal conflicts were dissimilar to their own. Chen and Davenport (2005) identified the main teachings of confucianism as the core of Chinese values, they include: filial piety, respect for familial and social hierarchy, discouragement of self-centeredness, emphasis of academic achievement, and importance of interpersonal harmony. In addition, it has been found that Asians—with interdependent construal of the self—tend to define the self based on their relationships with others (e.g., family members) and to derive self-worth through their abilities to be connected to the social environment (Markus and Kitayama 1991; 1994; Okazaki 1997). For example, a typical conflict experienced by adults in the mid-20s (or 30s) in Taiwan is whether to move away from home. Though desiring independence and freedom, many would also experience tremendous guilt for wanting to leave their parents. The situation of “wanting to find one’s own place” likely generates feelings of guilt, anxiety, depression, and anger, which may be related to dysfunctional thoughts such as “I am no good!” and “Everyone is disappointed in me!” This example highlights most of the core Chinese values noted by Chen and Davenport (2005), and makes evident the interdependent nature of Chinese individuals. It may be speculated that had examples like this been incorporated into the book, the participants might have identified with the characters more, and benefitted even more from bibliotherapy.

In addition, Hwang et al. (2006) offered 18 principles for adapting CBT to clients of Chinese origin (for a detailed discussion, see Hwang et al. 2006). Importantly, some of these principles also appear to be applicable to cognitive bibliotherapy, including: focusing on the psychoeducational aspects of the treatment (Principle 4); bridging CBT concepts and Chinese cultural beliefs (Principle 5); keeping in mind the highly family oriented nature of the Chinese culture (Principle 10); being sensitive to the “shame and stigma associated with having a mental illness” (p. 297) (Principle 11); recognizing that Chinese individuals do experience emotional and cognitive symptoms of depression (Principle 16); and helping clients to understand the relationships between “biomedical and psychosocial models of disease development” (p. 299) (Principle 17). If we examine MindOverMood (Chinese version) using these principles, we can see that while the principles that are related to CBT (Principle 4, 16, and 17) are reflected in the book, others (Principle 5, 10, and 11) are not. Given the results of the present study, it appears that cognitive-behavioral treatment, as disseminated in the form of a self-help book, can be effective for Chinese individuals. However, it is quite possible that a self-help book can be made even more effective if the Chinese values were integrated with the teaching of cognitive skills.

The mechanisms through which bibliotherapy may effect change have been explored in previous studies. Dysfunctional thinking (Ackerson et al. 1998; Scogin et al. 1989), hopelessness (McKendree-Smith 2000; Rohen 2002), and learned resourcefulness (Rohen 1999, 2002) have all been proposed as potential mechanisms of change, yet the results are far from being conclusive. The potential mediating role of learned resourcefulness was examined in the present study, and it was found that learned resourcefulness met the criteria for partial mediation.

Learned resourcefulness is conceptualized as a repertoire of cognitive-behavioral coping strategies for managing aversive stimuli (Rosenbaum 1980; Rosenbaum and Jaffe 1983). With these coping strategies, individuals would not try to avoid aversive stimuli (i.e., coming up with excuses to skip a class and avoid oral presentation); instead, they would try to face the challenges by engaging in cognitive and behavioral efforts (e.g., minimizing self-defeating internal dialogues, focusing on their positive abilities/attributes, providing self-reinforcers upon the achievement etc.). As suggested by the present findings, bibliotherapy may lower cognitive-affective symptoms by increasing an individual’s capacity to self-regulate negative internal responses (e.g., emotions, pain). It appears that these coping strategies are helpful for Chinese individuals. Thus, it is possible that interventions that target learned resourcefulness may be helpful for Chinese individuals with depressive symptoms.

Measured by participants’ self-report of the number of chapters and thought records completed at posttreatment, treatment compliance was not found to be related to the treatment outcome. It is possible that there exists tremendous variation in how meaningful the readings were perceived. An individual may complete only 3 chapters and 2 thought records and have meaningful gains, while another may complete 10 chapters and 10 thought records without appreciating their meaning. Further, although the participants seem to have acquired the major CBT principles through bibliotherapy, comprehension of the major points was also found to be unrelated to treatment outcome. The limited items on the test and its questionable validity may account for this finding.

The dissemination of evidence-based interventions for depression is worthy of discussion. In recent years, the stepped-care model proposed by Lovell and Richards (2000) has been given increasing attention (Mains and Scogin 2003; Gregory et al. 2004; Reeves and Stace 2005; Williams 2003). In this model, clients with mild to moderate depression may not need such intensive intervention, and may be appropriate targets of level 1 intervention—the less intensive treatments such as bibliotherapy. When necessary, these clients may be stepped to the next level of care (National Institute for Clinical Excellence 2002). In the present study, although some participants had no problem completing thought records on their own, some struggled with tremendous difficulty and gave up prematurely. Although they might have gotten some help from other parts of the book (e.g., information about depression), they apparently could not grasp cognitive-restructuring on their own. These participants may benefit more from some professional guidance.

In addition, more than half of the participants noted that they could not concentrate on reading when they were extremely depressed. Although some of them tried hard to continue reading, the words did not mean much to them at the time. Campbell and Smith (2003) suggested that it may be a good idea for a depressed individual to be given audio tapes rather than books if his/her energy level is low. As a depressed individual’s symptom state may fluctuate, it may be even more helpful to provide them with both a self-help book and an audio-visual tool.

Due to a number of limitations to this study, the findings reported need to be interpreted with great caution. First, the results are based on self-report measures of depression and learned resourcefulness, demand characteristics may have influenced participants’ responses. Further, the study is limited in terms of generalization of outcomes because of the specific sample. The present sample consisted of a very small group of young adults in Taiwan who reported depressive symptoms (not a full diagnosis of Major Depressive Disorder or Dysthymia) and who volunteered to go through bibliotherapy; as such, the present findings may not be generalizable to individuals who differ on these attributes. Finally, given the brief follow-up, it is not possible to observe the long-term effects of bibliotherapy.

Despite the limitations, the present study provides preliminary evidence for the utility of bibliotherapy in the Chinese population, which is consistent with the self-help cultural orientation. Although depressed individuals in Taiwan tend to rely on self-help, most self-help books available on the shelf are not based on sound psychological theories. As such, there is a need to provide them with evidence-based recommendations for self-help books. Grounded on the cognitive-behavioral approach, the Chinese version of MindoverMood was found to serve the functions of lowering participants’ cognitive-affective symptoms of depression, possibly by increasing their sense of resourcefulness. Cognitive bibliotherapy appears to be a promising treatment option for depressed, non-suicidal adults who are receptive to this treatment modality. Given the high accessibility and cost effectiveness, it may be of special value for the vast majority of Chinese individuals with depression who are undertreated in the mental health system. The current findings warrant further exploration in appropriate cross-cultural adaptation of cognitive bibliotherapy. The development of self-help books that are of more cultural relevance to depressed individuals in Taiwan remains an important task. Future studies are needed to expand on the present findings.


This study was funded by a research grant (94-2413-H-030-002) from the National Science Council (NSC) in Taiwan.

Copyright information

© Springer Science+Business Media, LLC 2008