A Survey on Chinese Scholars ’ Adoption of Mixed Methods

Since the 1980s when mixed methods emerged as “the third research methodology”, it was widely adopted in Western countries. However, inadequate literature revealed how this methodology was accepted by scholars in Asian countries, such as China. Therefore, this paper used a quantitative survey to investigate Chinese scholars’ perceptions and adoption of mixed methods in China. The data of the study were obtained from 247 Chinese scholars in higher education. Structural equation modelling was used to examine the relationship between participants’ perceptions and use of mixed methods. The results revealed that Chinese scholars’ research expertise of using quantitative and qualitative methods as well as their perceived advantage of using mixed methods has significantly influenced their adoption of mixed methods. This paper advanced the literature of the evolution of mixed methods by investigating the expansion and adaptability of mixed methods in an Asian context.


Introduction
Mixed methods have been called "the third research paradigm" along with traditional quantitative and qualitative methodologies (Johnson & Onwuegbuzie, 2004, p. 14).It was defined as the combination and integration of quantitative and qualitative methods in a single study or in a program of study (Creswell & Plano Clark, 2007;Creswell & Plano Clark, 2011;Greene et al., 1989;Tashakkori & Teddlie, 1998).The development and evolution of mixed methods has experienced several stages in the last 20 years: the formative stage (1980s and before), the paradigm stage (1980s to 1990s), the procedural stage (1980s to present), the advocacy stage (early 2000s to present), the reflective stage (2000s to present), and the expansion stage (2010s to present) (Creswell & Plano Clark, 2011;Tashakkori & Teddlie, 1998;Teddlie & Tashakkori, 2009).Although mixed methods have developed for 30 years, only the recent literature discussed its expansion in different countries (Creswell, 2009;Teddlie & Tashakkori, 2009).Most of the articles focused on the adoption of mixed methods in Western countries, whereas little literature discussed its adoption in non-Western cultural contexts such as East Asia (Creswell & Plano Clark, 2011;Teye, 2012).To address the scant literature, this study investigated the expansion of mixed methods in China, an East Asian country.
Expansion, or diffusion, is a special type of communication, in which an innovation is spread among the members of a social system (Rogers, 2003).An innovation was defined as an idea or a practice that is perceived as new by the population of adoption (Rogers, 2003), such as mixed methods being a new methodology to researchers.In this study, mixed methods were assumed to be the innovation that fit in Rogers' (2003) definition of innovation.Such assumption was supported by Johnson and Onwuegbuzie (2004, p. 15), "if one prefers to think categorically, mixed methods research sits in a new third chair, with qualitative research sitting on the left side and quantitative research sitting on the right side" As Rogers ' (2003) diffusion theory of innovations declared, an individual's decision of adopting an innovation was influenced by five intrinsic characteristics of the innovation: the relative advantage, compatibility, simplicity, trialability, and observability.Researchers have developed reliable measures of the five intrinsic attributes, including the Perceived Usefulness, the Compatibility, the Perceived Ease of Use, the Visibility, and the Trialability (Davis et al., 1989;Moore & Benbast, 1991, 2001).Besides the scales, Davis, Bagozzi, and Warshaw (1989) published the Technology Acceptance Model (TAM), which has been widely used in the field of innovation diffusion research (Al-Azawei, 2017;Agag & El-Masry, 2016;Davis, 1993;Hye-Young, 2016;Liaw, ies.ccsenet. 2002;Liaw 2003).
In this stud to investig existing m Moreover, this study is meaningful to research methodologists who are interested in quality practices of using mixed methods in specific disciplines.Historically, when methodologists drafted the guidelines for using mixed methods in specific disciplines, they mostly considered the situations in developed countries due to lack of relevant literature in other cultures.However, when one discipline adopts a new research method, it requires a complete analysis and best practices for using this method in different situations, either in developed or developing countries, and either in Western cultures or non-Western cultures.Thus, this study intended to provide the needed information of mixed methods' compatibility in a different context.In essence, the current study could contribute to the development of mixed methods in expanding its frontier of application.

Sampling
The population of the study was defined as Chinese scholars because they are the primary group using or intending to use mixed methods in China.A sample of 247 faculty members and senior graduate students were recruited from the top 300 comprehensive universities of 2012 in China.Multiple sampling strategies were used, including random sampling and criteria sampling.Firstly, three Chinese universities among the top 300 comprehensive universities of 2012 in China were randomly selected.Secondly, criteria sampling strategies were employed.Faculty members and senior graduate students were invited to participate in the survey if they had learned about research methodologies, qualitative and or quantitative approaches.Consequently, a total sample of 247 participants was recruited, averaging 80 at each research site.The sample information was reported in Table 1.
In addition to the four existing scales, another four scales were generated based on literature review and a previous qualitative study (Zhou & Creswell, 2012).The scale of Reasons to Use Mixed Methods (3 items; α= 0.88) was developed to measure participants' perception of the reasons to use mixed methods on a 7-point Likert scale from 1 (strongly disagree) to 7 (strongly agree).This scale consisted of the following three items: 1) The weaknesses of one research method can be offset by the strengths of the other research method.
2) Using mixed methods can solve complex research problems.
3) Using mixed methods can provide a complete understanding of research problems.
Another three scales meant to measure participants' expertise of using different research methods, namely Expertise of Using Qualitative Methods (6 items; α= 0.78), Expertise of Using Quantitative Methods (6 items; α= 0.79), and Use of Mixed Methods (6 items; α= 0.79).Participants were asked to give numeric response to the following six questions about their experience of using specific methods.The raw responses were then converted to the data within a range of 1 to 7, with higher score indicating higher-level expertise with the specific research methodology.Each of the three scales included the following six questions, such as: 1) How many courses did you take to learn about this methodology?
2) How many training occasions (including conferences, workshops, lectures, and seminars) did you attend to learn about this methodology?
3) How many studies did you participate in using this methodology?4) How many times did you present this type of research (including publishing papers, presenting at conferences, and speaking at lectures)?
5) How many articles (including book chapters) did you read about this methodology?6) How many times did you talk with others about using this methodology?

Data Collection
The survey questionnaire was originally worded in English and then was translated into Chinese for the participants of the study.The accuracy of translation was discussed by a panel of Chinese scholars (including five Chinese graduate students in the United States and five Chinese faculty in China).The Chinese versions of the questionnaire and the informed consent were approved by the Institutional Review Board of the PI's institute.
The survey was conducted in several sessions at each site.The procedure of data collection was the same for every session.First, the university administrators gathered the participants.Second, the PI briefly introduced the study to participants.For instance, the PI clarified the terms used in the questionnaire to make sure participants' understanding of these terms was accurate and consistent.The overall response rate was 91%, as the authors received 247 complete surveys out of 270.The average time for completing the survey was approximately 10 minutes.

Data Analysis
Data were analysed using descriptive statistics and structural equation modelling techniques.Data analysis was conducted in Mplus 7, which provided the maximum likelihood estimation with robust standard errors (MLR) for non-normal continuous variables with missing data.For all variables in the study, there was less than 4% missing data at the assumption of missing at random.To evaluate model fit, multiple model fit indices were used, including model chi-square (χ 2 ), Bentler comparative fit index (CFI; Bentler, 1990), standardized root mean square residual (SRMR), and Steiger-Lind root mean square error of approximation (RMSEA; Steiger, 1990).
According to Kline (2011), the combination of chi-square values accompanying p values greater than .05,CFI values greater than .95,RMSEA values less than .05,and SRMR values less than .08indicated a good model fit.
The sample size (N > 200) in the quantitative survey decreased the concerns of power in the study.The effect size of the model and coefficients estimates were evaluated and reported using the R 2 s.

Results and Discussion
The total sample size was 247, with 185 females (75%) and 62 males (25%).The demographic information of all the participants was summarized in Table 2.In summary, the average age of the participants was 27.35 with a standard deviation of 7.32.The youngest participant was 22 and the oldest was 66.Approximately 72% (n=179) of the participants were graduate students (about half of them being senior master students and the second half being senior doctoral students), and 28% (n=68) of the participants were university faculty members (50% of them being assistant professors, 26% being professors and associate professors, as well as 24% being research faculty).Participants came from a variety of disciplines, including public health (n=51, 20%), nursing (n=48, 19%), psychology (n=47, 19%), management (n=41, 17%), education (n=36, 15%), and arts and sciences (n=24, 10%).In the study, the hypothesized model was used to examine the influential factors on Chinese scholars' use of mixed methods.It consisted of eight variables.The descriptive statistics of all the variables were reported in Table 3. Note.* p<.05.

Chinese Scholars' Likelihood of Adoption
As the results indicated, the mean response to the scale of Intention to Use was high at 5.18 on a 7-point scale, with higher scores indicating a higher level of intention to use mixed methods in research.The results also indicate that participants' intention to use mixed methods were moderately correlated with their perceptions of adopting mixed methods, including perceived advantage, compatibility, reasons, and ease of using mixed methods (all rs > .36*),which were all reported with means higher than 4.5 on a 7-point scale.The above results provide us with the answers to the first research question of the study.That said, it is very likely that Chinese scholars will adopt mixed methods as a new methodology in their research.Meanwhile, Chinese scholars also have very positive perceptions toward using mixed methods.
Though Chinese scholars reported that they strongly intended to adopt mixed methods, their current use of it was less than frequent.According to the results, the mean of using mixed methods was 1.83 on a 7-point scale, with the standard deviation of 1.20.Such conflicting results were explained by Roger ( 2003) that people' intention to adopt an innovation and their actual adoption were two different issues.The above results implied that there were issues hindering the actual adoption of mixed methods in China.A question lingers as besides the intention to use, what other factors could influence Chinese scholars' adoption of mixed methods?
The final model was estimated to identify the influential factors of participants' intention to use mixed methods and their actual use of mixed methods.The model results with the standardized path coefficients are reported in Figure 2 The above findings are consistent with the adoption research in many other fields, where researchers found that perceived compatibility and advantage of innovation are two significant predictors to people's innovation adoption (Al-Azawei, 2017;Chang & Tung, 2008;Hardgrave, Davis, & Riemenschneider, 2003;Hye-Young, 2016;Lala, 2014;Lee, Hsieh, & Hsu, 2011;Powell & Wimmer, 2017;Wu & Wang, 2005).
One original finding of the study is that Chinese scholars' perceived reasons to use mixed methods is a significant predictor of their intention to adopt this method.This result implies that researchers are more likely to adopt mixed methods when they are fully aware of the needs and contexts to use this method in practice.That said, if researchers could read examples of using mixed methods in various cases, they might gain a better understanding of why mixed methods is needed in specific fields and inquiries.This finding delivers useful suggestions to methodologists who are drafting mixed methods research guides and trainings for practitioners.
Lastly, according to hypothesis 3, the possible mediation effects in the model were tested.The results indicated that Chinese scholars' perceived advantage of using mixed methods was a significant mediator on Chinese scholars' intention to adopt mixed methods.Therefore, hypothesis 3 of the study was supported (Table 4).Specifically, participants' perceived advantage partially mediated the relationship between perceived compatibility and intention to use (std. β* β=. 19, p<. 001).The 95% confidence interval of such mediation is [.10, .26]using the Monte Carlo resampling method (replication = 20,000).In addition, participants' perceived advantage also partially mediated the relationship between perceived reasons and intention to use (std.β* β=.09, p<.01).The 95% confidence interval of such mediation is [.05, .16]using the Monte Carlo resampling method (replication = 20,000).Additionally, participants' perceived advantage fully mediated the relationship between perceived ease of use and intention to use mixed methods (std. β* β=.05, p<.05).The 95% confidence interval of such mediation is [.01, .07]using the Monte Carlo resampling method (replication = 20,000).
In brief, the Relative Advantage is a unique factor predicting Chinese scholars' adoption of mixed methods.It has both direct effects and mediation effects on people's decision of adoption.The mediation effects exist between several predictors and the outcome of adoption so they cannot be ignored.This finding is new to the literature of diffusion of innovations.It suggests that such mediation effects should be examined in other innovation adoption studies in the future.H1.Chinese scholars' intention to use mixed methods is positively predicted by their perceived compatibility, advantage, reasons, and ease of using mixed methods, as well as their expertise of using qualitative and quantitative methods.

Partially supported
because not all the predictors were significant.
Significant factors: compatibility, advantage, and reasons of using mixed methods.
H2. Chinese scholars' use of mixed methods is positively predicted by their perceived compatibility, advantage, reasons, and ease of using mixed methods, as well as their expertise of using qualitative and quantitative methods.

Partially supported
because not all the predictors were significant.
Significant factors: expertise with qualitative methods, expertise with quantitative methods, and perceived ease of use.
H3. Chinese scholars' perceived advantage of using mixed methods has mediation effects on their intention to adopt mixed methods.Supported.
Significant mediation effects of perceived reasons, compatibility, ease of use on intention to use mixed methods.

Conclusion
The study makes original contribution to the literature of the evolution of mixed methods by quantitatively examining the expansion of mixed methods in an East Asian country.According to the results of the study, Chinese scholars' positive perceptions of using mixed methods have greatly helped with this methodology's expansion in China.However, though Chinese scholars have high level of intention to adopt mixed methods, they have not widely used this methodology in practice.The major challenges that have hindered mixed methods' adoption are due to researchers' insufficient expertise in using quantitative and qualitative methods.Moreover, people's perceived advantage of using mixed methods has significant mediation effects between perceptions and adoption of this method.That said, if researchers are not cognizant of the advantages of using mixed methods, they may not adopt this method despite their positive perceptions.Extra attention should be placed on this mediator in order to enhance the adoption of mixed methods in China.Another point of emphasis is that Chinese scholars' research expertise of both quantitative and qualitative methods should be improved, so that their strong intention of using mixed methods could be sustained.All these efforts would feed into the incremental expansion of mixed methods use in research in China.
Due to the limited time span for data collection and insufficient funding of the study, the researcher could not examine the topic in a wider geographic range of China and for a longer time.Data should be collected every other five years to update the results.Another limitation lies in that the participants were not randomly selected at each research site.Their research experience varied that could make the results biased.The above limitations should be considered when generalizing the results of the study.
Despite the limitations, the study investigated the expansion of mixed methods in China, a non-Western developing country.The study quantitatively confirmed the relationship between researchers' perceptions of using mixed methods and adoption, which is a new finding to the literature.In all, this study presents valuable information and suggestions on the expansion of mixed methods in a non-Western culture, thus contributing to the formation of international mixed methods community.More future research is necessary to examine the adoption of mixed methods in other non-Western contexts.
learn about this methodology?

使用混合方法能使我更快地完成研究的任务。
2. Using mixed methods makes it easier to do my job.

使用混合方法让我的研究变得简单。
3. Using mixed methods gives me greater control over my research process.

使用混合方法让我能更好地掌控我的研究过程。
4. Using mixed methods improves the quality of work I do.

使用混合方法能提高我的工作／研究质量。
5. Using mixed methods enhances my effectiveness on the job.

使用混合方法能提高我的工作／研究成效。
Compatibility of Using Mixed Methods (α=.80) 1.I think that using mixed methods fits well with the way I like to work.

我觉得使用混合方法十分符合我所喜欢的研究方式。
2. Using mixed methods is completely compatible with my current situation.

使用混合方法很符合我做科研的风格。
Ease of Use of Mixed Methods (α=.73) 1. Overall, I believe that mixed methods is easy to use.

总的来说，我认为混合方法很容易用。
2. Learning to use mixed methods is easy for me.

学会使用混合方法对我来说很简单。
3. The use of mixed methods is clear and understandable to me.

对我而言，混合方法的使用简单易懂。
4. I believe that it is easy to use mixed methods to get my research questions answered.

我相信使用混合方法能很容易地解决我的研究问题。
Reasons to Use Mixed Methods (α=.88) 1.The weaknesses of one research method can be offset by the strengths of the other research method. 在一项研究中，同时使用量化和质性方法能让他们互补不足。

Table 1 .
Sample size and information of three selected Chinese Universities

Table 2 .
Table of participants' demographic information

Table 3 .
Means, SDs, Ns and bivariate correlations of variables in the model .

Table 4 .
Hypotheses of Chinese scholars' adoption of mixed methods