Exploring social work students’ attitudes toward research courses: predictors of interest in research-related courses among first year students enrolled in a bachelor’s programme in Switzerland

Abstract

Research courses have become a taken-for-granted component of social work study programmes. Nonetheless, studies still confirm that social work students are reluctant to engage with research courses. They report considerably negative beliefs and attitudes as well as higher levels of anxiety. The present study aims to assess the interest in research courses among students entering a BSW programme in Switzerland. Further, it aimed to establish the relationship between students’ interest in research courses, their research orientation, and research anxiety. From 2014 to 2016, 708 first-year students were surveyed using an online questionnaire. Data were analysed using descriptive statistics and structural equation modelling. Analysis showed that the interest in research courses was predicted by research anxiety (β = −.29) as well as by the perceived importance (β = .27), the attributed usefulness (β = .15) and the perceived unbiased nature of research (β = .08). These variables were predicted, in turn, by research anxiety (β = −.10, β = −.23, β = −.13). Moreover, interest was predicted by age (β = .13). Research anxiety was predicted by age (β = −.10), female gender (β = .28) and the type of baccalaureate allowing admission (β = −.09). This first study on entering BSW students in Switzerland confirms that research orientation and research anxiety should be considered and addressed by teachers when conceptualizing research courses.

Keywords:

• Social work education
• research education
• attitudes
• interest
• predictors
• BSW students
• Switzerland

Introduction

Across many countries, research courses have become a taken-for-granted component of social work study programmes. In the USA, for instance, where a first social work study programme was launched at the University in1920, training in research has also been anchored in the Council of Social Work Education’s Curriculum Policy Statement since 1982 (Lazar, 1991; Montcalm, 1999). In Australia, where social workers have been trained at universities since the Second World War, the Social Work Education and Accreditation Standards require research courses to be part of a social work curriculum (AASW Australian Association of Social Workers, 2015). In Europe, where social work education has been integrated at the academic level (university or university of applied sciences) since the 1970s, research courses also have become a standard in social work programmes (Labonté-Roset, 2005). Significantly, the global standards for the education and training of the social work profession stipulate that students should acquire research skills (IASSW&IFSW, 2004, p. 7).

In Switzerland, from 1972 on, a social work degree was offered on a small scale by one university only while, in general, social work training was offered at a higher vocational level. Social work education was not established on a large scale at university level until 1998. The study programmes were then integrated into the universities of applied sciences, the higher education institutions newly established from 1995 on (Gredig & Truniger, 2005). Since then, bachelor’s and consecutive master’s programmes in social work have included research courses. Relevant points of reference are the requirements implied in the national qualification framework (swissuniversities, 2016) and specified in the national profile on social work education (EDK/FDK, 1999). The specific learning objectives are stipulated in the particular competence profiles created by the schools of social work at the universities of applied sciences.

Just as many social work educators have experienced and reported (see e.g. Bolin, Lee, GlenMaye, & Yoon, 2012, p. 226), teachers in Bachelor of Social Work (BSW) programmes in Switzerland hear students question the sense and adequacy of the research-related course work which forms an integral part of the required courses in their study programme. Looking back to the time when we started offering the programmes at university level, we have the impression that students’ attitudes towards research have changed and the utilization of scientific knowledge and reference to social work research findings have now become widely accepted. Thus, we would not go as far as to pretend that social work students would respond to research courses with ‘groaning, moaning, eye-rolling’ or ‘hyperventilation’ as Irwin Epstein put in his much-noticed article (Epstein, 1987, p. 71). Nonetheless, we still have the feeling that students are somewhat apprehensive and reluctant to engage with research courses.

In the social work literature, there are a number of suggestions as to how research teachers might meet this challenge and generate students’ appreciation of and enthusiasm for research (see e.g. Berger, 2002; Bolen, 2006; Harder, 2010; Jacobson & Goheen, 2006; Kapp, 2006; Knee, 2002). There are also trials evidencing that teachers can effectively encourage students and help them to manage their anxiety (see e.g. Einbinder, 2014; Maschi, Wells, Yoder Slater, MacMillan, & Ristow, 2013). This knowledge could have informed us about an adequate organization of the research courses, a suitable didactic approach and favourable attitudes of teachers. However, we did not want to initiate the further development of our research courses without a sound assessment of the current situation. We held that our own impressions of our students’ attitudes towards research courses alone would not provide a sufficient foundation for such an endeavour. Thus, cautioned by Lazar’s (1991) findings that social work students’ attitudes towards research content in the curriculum can be more favourable than perceived by faculty and practitioners, we set out to look for studies substantiating a valid view of students’ attitudes towards research courses. Recent studies confirmed that social work students were ‘apprehensive about research courses and about the process of learning research’ (Morgenshtern, Freymond, Agyapong, & Greeson, 2011, p. 552). A comparative study by Green, Bretzin, Leininger, and Stauffer (2001) also showed that BSW students attributed less importance to research for practice and reported more research anxiety than among students of psychology or business. Some authors did not explicitly support the view that social work students were ‘reluctant’. They did however report on, for example, students’ moderate interest in research courses, concluding that findings regarding the perceived importance of research for social work practice were ‘disappointing’ and attesting to students a ‘lack of enthusiasm for research courses’ (Bolin et al., 2012, p. 238). Maschi et al. (2007) observed that students’ negative thoughts and anxiety about the research process decreased during the course of a research course. Although positive, this finding indirectly affirms that students initially had negative feelings and experienced anxiety. Other authors, for example Secret, Ford, and Rompf (2003), reported higher degrees of interest in and moderate fear of research courses among social work students. They interpreted their data in a more positive way and concluded that the extant literature might overstate students’ negative attitudes towards research. Nevertheless, the students in their sample reported a fear of research.

In short, the studies produced ambiguous results. From this somewhat uneven ground, we may conclude—as Morgenshtern et al. (2011) summed it up regarding graduate students—that ‘social work students are anxious and frequently have negative feelings toward research’ (Morgenshtern et al., 2011, p. 555).

Several studies identified factors contributing to the reported attitudes and apprehensiveness. The appeal of research courses covaried with BSW students’ research orientation in terms of the perceived importance and the attributed usefulness of research for social work practice as well as the perceived unbiased nature of research. It was also associated with age and gender. Other correlates were fear of research courses, statistics anxiety, statistical knowledge, computer anxiety, social work empowerment or self-efficacy (for more recent findings see e.g. Bolin et al., 2012; Green et al., 2001; Secret et al., 2003).

In order to develop an adequate understanding, Bolin et al. (2012) modelled the factors influencing social work students’ interest in research courses. Their model posited that both students’ research interest (interest in research courses) and their research anxiety (fear of research courses) were predicted by students’ research orientation. Their study confirmed this relationship and showed that research anxiety was also predicted by students’ self-efficacy while research interest was predicted by students’ self-efficacy and age.

From this, we could derive that entering students enrolled in a BSW programme in Switzerland might also have negative attitudes towards research courses and experience adverse feelings such as concern, worry and strain. Further, we could assume that these feelings were determined mainly by their research orientation.

However, such a conclusion could be risky. The available studies investigated populations of undergraduate and graduate students ranging from BSW students entering their first research course (Secret et al., 2003), to third-year students attending the last compulsory research course (Lazar, 1991) or from BSW and MSW students having successfully completed up to two research courses (Bolin et al., 2012) to graduate students (Morgenshtern et al., 2011) or groups including research course-experienced and inexperienced students (Maschi et al., 2007). However, we did not find studies investigating students entering a BSW programme.

Further, the studies to which we had access investigated entering students enrolled in programmes in the USA, Canada and Israel. We did not find studies about entering BSW student’s attitudes towards research courses in programmes in any European countries, let alone Switzerland.

Thus, we felt that it was not sound to draw direct conclusions from the available body of knowledge for students entering a BSW programme in Switzerland. In agreement with Secret et al. (2003), we instead preferred to refrain from immediate generalizations and to ask entering BSW students directly about their feelings regarding the upcoming research courses.

The curricular context

The Swiss education system implements the structures and principles of the European Higher Education Area. Applicants for a BSW programme must have obtained a general, professional or specialized baccalaureate, professional college diploma, intermediate diploma school certificate or undergo a specific admission procedure. Additionally, they have to document work experience of at least one year including a minimum of six months experience in social work (pre-study internship).

The BSW programme should prepare BSW students for the critical utilization of research findings and scientific knowledge. For this purpose, the programme of the school under investigation includes a strand of four required modules directly related to research, techniques of scientific work, social scientific research methods and research skills.

The first module provides students with basic notions of science and research. They are requested to attend this module in their first semester. The second module introduces students to qualitative and quantitative research methods. The third element of the research-related course work is integrated in the casuistic courses. Inspired by a case, students choose a topic and write a scientific paper on theory relevant to it. Finally, all bachelor students write a thesis, which is a monographic scientific paper on a topic chosen by the students. More in-depth research teaching continues in the consecutive master’s programme, which includes five required research-related modules.

This curriculum introduces research and research skills to the students from the beginning of their studies. The early introduction and continuous research teaching is an expression of the understanding underpinning the programme: the engagement with research is conceived as an integral part of the (basic) professional skills of social workers rather than a matter of refinement or specialization for students with specific career plans like the ‘icing on the cake’ as Lorenz (2003) tellingly put it. In this way, the research teaching is placed in a framework of application. Teachers give numerous examples. They repeatedly highlight how scientific knowledge and other forms of knowledge are intertwined in professional practice. They illustrate the importance of research findings for courses of action and show on many occasions how social work research may lay the foundations for social change.

This BSW programme is available on two campuses at a distance of 45 kilometres from each other and situated in diverging socio-cultural and demographic contexts. It can be completed in three modalities: full time, part time with field placements in blocks or part time with a concurrent field placement.

Study objectives and research question

Aims

Given the difficulties regarding generalizing extant research findings for social work students enrolled in a BSW in Switzerland, we aimed to gain a greater understanding of first-year students’ attitudes towards research courses in order to provide us with a strong foundation for the further development and adjustment of the research teaching in our programme. The present study therefore aimed to assess students’ interest in research courses. Further, it aimed to gain an understanding of factors influencing BSW students’ attitudes towards research-related course work and to contribute to the development of a model explaining the research interest in students entering BSW.

Modelling BSW students’ interest in research courses

We designed a model of the relationship between interest in research courses, research orientation and fear of research courses. The three variables had proved to be of importance in previous research. The model drew mainly on the work by Bolin et al. (2012) but, in contrast, modified the constellation of the implied variables by assigning fear of research courses a different position. The extant social work literature considers the experience of anxiety as a factor impacting students and in particular their learning and performance (see, for instance, Baird, 2016). Accordingly, we assumed that fear of research courses would be a factor that fuels a self-distancing from the frightening stimulus, leads to depreciation of research and, finally, low research orientation. Thus, the model posits that students’ fear of research courses predicts students’ interest in research courses directly as well as mediated by their research orientation. Further, it considers that age, gender, study modalities, the type of qualifications allowing admission to the programme, and campus could have an influence on fear of research courses as well as on research orientation (see Figure 1).

Figure 1. The hypothetical model explaining BSW students’ interest in research courses.

Objectives

Against this background, the objectives of this study were

(a)	to assess the interest in research courses, the fear of research courses and the research orientation of students entering a BSW programme in Switzerland, and
(b)	to test the proposed model explaining these students’ interest in research courses.

Hypotheses

Drawing on the model presented above, we hypothesized that

(a)	students’ interest in research courses was predicted directly by student’s fear of research courses and their research orientation,
(b)	student’s research orientation was predicted by their fear of research courses
(c)	the implied control variables (age, gender, study modality, the type of credentials allowing admittance to the programme, and campus) had an influence on fear of research courses as well as on research orientation.

Methods

Design and data collection

The study had a cross-sectional design and was based on data collected from three cohorts of students entering a BSW programme in three consecutive years. Data were collected by means of an anonymous, standardized self-administered questionnaire. The questionnaire was constructed using ‘EFS Survey’ (Quest Back AG) and was accessible online with an open access link.

Sampling

We gathered a convenience sample of entry-level students enrolled in a BSW programme offered by a major school of social work in Switzerland. It included entering students of the years 2014, 2015 and 2016. In 2014, we asked the students on both campuses to complete the questionnaire during the first class of the required course on fundamental concepts of (social work) science and research in the very first week of their bachelor’s programme. In 2015, we launched the survey during the then piloted introduction event for entrants at the North Campus in the week before the semester and the programme commenced. In 2016, we surveyed the entering students on both campuses during the introduction event in the week before the programme began.

In accordance with national legislation, we did not have to seek formal approval from an ethical committee.

Questionnaire development and measures

The questionnaire opened with information about the study objectives and instructions. It reaffirmed that participation was voluntary, confirmed anonymity, and explained to the participants that in sending in the questionnaire they gave their consent to the inclusion of their data in the analysis.

To measure the independent variable ‘interest in research courses’ we drew on an instrument developed and validated by Bolin et al. (2012). We included six items that were applicable to the situation of entering students who were anticipating attending research courses. We did not consider three items that referred to lived experiences of attending research courses and the utilization of skills developed as a result of completing such course work. The items presented the beginning of a sentence (an example of an item was: ‘When I imagine that I must read research textbooks or articles, I am …’) and the respondents were asked to complete the sentence with a response from a five-point Likert scale ranging from 1 = ‘strongly interested’ to 5 = ‘strongly disinterested’. The construct had a high internal consistency. Cronbach’s α was .812.

During data analysis, we recoded the scale. In order to facilitate understanding of the findings, we reverse coded the values with the effect that high values expressed high interest (5 reads ‘strongly interested’). The construct was calculated using the mean of the responses to the corresponding items.

Students’ research orientation was operationalized by the perceived importance and the attributed usefulness of research for social work practice, and the perceived unbiased nature of research. To measure these constructs, we drew on the measures validated by Bolin et al. (2012).

The perceived importance of research for social work practice was measured by five items (an example for an item was: ‘Social workers should base their practice on knowledge gained from research’) with a 5-point Likert response scale ranging from 1 = ‘strongly disagree’ to 5 = ‘strongly agree’. The internal consistency of this construct was not fully satisfactory. Cronbach’s α was .60.

In the original study by Kirk and Rosenblatt, Cronbach’s α of this scale was .65 (Kirk & Rosenblatt, 1981) while it was .75 in the study by Bolin et al. (2012). The analysis of our data showed that the inter-item correlations of the five items were rather weak and ranged from r = .36 to r = .12. However, the exclusion of any item would have resulted in an even smaller Cronbach’s alpha. A reduction to three items based on content did not result in better internal consistency (Cronbach’s α = .55). Finally, we performed an explorative factor analysis which included the five items, in order to test the structure of the measure. The principal component analysis, using oblimin rotation, extracted only one factor with an eigenvalue higher than one. Specifically, the eigenvalue of this factor was 1.96 and explained 39.71% of the variance (Kaiser–Meyer–Olkin measure = .69; Barlett’s test p ≤ .001). The five items loaded with values ranging from .55 to .74. We interpreted this result as confirmation that the instrument measured one construct. Thus, we decided to accept the low internal consistency and include the construct in this form in the further analysis. The construct was calculated using the mean of the responses to the corresponding items.

The attributed usefulness of research for social work practice was captured by five items (e.g.: ‘Social work research is not particularly useful to practitioners providing direct services’) with a five-point Likert response scale ranging from 1 = ‘strongly disagree’ to 5 = ‘strongly agree’. As all items were in a negative form, we reverse coded the values with the effect that high values expressed high levels of attributed usefulness. The internal consistency of this construct was not fully satisfactory. Cronbach’s α was .53.

In the original study by Kirk and Rosenblatt, Cronbach’s α of this scale was .71 (Kirk & Rosenblatt, 1981). In the study by Bolin et al. (2012) the scales had a Cronbach’s α of .83 for usefulness. The analysis of our data showed that the inter-item correlations of the five items were weak, ranging from r = .25 and r = .21. Again, the exclusion of any item would have resulted in an even smaller Cronbach’s alpha. A reduction to three items based on content did not result in better internal consistency (Cronbach’s α = .43). The principal component analysis included the five items. Using oblimin rotation, the analysis extracted only one factor with an eigenvalue higher than one. The eigenvalue of this factor was 1.78 and explained 35.56% of the variance (Kaiser–Meyer–Olkin measure = .69; Barlett’s test p = ≤ .001). The five items loaded with values ranging from .48 to .69. We interpreted this result as a confirmation of the measurement model. Thus, again, we decided to accept the low internal consistency and include the construct in this form in the further analysis. The construct was calculated using the mean of the responses to the corresponding items.

To measure the perceived unbiased nature of research we used six items (e.g.: ‘Many research findings are skewed in order to appeal to funding resources’) with a five-point Likert response scale ranging from 1 = ‘strongly disagree’ to 5 = ‘strongly agree’. As all items were in a negative form, we reverse coded the values with the effect that high values expressed high levels of attributed usefulness. The internal consistency of this construct was acceptable. Cronbach’s α was .71. The construct was calculated using the mean of the responses to the corresponding items.

Fear of research courses was operationalized in terms of worries, concerns and strain (Baird, 2016) in the face of research courses and involving ‘fear that one would not be able to cope with the demands of the course’ (Bolin et al., 2012, p. 226). It was assessed using six items. The items presented the beginning of a sentence (an example of an item was: ‘If I imagine that I must attend a lecture about research, my feelings of strain are …’) and the respondents were asked to complete the sentence with a response from a 5-point Likert scale ranging from 1 = ‘extremely strong’ to 5 = ‘extremely weak’. The construct had a high internal consistency. Cronbach’s α was .87. We recoded the values with the effect that high values expressed a high level of fear. The construct was calculated using the mean of the responses to the corresponding items.

Further, the questionnaire captured the age of the students and the gender they identified with, the campus where the students were enrolled (North Campus vs. South Campus), the programme modality students had chosen (full time, part time with field placement in blocks, part time with concurrent field placement), and the qualifications that allowed admission to the BSW programme.

The questionnaire was in German. When drawing questions, items and response categories from original questionnaires in English, we translated them into German. To verify the outcome we used a back-translation procedure. In order to avoid position effects, we had the items measuring a specific construct displayed in a randomized sequence. Further, we programmed the questionnaire to fit for laptop computers, tablet computers or smart phones. We tested the questionnaire in a pre-test with 35 second-year students to verify its comprehensibility and usability. In the light of the feedback, we slightly adjusted the wording of the questionnaire. In the pre-test with second-year students in 2014, the internal consistency of all scales reached an acceptable level.

Data entry and analytic strategy

Data were imported from the EFS-Survey into IBM SPSS 22. First, we performed descriptive analyses. This included testing the normal distribution of the scores on the construct variables. Based on Shapiro–Wilk’s test of normality, we concluded that the data on the construct variables were non-normally distributed. Second, we determined the correlations of these variables using bivariate analysis (i.e. Spearman’s rho). To make comparisons between groups, we used the Mann–Whitney U Test (Abu-Bader, 2011). For multivariate analysis, we used multiple standard regression analysis. Third, and in order to test the hypothesized predictor model, we performed a path analysis. Variables that were measured on a nominal level (gender, campus, study modality and the credentials for admission) were transformed into dummy variables with the values 0 and 1. The three study modalities were transformed into two dummy variables: the value of the variable was 1 for ‘studies with concurrent field placement’ and 0 for full time and part time studies. The credentials were transformed into six variables. For instance, the value was 1 for having been awarded a ‘professional baccalaureate’ and 0 for all other credentials. The variable labels displayed in the tables and graphs relate to the value 1. (For instance, the value of the variable ‘gender’ was 1 for female and 0 for male; the variable is labelled ‘gender: female’.)

Structural equation modelling requires complete data-sets. Missing data result in the exclusion of a number of participants, which would entail bias. In order to avoid this, we performed a multiple imputation as suggested by Lüdtke, Robitzsch, Trautwein, and Köller (2007), including all variables in the predictor model.

Preparing structural equation modelling, we performed a series of multiple standard regressions (method: enter). In the first analysis, interest in research courses was the dependent variable. The independent variables were the construct variables constituting research orientation and fear of research courses as well as the control variables age, gender, study modality, type of credential allowing admittance to the BSW programme and campus. In consecutive steps, we performed three regression analyses. In each analysis, we regressed one of the three construct variables constituting research orientation on fear of research courses and the above-mentioned control variables as independent variables. Finally, we regressed fear of research courses on the above-mentioned control variables.

Structural equation modelling was performed using the Generalized Least Square Estimates method in IBM AMOS 22. We included the variables and modelled the paths according to the results of the previous multiple regression analyses. Certain variables had turned out not to have a significant association with either interest in research courses or the variables constituting research orientation or fear of research courses. These variables therefore did not form part of a causal path and were not included into the model. Thus, campus and study modality were excluded.

We entered the remaining eligible variables into the equation at the same time and did not carry out modifications.

We used several indices for the assessment of model fit. Given that each index has its flaws, we considered various tests and observed whether the different indices would point to the same conclusion. We used the Goodness of Fit Index (GFI) and the Adjusted Goodness of Fix Index (AGFI) that adjusts for model complexity (i.e. number of parameters). Both indices are adequate for path models and considered satisfactory when >.90. We also consulted the Chi-Square (in AMOS referred to as CMIN). A value of 0 represents a perfect fit. The related p-value should be >.05. (Nieuwenboom, 2008, p. 244). We also calculated the normed Chi-Square, which is the CMIN-degree of freedom (df)-ratio. Given the sample size, this ratio should result in a value smaller than 2 (Ullman, 2001). As these indices are affected by sample size, we also considered the standardized root-mean-square residual (SRMR). It might correct a possible overstatement of fit yielded by the other indices used and is especially suitable for the GLS method. Its value should be <.08. (Browne & Cudeck, 1993; Hu & Bentler, 1998). Finally, we also wanted to express in an accessible way the extent to which the model was able to explain the variance in the dependent variable. For this purpose, we calculated the coefficient of determination, R squared, and adjusted it to the number of predictors of the dependent variable (adj. R²). It yielded the percentage of explained variance in interest in research courses.

Results

Sample description

In total, 708 entering students administered and sent in a questionnaire: 307 (43.4%) in 2014, 105 (14.8%) in 2015 and 296 (41.8%) in 2016. Given that the total number of students enrolled in the BSW programme offered by the same school in these years was 760, the response rate was 93.16%. Participants’ ages ranged from 19 to 56 years with a mean age of 25.6 years (SD = 6.24) and a median age of 23 years. Table 1 displays further characteristics of the sample.

Table 1. Characteristics of participants (N = 708).

Variable	Descriptor	n	%	N
Gender				690
	Female	501	27.4
	Male	189	72.6
Credentials for access to the BSW programme				701
	Professional baccalaureate	230	32.8
	General baccalaureate	131	18.7
	Specialized baccalaureate	232	33.1
	Professional college diploma	55	7.8
	Intermediate diploma school certificate	10	1.4
	Other	43	6.1
Study modality				702
	Full-time study	275	39.2
	Part-time study	168	23.9
	Study with concurrent field placement	259	36.9
Campus				706
	South Campus	444	62.9
	North Campus	262	37.1

Interest in research courses

Respondents’ interest in research courses varied widely. Scores ranged from 1.5 to 5 on a scale ranging from 1 to 5. With a mean score of 3.6 (SD = .55), a median score of 3.67, and a modal value of 4, students reported somewhat elevated levels of interest. Scores were not normally distributed (Shapiro–Wilk Test = .97, p ≤ .001).

Among the items measuring interest in research courses, the item with the highest mean score referred to the student’s interest in asking a lecturer a research-related question (M = 3.98; SD = .60). The item with the lowest mean score referred to the interest in writing a research paper (M = 3.2; SD = .92).

A comparison of the reported interest in research courses among the three cohorts of entering students showed a significant difference between the groups beginning the programme in 2014 and 2016. Among the students entering in 2016 the median score for interest was significantly higher, which is more positive than that among the students entering in 2014 (for details, see Table 2).

Table 2. Comparisons between the cohorts entering 2014 and 2016.

Variable	Cohort	N	Median	Mann–Whitney U	p
Interest in research courses	2014	307	3.52
	2016	296	3.67	38905.0	.002
Perceived importance of research	2014	307	3.08
	2016	296	3.20	40102.0	.012
Attributed usefulness of research	2014	307	3.30
	2016	296	3.40	40326.0	.016
Perceived unbiased nature of research	2014	307	3.25
	2016	296	3.33	42501.0	.168
Fear of research courses	2014	307	3.13
	2016	296	3.00	38111.5	.001

Research orientation and fear of research courses

Research orientation included the perceived importance and the attributed usefulness of research for social work practice. The third component was the perceived unbiased nature of research.

The perceived importance of research for social work practice obtained balanced but generally moderate scores. It covered the full range from 1 to 5. The mean score was 3.12 (SD = .49), the median score 3.2. Scores were not normally distributed (Shapiro–Wilk Test = .98, p ≤ .001).

The two items with the highest mean scores stated the view that social workers should base their practice on knowledge gained from research (M = 3.78; SD = .68) and should always keep abreast of research in their field of practice (M = 3.72; SD = .78). The item with the lowest score referred to the view that a major part of social work education should be research training (M = 2.49; SD = .82).

Compared with students who had entered in 2014, students entering in 2016 reported significantly higher median scores for the perceived importance of research for social work practice. Their evaluation of the importance of research was more positive than that of their colleagues entering the programme two years before (for details, see Table 2).

The attributed usefulness of research for social work practice obtained higher scores then the other components of research orientation. They ranged from 2 to 5 on a scale ranging from 1 to 5. The mean score was 3.34 (SD = .51) and the median score was 3.4. Scores were non-normally distributed (Shapiro–Wilk Test = .99, p ≤ .001).

Among the items measuring the attributed usefulness, the item with the highest mean score referred to the statement that social work research is useful to practitioners providing direct service (M = 3.83; SD = .71). The item with the lowest mean score conveyed the view that research findings had limited applicability to complex practice situations (M = 2.52; SD = .94).

The analysis of the entering cohorts revealed that students entering in 2016 attributed significantly higher levels of usefulness to research than the cohort of students entering in 2014 (for details, see Table 2).

The perception of research as unbiased obtained scores with a range from 1.83 to 5 on a scale ranging from 1 to 5. The mean score was 3.28 (SD = .49) and the median score was 3.3. Scores were non-normally distributed (Shapiro–Wilks Test = .99, p ≤ .001).

Among the items measuring the perceived unbiased nature of research, the item with the highest mean score expressed the view that social work research often examines significant questions (M = 3.66, SD = .73) while the item with the lowest mean score stated that many research findings were skewed in order to appeal to funding sources (M = 2.98;SD = .85).

Comparative analysis of the scores of the student cohorts entering in 2014 and 2016 did not reveal a significant difference in the median level of the perceived unbiased nature of research (for details, see Table 2).

The three components of research orientation showed moderate correlations. The perceived importance and the attributed usefulness of research correlated with r_s = .28 (p = ≤ .001) the attributed usefulness and the perceived unbiased nature correlated with r_s = .39 (p = ≤ .001) the perceived importance and the perceived unbiased nature correlated with r_s = .20 (p = ≤ .001).

The scores for fear of research courses covered the full range of the scale from 1 to 5. The mean score was 3.01 (SD = .68) and the median was 3. The scores on this scale were non-normally distributed neither (Shapiro–Wilk Test = .99, p ≤ .001).

Among the items assessing worries, concerns and strain, the item with the highest mean score asked the respondents to imagine having to complete an in-class quiz on research (M = 3.77; SD = .93). The item with the lowest mean score evoked a situation in which the respondents had to listen to another student explain a research topic (M = 2.38; SD = .76).

A comparison of the reported levels of fear of research courses showed a significant difference between the groups beginning the programme in 2014 and 2016. The mean scores for fear reported by the students entering in 2016 was significantly lower than that of the students entering in 2014 (for details, see Table 2).

Table 3 provides an overview of the descriptors of the five construct variables in the model.

Table 3. Description of the construct variables (N = 708).

Variable	Scale	Minimum/Maximum	Range	Mean	SD	Median
Interest in research courses	1–5	1.5/5	3.5	3.6	.55	3.67
Perceived importance of research	1–5	1/5	4	3.12	.49	3.20
Attributed usefulness of research	1–5	2/5	3	3.34	.51	3.40
Perceived unbiased nature of research	1–5	1.83/5	3.17	3.28	.49	3.30
Fear of research courses	1–5	1/5	4	3.01	.68	3.00

Findings from the testing of the model

The analysis shows that participants’ interest in research courses was predicted by the perceived importance of research for social work practice (β = .27, p ≤ .001), the attributed usefulness of research for social work practice (β = .15, p ≤ .001), and the perceived unbiased nature of research (β = .08, p = .023). On all these variables, the higher the participants scored, the higher was the reported interested in research courses. The interest in research courses was also directly predicted by fear (β = −.29, p ≤ .001). The higher the reported level of fear, the lower was the reported interest in research courses. The fifth predictor of interest in research courses was students’ age (β = .13, p ≤ .001). Notably, the strongest predictor of the interest in research courses was fear, which turned out to have both the above-mentioned direct effect and an indirect effect mediated through the three components of research orientation. Fear of research courses proved to be a predictor of the perceived importance of research for social work practice (β = −.10, p = .009), the attributed usefulness for social work practice (β = −.23, p ≤ .001) and the perceived unbiased nature of research (β = −.13, p ≤ .001). In other words, higher levels of fear of research courses were associated with lower scores on the three components of research orientation.

The perceived importance of research was also predicted by the respondents’ gender (β = −.11, p = .005). Female participants reported lower scores on the perceived importance of research for social work practice. The attributed usefulness of research for social work practice was predicted by age (β = .12, p ≤ .001). Older respondents attributed higher degrees of usefulness to research then younger participants did.

The reported level of fear of research courses in turn was also predicted by the participants’ gender (β = .28, p ≤ .001) and age (β = −.10, p = .003). Female respondents reported higher levels of fear of research courses while older respondents reported less fear. The third but minor predictor of fear was a specific qualification allowing access to the BSW programme (β = −.09, p = .013). Entering students who had obtained a general baccalaureate reported less fear of research courses then students with a different type of secondary education.

Regarding fit, all the indices pointed to a good model fit. Both the GFI (.997) and the AGFI (.988 yielded values well above .90, which implied a very good model fit. CMIN was 7.568(df = 8) and the p-value .477. The normed Chi-Square (CMIN/df) was .946. These indices also denoted a good fit. Finally, SRMR was .0156 and considerably lower than the suggested threshold of .08. Based on these indices, we conclude that the model showed a good fit and, thus, is plausible. The adjusted R² was .312. The predictors established in the model explained 31.2% of the variance of the interest in research courses. The standardized regression weights implied are available in Table 4. Figure 2 provides a visualization of the paths described here.

Table 4. Results of the path analysis including all variables in the equation.

Variable			β	Confidence interval (95%)		p Value
				Lower	Upper
Fear of research courses	←	Gender: female	.284	.205	.367	≤.001
Fear of research courses	←	Age	−.105	−.186	−.031	.003
Fear of research courses	←	General baccalaureate	−.089	−.161	−.016	.013
Perceived importance of research	←	Gender: female	−.106	−.185	−.039	.005
Perceived importance of research	←	Fear of research courses	−.101	−.173	.008	.009
Attributed usefulness of research	←	Fear of research courses	−.233	−.305	−.169	≤.001
Perceived unbiased nature of research	←	Age	.095	.025	.165	.012
Perceived importance of research	←	Age	.067	.003	.138	.072
Attributed Usefulness of Research	←	Age	.121	.37	.196	≤.001
Perceived unbiased nature of research	←	Fear of research courses	−.130	−.214	−.029	≤.001
Interest in research courses	←	Perceived importance of research	.269	.201	.345	≤.001
Interest in research courses	←	Attributed usefulness of research	.155	.069	.223	≤.001
Interest in research courses	←	Age	.132	.069	.202	≤.001
Interest in research courses	←	Fear of research courses	−.292	−.358	−.219	≤.001
Interest in research courses	←	Perceived unbiased nature of research	.078	.009	.152	.023

Notes:

N = 708; Method: Generalized Least Squares Estimates; GFI = .997, AGFI = .988; X² = 7.568, p = .477; CMIN/df = .946; SRMR = .0156; adj. R² = .312.

Figure 2. Visualization of the path model.

Notes: N = 708; Method: Generalized Least Squares Estimates; GFI = .997, AGFI = .988; X² = 7.568 p = .477; CMIN/df = .946; SRMR = .0156; adj. R² = .312. Non-significant standardized coefficients (p > .050) and correlations among the predictors are not displayed.

Discussion

The assessment of entering students’ attitudes towards research courses evidences somewhat elevated scores regarding their interest in research courses. This echoes the findings by Secret et al. (2003) regarding undergraduate students enrolled in an introductory research course at a university in a south-eastern state of the USA. At the same time, the reports of the students participating in this study confirm the noteworthy prevalence of worries, concerns and strain in relation to research courses. This corresponds with the findings of many previous studies regarding diverse student populations in the USA and Canada (Bolin et al., 2012; Einbinder, 2014; Green et al., 2001; Maschi et al., 2007; Morgenshtern et al., 2011; Secret et al., 2003).

The data of this sample found support for the hypothesized model explaining interest in research courses among entering students of a BSW programme in Switzerland. The analysis confirmed the expected relations between the participants’ interest in research courses, their research orientation and the reported level of fear of research courses (hypotheses a and b). Higher levels of worries, concerns and strain were associated with lower scores on the three components of research orientation, namely the perceived importance of research for social work practice, the attributed usefulness of research for social work practice and the perceived unbiased nature of research. Lower scores on research orientation in turn were associated with less interest in research courses. Further, analyses evidence that fear also directly influences the level of interest in research courses. Thus, the data from this sample provide evidence for the expected direct and mediated impact of fear on interest in research courses.

However, the data did not fully support the assumptions regarding the demographic and study-related characteristics (hypothesis c). Students’ age turned out to be an important variable as it predicted fear of and interest in research courses as well as two components of research orientation. The association with the perceived importance of research, however, was not significant. All the same, these findings regarding age allow for the conclusion that fear of research courses, research orientation and interest in research courses show an age gradient. Gender, on the other hand, seems less influential. In this sample, there is a comparatively strong association of fear of research courses and gender. However, two of the components of research orientation and, worthy of note, interest in research courses were not associated with gender. The qualification allowing access to the BSW programme impacted exclusively fear of research courses and with a small effect. The study-related characteristics of the respondents considered in the model were not significantly associated with the construct variables. The modality of the study programme (implying strong differences regarding the degree of involvement in practice) and the campus (implying differences in the socio-cultural and economic context of the school of social work concerned) did not covary with fear of research courses, research orientation or interest in research courses.

In contrast to the findings by Bolin et al. (2012), data from this sample support the understanding that research orientation does not predict fear of research courses. Rather, they support the view that fear of research courses is a predictor of research orientation and interest in research courses. We interpret this finding as a confirmation of the assumption underlying our model that fear is an important driver of perceptions regarding research as well as the readiness and openness to engage with research-related curricular content rather than an effect of research orientation.

This study has some limitations. We gathered an extensive sample including 93.15% of three cohorts of entering students of one major school of social work. Nevertheless, it is restricted to 1 of the 10 BSW programmes available in Switzerland. Thus, generalizations can be made with caution only.

Second, although we used the scales developed and tested by Bolin et al. (2012), the internal consistency of the scales measuring the perceived importance and attributed usefulness of research for social work practice was not satisfactory. Given the analysis described in the method section, we speculate that the rather weak internal consistency of the two scales could be due to the fact that the views of entering students on these specific issues might still not be highly profiled. In contrast to students participating in previous research using these measures, the population investigated in this study had not yet been exposed to the programme content and the teacher’s expectations. In the first few days of the BSW programme, it seems likely that their views on the importance and usefulness of research for social work practice were not yet founded in knowledge and experience, or reflected and clearly contoured in a way that produces highly consistent answers. For future research on entering students, it will be important to develop measures with good psychometric qualities.

Regarding future research, we advocate continuing the endeavour to explore and model entering students’ attitudes relating to research courses. Given the many comments on the limited generalizability of the findings in previous studies as well as in this investigation, we recommend samplings including students from various programmes in the Swiss higher education area. Further, we suggest comparing students in programmes in different higher education areas living and studying in diverse societal, professional and educational contexts.

The programmes in Switzerland do not focus on the use of findings from quantitative studies or restrict their teaching to quantitative methods for social research. Nonetheless, statistics are a component of the study programmes. Further, we have to be aware that secondary education seldom provides students with a notion of qualitative research. Mentioning research courses could trigger a certain statistics anxiety in entering students and influence their feelings and attitudes regarding research courses in a BSW programme. Therefore, when it comes to modelling entering students’ attitudes towards research courses in the future we also recommend including variables such as general self-efficacy (Montcalm, 1999) or feeling empowered (Secret et al., 2003), which have proven to be associated with research interest in studies on students in the USA and Canada. Given the findings on statistics anxiety in students in general and in social work graduates (Davis, 2003; Onwuegbuzie & Wilson, 2003), we also suggest including this variable.

This study was the first in Switzerland to investigate BSW students’ attitude towards research interest. It affirms faculty members’ perception that students have a certain apprehension regarding engaging in research-related course work and allows for the assumption that students enrolled in BSW programmes in Switzerland do not differ from their peers enrolled in programmes in North America or Israel. The study also contributes to a deeper understanding of the dynamics of fear of research courses, interest in research courses and the research orientation of BSW students.

Although conclusions have to be cautious and more research is needed, we may already learn from this study that research teachers also have to be aware of the fact that students may experience worry, concern and strain regarding research courses. It is not enough to highlight and demonstrate the relevance, importance and utility of research for social work practice as we have done in our curriculum. It also seems to be important to address students’ fear of research courses although students might declare such a feeling less frankly than disinterest or disregard for research. As Baird (2016) highlights, anxiety does not always have a debilitating effect on the learning processes and outcomes. However, research teachers should support students in managing their anxiety so that it reaches a level that instigates their growth and does not hinder the development of their skills. There is evidence that, among other things, fear of research courses can be addressed, for instance, by acknowledging these feelings openly and repeatedly (Einbinder, 2014). Thus, the present assessment illustrates the potential of an assessment of the student cohorts we teach. It may open ways to further develop and improve our research teaching in accordance with current students’ needs.

Notes on contributors

Daniel Gredig, PhD and trained social worker, is a professor of social work and head of the MSW programme. His recent work focused on HIV prevention, research-based intervention development, stigma and discrimination against people living with diabetes mellitus and social work students’ attitudes.

Annabelle Bartelsen-Raemy, Lic Phil, studied educational sciences and educational psychology. She is an assistant head of the MSW Programme. Her recent work focused on parenthood and concepts of parenthood, stigma and discrimination against people living with diabetes mellitus, entering BSW students’ motivation to study social work and social work students’ attitudes.

Disclosure statement

We, the authors of this study, would like to declare that we teach in the programme in which the students under investigation are enrolled. This said, we also state that we do not perceive a conflict of interest or feel conflicting loyalties when assessing entering student’s attitudes towards courses they will attend in future.

Ethical approval

In accordance with national legislation, the authors of this study did not have to seek formal approval from an ethical committee.

Funding

This work was funded by the School of Social Work at the University of Applied Sciences and Arts Northwestern Switzerland.

Acknowledgements

We thank the participants of this study. We appreciate their readiness to share their views and experiences.