Assessing the learning environment of a faculty: Psychometric validation of the German version of the Dundee Ready Education Environment Measure with students and teachers

Abstract

Aims: The teachers’ perspectives of the educational environment have as yet only been sparsely considered. This study aimed at validating the first German version of the Dundee Ready Education Environment Measure (DREEM) from the points of view of both students and teachers.

Methods: Data from 1119 students and 258 teachers were available for analysis. Psychometric validation included the analysis of homogeneity and discrimination at item level as well as reliability (Cronbach's α), criterion and construct validity at test level. Effect sizes were calculated and the independent samples t-test was used for statistical inference testing of mean differences between two groups.

Results: Item characteristics were satisfactory in both samples. Reliability was high with α = 0.92 (students) and 0.94 (teachers), respectively. Factor analyses revealed five dimensions which slightly diverged from the five subscales postulated by the DREEM authors though. The environment was evaluated significantly (p < 0.001) more positively by teachers (M = 117.63) than by students (M = 109.75). Further significant differences were observed with regard to gender, mother language, stage of studies and previous professional training among others.

Conclusions: With convincing psychometric properties at item and test levels, the suitability of DREEM not only for students but also for teachers to assess the educational environment has been demonstrated.

Introduction

Knowledge about the teaching and learning environment is an important prerequisite for the successful conception and implementation of a new curriculum (Bouhaimed et al. 2009) and should become a part of educational institutions’ good practices (Soemantri et al. 2010). Although the educational environment is rather a fuzzy construct which makes a binding definition difficult (Genn & Harden 1986), it can be deduced from the numerous attempts at definition that here not only external parts in the sense of an objective environment but also internal parts in the sense of a personality trait are important. The individual perception of one and the same objective fact is known to vary widely (Dornic & Ekehammar 1990). Furthermore, the individual perception of the learning environment by the student has a significant impact on the educational results realised in this learning climate: Desirable learning outcomes and gains are positively associated with favourable aspects of socio-psychological learning environments, i.e. the better the environment, the better the results (Haertel et al. 1981). Besides the performance behaviour, the student's gender also correlates with his/her perception of the environment (Chaput de Santonge & Dunn 2001): Some authors found female students to assess their learning environment more negatively than male students (Mayya & Roff 2004), others, however, found the opposite results, i.e. female students evaluating the climate more positively than their male counterparts (Bassaw et al. 2003). Apparently, female and male students often experience their learning environments differently (Roff et al. 2001) which is, in light of an increasing proportion of female students not only in Germany but also worldwide, of interest for the traditionally male-dominated medical teaching facilities (Genn 2001b).

The teachers’ perspectives of the educational environment have as yet only been sparsely considered in spite of the many curriculum reforms that have been initiated worldwide (Fraser 1986; Genn 2001a, 2001b; Miles & Leinster 2009). This is even more remarkable since curriculum changes represent a challenge for the teachers and, just for this group of people, an evaluation appears to be particularly important (Till 2005).

In German (medical) educational research, the educational environment has received only little attention and there is a lack of suitable measurement tools in the German language. An international, widely used and validated measurement tool that has been translated into several other languages is the Dundee Ready Education Environment Measure (DREEM; Roff et al. 1997; cf. Soemantri et al. 2010). The test construction included a Delphi panel of almost 100 international health-care educators and the combination of quantitative with qualitative test construction techniques. The questionnaire was thus conceived as a non-culturally specific instrument for assessing the teaching and learning environment in the perception of students of the health-care professions (Roff 2005). Up to now, it has been used only once to compare staff and student perceptions of the learning environment at a given medical school. To explore potential perceptional discrepancies between these two groups, Miles and Leinster (2009) administered DREEM to students and – in a reworded version – to teachers who were thereby able to complete the questionnaire with their opinion about the student experience. Although the overall results were closely aligned, there were significant differences at subscale level, i.e. teachers perceived themselves and learning more positively than students did, whereas students’ social perception of themselves was more positive than in the teachers’ view. In addition, the results demonstrated staff's unfamiliarity with student aspects of the learning environment. In its present form, DREEM consists of 50 items that must be answered using a scale of 0 (“strongly disagree”) to 4 (“strongly agree”); thus, a maximum of 200 points can be achieved. A high number of points is indicative of a good environment. The items are divided into the five subscales: perception of teaching, perception of teachers, academic self-perception, perception of the atmosphere and social self-perception.

The aim of this study is to psychometrically validate DREEM for the first time in the German-speaking area as a standardised questionnaire for assessing the learning environment from the points of view of both students and teachers.

Table

Practice points

The suitability of DREEM not only for students but also for teachers to assess the educational environment has been demonstrated. The exploration and communication of discrepancies on group immanent self- and other-perception for students and teachers create chances for the improvement of the learning environment, in particular, the student–teacher relationship. More investigation of the factorial structure of DREEM, the relationship between the perception of climate by means of DREEM and academic performance and its retest reliability with use of an adequate retest interval would be of interest.

Methods

A German version of DREEM was completed online by 1119 medical students and 258 teachers of the Medical Faculty of the Heinrich-Heine-University (HHU) Duesseldorf. In addition, the students and teachers were asked extensive questions about their demographic background.

First, a German translation of DREEM was prepared using the back-translation method of Brislin (1970) as modified by Jones et al. (2001). This was then translated back by two independent, bilingual experts, discrepancies were discussed and a final version agreed upon. For the preparation of the teachers’ version of the questionnaire, items formulated in the original in the “I” form (e.g. “I am encouraged to participate actively in the classes”) were adapted for teachers according to the procedure of Miles and Leinster (2009; e.g. “The students are encouraged to participate in class”; see Appendix 1).

Medical education in Duesseldorf currently follows a traditional curriculum that is divided into a 2-year, preclinical period (science and basic subjects) and a 4-year clinical period. The first state examination is held at the end of the preclinical period while the second and final state examination is held at the end of the clinical period.

Data collection was carried out as an online procedure at the end of the summer semester 2010 for all students and teachers. Participation was voluntary and the data were stored in anonymous form. Details of demographic variables were collected prior to the questionnaire. Hereby the students were, among others, asked to provide information about their school leaving certificates and their grades in the first state examination.

Altogether, questionnaires from 1119 students (=55.0%) and 258 teachers were analysed. As it was not known who in the faculty was actually really involved in teaching, the questionnaires were first sent to all scientific personnel in the faculty (N = 1294). Staff members who could be safely ignored as not being involved in teaching, such as clerks, technicians or nurses, were omitted from the very beginning. The accompanying letter was addressed specifically to teachers and demographic questions on the topic “teaching activity” were posed. Even if staff members, such as scientific personnel solely involved in research, had begun to complete the questionnaire, they would not have been able to provide pertinent answers to most questions because of their lacking teaching experience and would have therefore been rather likely to cancel the survey. This is why we only evaluated completed surveys and additionally run post hoc plausibility data checks to identify respondents – both students and staff – who had not answered the questions in a meaningful way. Thanks to these precautions, it can thus be safely assumed that the questionnaires we analysed were only completed by personnel actually involved in teaching. Because of these circumstances, however, no statements about the proportion of all teaching staff who participated are possible. Table 1 contains demographic information on both samples.

Table 1.  Demographic data of the student and the teacher sample

		Students (N = 1119)	Teachers (N = 258)
Gender	Female	739 (66.0%)	72 (27.9%)
	Male	380 (34.0%)	186 (72.1%)
Age	M (SD)	24.1 years (3.8)	41.7 years (9.5)
Mother tongue	German	985 (88.0%)	249 (96.5%)
	Other	134 (12.0%)	9 (3.5%)
Mark in school leaving examination^a	M (SD)	1.8 (0.6)	–
Stage of course: studying (students)/teaching (teachers)	Pre-clinical^b	509 (45.8%)^c	30 (13.4%)
	Clinical	602 (54.2%)	152 (67.9%)
	Both	–	42 (18.7%)
If clinical, mark first state examination*	M (SD)	2.7 (0.7)	–
Have you completed a professional occupational training (e.g. as a nurse)?	Yes	241 (23.2%)	–
	No	796 (76.8%)
Do you live in Duesseldorf?	Yes	836 (75.0%)	144 (56.9%)
	No	278 (25.0%)	109 (43.1%)
How far away from school do you live?	M (SD)	10.2 km (17.1)	13.2 km (17.2)

Notes: M (SD) = mean value (standard deviation).

^aThe marks in the school leaving examination and in the first state examination range from 1 (“excellent”) to 5 (“failed”).

^bPreclinical = semesters 1–4; Clinical = semester 5–12.

^cThe frequencies do not always add up to 1119 and 258, respectively, as some participants have not replied to all demographic questions. Thus, due to these missing values the percentages are not always based on the total sample size of N = 1119 and N = 258, respectively.

Within the framework of item and test analysis, besides the mean values and the discrimination indices of the items, also item homogeneity as well as the reliability and validity of the entire test were to be examined.

A control of the administration and scoring objectivity was not necessary because the whole questionnaire contained standardised instructions and answering possibilities. Thanks to the interpretation guidelines for this questionnaire by Lai et al. (2009) as well as McAleer and Roff (2001), the interpretation objectivity of DREEM was also high and thus did not need to be examined. Content validity did not require any examination either because DREEM was developed by experts in a Delphi process.

• Homogeneity: Explorative factor analyses were performed to analyse the homogeneity of the items and the factorial validity of DREEM.

• Discrimination: Item discrimination was assessed by means of part-whole-corrected discrimination indices.1

• Reliability: Examination of the reliability was limited to an analysis of the internal consistency by means of Cronbach's α.

• Criterion validity: With regard to criterion validity, we investigated both predictively (on the basis of the school leaving grades) and concurrently (on the basis of the grades in the first state examination) whether good students evaluated the environment better than did their weaker counterparts (Pimparyon et al. 2000; Sun 2003; Mayya & Roff 2004; Carmody et al. 2009) and whether the perception of the environment became more negative with increasing number of semesters, even when age was held constant (Hutchins 1964; Till 2004; Zaini 2005; Bouhaimed et al. 2009; Riquelme et al. 2009). Furthermore, possible relationships with mother language and place of residence were studied. Also, an answer to the question of the discrepancy between the evaluation of the environment by the students in comparison to the evaluation by the teachers was sought in an undirected manner, since the few available results on this subject are inconsistent (Sheehan 1970; Fraser 1986; Miles & Leinster 2009) and allow the assumption of a series of difficultly predictable factors. Finally, an evaluation of possible gender differences was undertaken.

• Construct validity: With regard to construct validity, we wanted to examine whether DREEM, also in its German version, exhibits the five-factorial structure (“perception of teaching”, “perception of teachers”, “academic self-perception”, “perception of atmosphere”, and “social self-perception”) as postulated by the authors (Roff et al. 1997) but as yet hardly ever tested (de Oliveira Filho et al. 2005; Wang et al. 2009).

All data were analysed with the help of the programme Statistical Package for the Social Sciences 17.0 for Windows (2008). Effect sizes were calculated with the freely available programme G*Power 3 (Faul et al. 2007). For inference testing of mean differences between two groups (e.g. students vs. teachers), the independent samples t-test was used. Pearson's correlations were calculated to examine relationships. Cohen's (1988) guidelines were used for the interpretation of effect sizes. For the effect size measure d in the t-test, it holds that d ≥ 0.20 = small effect, d ≥ 0.50 = medium effect, d ≥ 0.80 = large effect. The product–moment correlation coefficient r, on the other hand, is already a measure of effect size, it holds that r ≥ 0.10 = small effect, r ≥ 0.30 = medium effect, r ≥ 0.50 = large effect.

Explorative factor analyses were conducted to analyse the homogeneity of the items and the factorial validity of the employed questionnaire. Besides theoretical expectations, the Kaiser–Guttman criterion (Gorsuch 1983) and the scree test (Cattell 1966) were used as criteria for the extraction of factors.

Results

Item analysis

DREEM item means dispersed sufficiently in both the student and teacher samples. In the student group, the total item mean amounted to M = 2.19 (SD = 0.50), only just above the limit of 2, which in DREEM indicates a region worthy of improvement at item level (Vieira et al. 2003; Whittle et al. 2007). Altogether, the means for 20 items were less than 2. For the teachers, the mean values were somewhat less dispersed and on the whole somewhat higher with a total mean of M = 2.35 (SD = 0.35). In this group, only six items had a mean value of less than the limit of 2. In both samples, there were no ceiling or floor effects at item level.

With regard to homogeneity, for students all, and for teachers most, of the items loaded highly (>0.30) on the dimensions identified by us by means of factor analyses (Appendix 2). Exceptions in the teachers’ questionnaire were items 24 (“The students are able to memorise all that is needed”), 38 (“The students are able to concentrate well”) and 42 (“Cheating in exams is a problem at this university”), which did not exhibit satisfactory loadings (i.e. >0.30) on any of the identified dimensions.2 The part-whole-corrected discrimination indices for the students were all positive and amounted on average to M = 0.42 (SD = 0.13); in the teacher sample, the discrimination indices were on average M = 0.46 (SD = 0.16; see Table 2).

Table 2.  Mean values, standard deviations and part-whole-corrected discrimination indices of DREEM items in the student and the teacher sample.

		Students (N = 1119)		Teachers (N = 258)
Subscale	Item	M (SD)	DI	M (SD)	DI
Perception of teaching	1 I am encouraged to participate in class	1.98 (0.95)	0.50	2.73 (0.94)	0.48
	2 The teaching is sufficiently concerned to develop my confidence	1.95 (1.05)	0.62	2.13 (0.91)	0.68
	3 The teaching encourages me to be an active learner	1.90 (1.01)	0.58	2.21 (0.95)	0.71
	4 The teaching is well focused	2.05 (0.93)	0.63	2.15 (0.84)	0.58
	5 The teaching is sufficiently concerned to develop my competence	2.39 (0.92)	0.61	2.61 (0.79)	0.59
	6 I am clear about the learning objectives of the course	2.72 (0.89)	0.50	2.31 (0.90)	0.62
	7 The teaching is often stimulating	1.82 (0.93)	0.60	2.52 (0.88)	0.54
	8 The teaching time is put to good use	1.90 (0.97)	0.49	2.45 (0.82)	0.62
	9 The teaching is student centred	1.71 (0.96)	0.57	2.10 (0.94)	0.55
	10 Long-term learning is emphasized over short term	1.33 (1.06)	0.45	1.87 (0.94)	0.61
	11 The teaching is too teacher-centred	1.73 (0.88)	0.26	2.05 (0.98)	0.37
	12 The teaching over-emphasizes factual learning	1.43 (1.05)	0.25	1.91 (1.07)	0.35
Perception of teachers	13 The teachers are good at providing feedback to students	1.45 (0.91)	0.49	2.04 (0.93)	0.57
	14 The teachers have good communications skills with patients	2.06 (0.73)	0.45	2.57 (0.78)	0.54
	15 The teachers are knowledgeable	2.87 (0.70)	0.35	2.70 (0.87)	0.29
	16 The teachers give clear examples	2.41 (0.81)	0.52	2.63 (0.72)	0.58
	17 The teachers are well prepared for their classes	2.39 (0.92)	0.36	2.57 (0.86)	0.52
	18 The teachers provide constructive criticism here	1.88 (0.91)	0.55	2.36 (0.82)	0.54
	19 The teachers ridicule the students	2.56 (0.95)	0.41	3.47 (0.72)	0.47
	20 The teachers get angry in class	2.71 (0.90)	0.32	3.10 (0.88)	0.39
	21 The teachers are authoritarian	1.64 (0.94)	0.21	2.41 (0.95)	0.41
	22 The teachers are patient with patients	2.19 (0.77)	0.42	2.52 (0.74)	0.40
	23 The students irritate the teachers	2.16 (0.95)	0.30	2.73 (0.93)	0.43
Academic self-perception	24 I am able to memorise all I need	2.17 (1.10)	0.25	2.21 (0.93)	−0.01
	25 Much of what I have to learn seems relevant to a career in medicine	2.28 (1.03)	0.50	2.41 (1.00)	0.44
	26 I feel I am being well prepared for my profession	1.97 (1.04)	0.63	1.76 (0.89)	0.66
	27 Last year's work has been a good preparation for this year's work	2.08 (1.00)	0.45	2.12 (0.73)	0.43
	28 My problem-solving skills are being well developed here	1.70 (0.90)	0.56	2.21 (0.83)	0.66
	29 I am confident about passing this year	2.71 (0.90)	0.27	2.59 (0.66)	0.35
	30 I have learned a lot about empathy in my profession	1.86 (1.10)	0.45	1.83 (0.91)	0.51
	31 Learning strategies which worked for me before continue to work for me now	2.60 (0.93)	0.33	2.39 (0.87)	0.34
Perception of atmosphere	32 The atmosphere is relaxed during lectures	2.66 (0.83)	0.47	2.75 (0.73)	0.58
	33 I feel able to ask the questions I want	2.48 (0.95)	0.40	2.85 (0.73)	0.50
	34 I feel comfortable in class socially	2.70 (0.85)	0.50	2.52 (0.68)	0.59
	35 There are opportunities for me to develop interpersonal skills	2.20 (1.07)	0.48	2.41 (0.83)	0.35
	36 The enjoyment outweighs the stress of studying medicine	2.05 (1.10)	0.52	2.22 (0.78)	0.57
	37 The atmosphere motivates me as a learner	1.84 (0.99)	0.63	2.25 (0.76)	0.73
	38 I am able to concentrate well	2.65 (0.90)	0.30	2.22 (0.83)	0.31
	39 The atmosphere is relaxed during ward teaching	2.29 (0.72)	0.37	2.66 (0.76)	0.35
	40 This school is well timetabled	1.49 (1.08)	0.43	1.70 (0.92)	0.57
	41 I find the experience disappointing	1.83 (1.18)	0.41	2.04 (0.91)	0.53
	42 Cheating is a problem in this school	2.73 (1.04)	0.08	2.51 (0.84)	0.19
	43 The atmosphere is relaxed during seminars/tutorials	2.45 (0.98)	0.50	2.87 (0.73)	0.53
Social self-perception	44 I have good friends in this school	3.19 (0.91)	0.25	2.24 (0.79)	0.47
	45 There is a good support system for students who get stressed	1.02 (0.88)	0.44	1.89 (0.78)	0.51
	46 I am too tired to enjoy this course	1.78 (1.17)	0.39	2.17 (0.97)	0.35
	47 I am rarely bored on this course	2.78 (1.12)	0.19	2.05 (0.95)	0.38
	48 My accommodation is pleasant	3.30 (0.83)	0.23	2.11 (0.53)	0.09
	49 My social life is good	3.01 (0.96)	0.35	2.44 (0.67)	0.14
	50 I seldom feel lonely	2.60 (1.07)	0.33	2.08 (0.61)	0.14

Notes: M (SD) = mean value (standard deviation); DI = part-whole-corrected discrimination index.

Items in italics are keyed negatively. After reversing the polarity for analysis, a low value now corresponds to prior consent and therefore a poor teaching and learning environment.

The validated German version of the DREEM can be obtained from the authors on request.

Table 3.  DREEM total and subscale scores in the student and the teacher sample.

	Students (N = 1119)				Teachers (N = 258)
Questionnaire	Maximal score	Internal consistency (α)	M (SD)	Min–Max	Internal Consistency (α)	M (SD)	Min–Max	t(1375)	p	d
DREEM – Total score (50 items)	200	0.92	109.75 (21.71)	28–182	0.94	117.63 (20.80)	52–182	−5.30	<0.001	0.37
Perception of teaching (12 items)	48	0.84	22.92 (6.91)	2–47	0.86	27.05 (6.85)	5–47	−8.66	<0.001	0.60
Perception of teachers (11 items)	44	0.75	24.33 (5.08)	5–44	0.80	29.10 (5.32)	12–44	−13.46	<0.001	0.92
Academic self-perception (8 items)	32	0.68	17.39 (4.47)	0–31	0.67	17.50 (3.78)	5–26	−0.37	0.71	0.03
Perception of atmosphere (12 items)	48	0.75	27.39 (6.04)	5–48	0.80	29.01 (5.40)	13–45	−3.96	<0.001	0.28
Social self-perception (7 items)	28	0.57	17.71 (3.68)	2–26	0.53	14.97 (2.76)	6–23	11.26	<0.001	0.84

Note: α = Cronbach's α; M (SD) = mean value (standard deviation); Min–Max = minimal and maximal scores achieved; p = empirical α-level of two-tailed testing; t(1375) = t-value with 1375 degrees of freedom; d = effect size for independent sample t-test with d ≥ 0.20 – small effect, d ≥ 0.50 – medium effect and d ≥ 0.80 – large effect.

Table 4.  DREEM total and subscale scores in the student sample depending on students’ marks in the school leaving examination and the first state examination.

	Students with school leaving examination					Students with first state examination
Questionnaire	<1.60 (N = 518) M (SD)	>1.60 (N = 557) M (SD)	t(1073)	p	d	<2.50 (N = 308) M (SD)	>2.50 (N = 275) M (SD)	t(581)	p	d
DREEM – total score	111.26 (21.06)	108.74 (22.01)	1.92	<0.05	0.16	109.81 (21.82)	109.46 (22.43)	0.19	0.43	0.02
Perception of teaching	23.24 (6.78)	22.76 (6.89)	1.15	0.13	0.07	21.69 (7.11)	22.41 (7.08)	−1.22	0.11	0.10
Perception of teachers	24.56 (4.93)	24.19 (5.13)	1.18	0.12	0.07	24.32 (5.24)	24.13 (5.49)	0.42	0.38	0.04
Academic self-perception	17.68 (4.34)	17.18 (4.51)	1.87	<0.05	0.11	17.48 (4.51)	17.20 (4.39)	0.78	0.22	0.06
Perception of atmosphere	27.86 (5.90)	27.06 (6.12)	2.15	<0.05	0.13	28.28 (5.87)	27.93 (6.24)	0.70	0.24	0.06
Social self-perception	17.93 (3.50)	17.54 (3.85)	1.72	<0.05	0.11	18.04 (3.59)	17.80 (3.89)	0.77	0.22	0.06

Note: The marks in the school leaving examination and in the first state examination range from 1 (“excellent”) to 5 (“failed”).

p = empirical α-level of one-tailed testing; t(1073) and t(581) = t-value with 1073 and 581 degrees of freedom, respectively; d = effect size for independent samples t-test with d ≥ 0.20 – small effect, d ≥ 0.50 – medium effect and d ≥ 80 – large effect.

Table 5.  Association between DREEM total and subscale scores and stage of course in the student sample.

			Students
Questionnaire	Correlation with number of semester (r)	(Partial correlation)	Preclinical (N = 509) M (SD)	Clinical (N = 602) M (SD)	t(1109)	p	d
DREEM – Total score	−0.07*	(−0.01)	109.82 (21.07)	109.71 (22.28)	0.08	0.47	0.01
Perception of teaching	−0.18**	(−0.13***)	23.91 (6.52)	22.08 (7.14)	4.42	<0.001	0.27
Perception of teachers	−0.09**	(0.07**)	24.48 (4.59)	24.20 (5.40)	0.91	0.18	0.06
Academic self-perception	−0.04	(0.02)	17.40 (4.46)	17.38 (4.48)	0.07	0.47	0.00
Perception of atmosphere	−0.05	(0.11***)	26.56 (5.89)	28.12 (6.07)	−4.32	<0.001	0.26
Social self-perception	−0.03	(0.10***)	17.46 (3.64)	17.93 (3.71)	−2.10	0.05	0.13

Note: Partial correlation = correlation between respective DREEM score and number of semesters after partialling out the factor age; M (SD) = mean value (standard deviation); p = empirical α-level of one-tailed testing; t(1109) = t-value with 1109 degrees of freedom; d = effect size for independent samples t-test with d ≥ 0.20 – small effect, d ≥ 0.50 – medium effect and d ≥ 0.80 – large effect; *p < 0.05; **p < 0.01; ***p < 0.001 in the respective t-test for (partial) correlations (one-tailed).

Table 6.  DREEM total and subscale scores in the student sample depending on students’ mother tongue.

	Students
Questionnaire	Native speakers (N = 985) M (SD)	Non-native speakers (N = 134) M (SD)	t(1117)	p	d
DREEM – Total score	109.04 (21.68)	114.95 (21.24)	−2.97	<0.01	0.28
Perception of teaching	22.54 (6.86)	25.77 (6.60)	−5.14	<0.001	0.48
Perception of teachers	24.10 (5.02)	25.98 (5.25)	−4.03	<0.001	0.36
Academic self-perception	17.15 (4.49)	19.19 (3.83)	−5.01	<0.001	0.49
Perception of atmosphere	27.34 (6.08)	27.75 (5.70)	−0.74	0.46	0.07
Social self-perception	17.91 (3.60)	16.26 (3.99)	4.92	<0.001	0.43

Note: M (SD) = mean value (standard deviation); p = empirical α-level of two-tailed testing; t(1117) = t-value with 1117 degrees of freedom; d = effect size for independent samples t-test with d ≥ 0.20 – small effect, d ≥ 0.50 – medium effect and d ≥ 0.80 – large effect.

Table 7.  Association between DREEM total and subscale scores and distance between respondents’ residence and university (in km) in the student and teacher sample.

	Students (N = 1119)	Teachers (N = 258)
Questionnaire	Correlation with distance between residence and university (r)
DREEM – Total score	0.02	−0.13*
Perception of teaching	0.05	−0.12
Perception of teachers	0.07*	−0.06
Academic self-perception	−0.01	−0.09
Perception of atmosphere	−0.01	−0.14*
Social self-perception	−0.03	−0.19**

Note: *p < 0.05; **p < 0.01 in respective t-test for correlations (two-tailed).

Table 8.  DREEM total and subscale scores depending on students’ previous occupational training.

	Students with completed previous occupational training (N = 241)	Students without completed previous occupational training (N = 972)
Questionnaire	M (SD)	M (SD)	t(1035)	p	d
DREEM – Total score	104.67 (22.48)	111.14 (21.04)	−4.11	<0.001	0.30
Perception of teaching	21.66 (7.24)	23.32 (6.72)	−3.31	<0.001	0.24
Perception of teachers	23.76 (5.47)	24.51 (4.88)	−2.02	<0.05	0.15
Academic self-perception	16.47 (4.75)	17.63 (4.32)	−3.55	<0.001	0.26
Perception of atmosphere	25.79 (5.87)	27.83 (5.95)	−4.67	<0.001	0.35
Social self-perception	16.99 (3.89)	17.86 (3.62)	−3.20	<0.01	0.23

Note: M (SD) = mean value (standard deviation); p = empirical α-level of two-tailed testing; t(1035) = t-value with 1035 degrees of freedom; d = effect size for independent samples t-test with d ≥ 0.20 – small effect, d ≥ 0.50 – medium effect and d ≥ 0.80 – large effect.

Table 9.  Gender differences in the assessment of teaching and learning environment in the student sample.

	Male students (N = 380)	Female students (N = 739)
Questionnaire	M (SD)	M (SD)	t(1117)	p	d
DREEM – Total score	107.41 (23.35)	110.96 (20.72)	−2.60	<0.01	0.16
Perception of teaching	21.73 (7.44)	23.54 (6.54)	−4.17	<0.001	0.26
Perception of teachers	23.97 (5.36)	24.51 (4.93)	−1.69	<0.09	0.10
Academic self-perception	17.29 (4.55)	17.44 (4.42)	−0.54	<0.59	0.03
Perception of atmosphere	27.13 (6.30)	27.53 (5.90)	−1.06	<0.29	0.07
Social self-perception	17.29 (4.03)	17.93 (3.48)	−2.79	<0.01	0.15

Note: M (SD) = mean value (standard deviation); p = empirical α-level of two-tailed testing; t(1117) = t-value with 1117 degrees of freedom; d = effect size for independent samples t-test with d ≥ 0.20 – small effect, d ≥ 0.50 – medium effect and d ≥ 0.80 – large effect.

Reliability

At the level of the entire test, the results of the reliability analyses using Cronbach's α were similar in both groups, and DREEM proved highly reliable with α = 0.92 (students) and α = 0.94 (teachers), respectively. At subscale level, the coefficients ranged between 0.53 and 0.86 (Table 3). In accordance with the interpretation guidelines recommended by Lai et al. (2009) as well as McAleer and Roff (2001), the educational environments as measured by means of DREEM exhibited both for the students (M = 109.75, SD = 21.71) and for the teachers (M = 117.63, SD = 20.80) – albeit only marginally – more positive than negative aspects (100 − 149 points). However, this is far removed from an excellent climate (more than 150 points) and lies in the typical region for traditional curricula (<120; Roff 2005).

Criterion validity

For the total DREEM score, the analyses showed significant mean differences between students with a school leaving grade below the median of 1.60 (M = 111.26, SD = 21.06) and those with a school leaving grade above the median of 1.60 (M = 108.74, SD = 22.01), t(1073) = 1.92, p < 0.05 and d = 0.16. With regard to the grade in the first state examination, there were no longer any significant differences between the good and less good students (Table 4).

Students in later semesters (clinical part of course) evaluated the environment as being significantly poorer than did their colleagues in earlier semesters (preclinical part of course) on the DREEM subscale “perception of teaching” (preclinical: M = 23.91, SD = 6.52 vs. clinical: M = 22.08, SD = 7.14; t(1109) = 4.42; p < 0.001; d = 0.27). In contrast, the atmosphere in the clinical stage was considered to be significantly better (preclinical: M = 26.56, SD = 5.89 vs. clinical: M = 28.12, SD = 6.07; t(1109) = −4.32; p < 0.001; d = 0.26]. The significant negative correlations between number of semesters and evaluation of the environment (e.g. “perception of teaching”: r = −0.18, p < 0.001) remained significant even after partialling out the factor age which, as expected, correlated positively with the number of semesters (r = 0.47, p < 0.001), but, however, did become weaker (“perception of teaching”: r* = −0.13, p < 0.001; see Table 5).

In the student sample, native speakers tended to evaluate the environment (e.g. DREEM total score: M = 109.04, SD = 21.68) more negatively than non-native speakers (M = 114.95, SD = 21.24), t(1117) = −2.97, p < 0.01 and d = 0.28. The non-native speakers also considered themselves to be significantly more positive in academic self-perception in DREEM (M = 19.19, SD = 3.83) than the native speakers (M = 17.15, SD = 4.49), t(1117) = 5.01, p < 0.001, d = 0.49, and this even though – in the subgroup of students who had already completed the first state examination – the non-native speakers (M = 2.92, SD = 0.71) achieved significantly poorer results than the German native speakers (M = 2.66, SD = 0.74), t(581) = −2.77, p < 0.01, d = 0.36 (Table 6).

Students who did not live in Duesseldorf (M = 25.05, SD = 4.94) evaluated the teachers in DREEM significantly better than those students who lived in Duesseldorf (M = 24.11, SD = 5.11), t(1112) = 2.68, p < 0.01, d = 0.18. The further the teachers lived away from the university, the lower was their total score in DREEM (r = −0.13, p < 0.05) and the poorer did they evaluate the atmosphere in DREEM (r = −0.14, p < 0.05) as well as the social self-perception of the students (r = −0.19, p < 0.01; see Table 7).

Students who had completed a professional training programme, e.g. in nursing or geriatric care (DREEM total score: M = 104.67, SD = 22.48), evaluated the environment in all dimensions of DREEM as being significantly poorer than did their fellow students without a completed profession training (M = 111.14, SD = 21.04), t(1098) = −4.11, p < 0.001, d = 0.30 (Table 8).

Comparison of student perspectives versus teacher perspectives

On comparison between teachers and students, the overall environment was evaluated significantly more positively by our teachers (DREEM total score: M = 117.63, SD = 20.80) than by our students (M = 109.75, SD = 21.71), t(1375) = −5.30, p < 0.001, d = 0.37. In both groups, however, the DREEM total score was significantly poorer than the average score (weighted according to sample size) of the available international studies with medical students (DREEM total score: M = 121.04) in which DREEM has been used so far (see Ostapczuk et al. (2011) for an overview of all studies); students: t(1118) = –17.40, p < 0.001, d = 0.52, and, respectively, teachers: t(257) = −2.63, p < 0.01, d = 0.16.

Examining the different DREEM subscales in this study, teachers’ assessments were more positive than students’ assessments with regard to the subscales “perception of teaching”, “perception of teachers” and “perception of atmosphere”, i.e. our teachers were more satisfied with themselves as teachers, their teaching and the atmosphere in comparison with their students. Students evaluated only their “social self-perception” more positively than teachers did. On the DREEM subscale “academic self-perception”, there was no significant difference between student and teacher perceptions (Table 3).

Gender differences

In the student sample, female students evaluated the educational environment as being significantly more positive in practically all dimensions than did their male counterparts; however, the effects were small (Table 9). In contrast, there were absolutely no over-coincidental gender differences among the teachers. In order to clarify whether or not the gender difference among the students was due to a better academic performance of the female students, male and female students were analysed with regard to their school leaving grades and grades in the first state examination. It was found that female students (M = 1.81, SD = 0.55) on average achieved a significantly better school leaving grade than the male students (M = 1.92, SD = 0.59), t(1073) = −2.97, p < 0.01, d = 0.20, but then performed on average significantly more poorly in the first state examination (M = 2.76, SD = 0.50) than their male counterparts (M = 2.54, SD = 0.80), t(581) = 3.44, p < 0.001, d = 0.33.

Construct validity

In order to examine the factorial validity of DREEM, we investigated in both groups whether the five DREEM dimensions postulated by Roff et al. (1997) could be identified by means of an explorative principal components analysis. Since the Kaiser–Meyer–Olkin (KMO) value (Kaiser 1970, 1974) was 0.93 and Bartlett's (1954) test of sphericity significant, χ²(1225) = 18,561.63, p < 0.001, in the student sample, the factor analysis was carried out.3 After successful factorisation of the 50 DREEM items, there were 10 factors with an eigenvalue > 1; after the analysis of the scree plot and according to expectations, five factors were extracted and subjected to an orthogonal rotation. In total, the five factors explained 41.3% of the variance and exhibited a satisfactory simple structure. However, the five dimensions did not correspond exactly to the five DREEM subscales (Appendix 2): we interpreted them as “teaching and learning”, “self-perception”, “interpersonal relations”, “prevailing social conditions”, and a “method factor”, on which mainly items that were keyed negatively had high loadings. A factor analysis could also be carried out in the teacher sample, KMO = 0.91, Bartlett's test: χ²(1225) = 5066.23, p < 0.001. In analogy to the situation with students, of the original 13 factors with an eigenvalue >1, five components were extracted (variance explanation: 41.4%) and rotated orthogonally. As with the students, the five factors did not fully correspond with the expected subscales: we interpreted them as “teaching and academic perception”, “teachers and atmosphere”, “social perception” and “communication with patients” and, as already the case with the students, a “method factor”. Altogether, some of the extracted factors (e.g. “self-perception” for the students) represented mixed factors of the original scales, whereas others (e.g. “communication with patients” for the teachers) could not be assigned to any of the original DREEM subscales.²

Discussion

In this present investigation, a German version of DREEM (Roff et al. 1997) was subjected to a psychometric validation at item and test levels in a sample of students and teachers of human medicine, respectively. We were able to demonstrate convincing psychometric properties at both item and test levels, so that DREEM is now available in German and can be used to assess the teaching and learning environments for both students and teachers. With regard to other published studies on DREEM, ours involves the as-yet largest investigated samples.

The item means of DREEM were in a similar range as those in international studies in which DREEM had previously been employed (1.50–3.40; Whittle et al. 2007), whereby the dispersion among the teachers was somewhat lower than among the students. Items with a mean <2 indicate regions in need of improvement (Vieira et al. 2003; Whittle et al. 2007). In our samples, 20 items for the students and 6 items for the teachers were in this region.

The low discrimination of the DREEM items 24 (“I am able to memorise all I need”) and 48 (“My accommodation is pleasant”) for the teachers is most probably related to the fact that the teachers did not have the necessary information to answer these questions, so that their responses to these items were less predictive in regard to the DREEM total score. Item 42 (“Cheating is a problem at this university”), on the other hand, had already demonstrated a low discrimination in another study employing DREEM (Wang et al. 2009). Analyses of item discrimination have only been reported in two studies to date. In the above-mentioned work, the item discrimination indices of 49 items were in an acceptable range 0.27–0.77, solely the mentioned item 42 exhibited a discrimination index of <0.20 (Wang et al. 2009). Unfortunately, the authors do not clearly state whether or not they had performed the necessary part-whole corrections of the discrimination index; de Oliveira Filho et al. (2005), on the other hand, reported that the discrimination had been examined, but did not provide any results of these analyses. The homogeneity of our items was, above all, in the student sample very good.

The reliability of DREEM was as yet most frequently tested in the sense of internal consistency by means of Cronbach's α. In the previously published studies, it varied from α = 0.90–0.95 for the entire test and from α = 0.51–0.90 for the five subscales (Mayya & Roff 2004; de Oliveira Filho et al. 2005; O’Brien et al. 2008; Lai et al. 2009; Riquelme et al. 2009; Wang et al. 2009). The retest reliability amounted to r_tt = 0.43 for the entire test at a retest interval of 6 months for the original 50-item version. In our study, the reliability indices (internal consistency) in both samples were comparable and very good.

Criterion validation revealed weak relationships between the evaluation of the teaching and learning climates and the school leaving grades: students with higher grades assessed the climate to be somewhat better than did students with lower grades. In contrast, there were practically no significant associations between perception of the climate and performance in the first state examination. These results are in contradiction to those of other studies in which students with good examination results assessed the climate as being better than did students with poorer examination results (Pimparyon et al. 2000; Sun 2003; Mayya & Roff 2004). One reason for our results could be that the learning environment was on the whole rather negatively evaluated in the present sample and that the expected relationship between a good learning environment and good performance would rather develop in a positive atmosphere. Another reason could be the assessment of performance by means of objective and, above all, written theoretical measures (first state examination), for which Carmody et al. (2009) had previously found no relationship with DREEM in contrast to clinical practical performance. However, this was not confirmed by other authors (Pimparyon et al. 2000; Sun 2003; Mayya & Roff 2004). Further clarification of this situation may be provided by planned investigation with a further German-speaking sample in which the environment will be assessed on the whole as better and/or the grades achieved in the final state examination will be available, as in this examination, also clinical practical performance is being tested.

In our study as well, the perceived learning environment is found to be poorer on average with increasing duration of studies. Students in the clinical phase of their education, however, assessed the atmosphere as being better than did their counterparts in the preclinical stage. Above all, teaching in the clinic received a poorer assessment. The differences remained weakened but still significant when age was partialled out of the correlation between number of semesters and perception of the learning environment; this has been investigated for the first time in our analysis of DREEM. Thus, it can be deduced that the perception of an increasing deterioration of the educational environment is not solely due to the circumstances of education but also to personal factors such as, possibly, “becoming older and more critical”. The initial enthusiasm in medical education seems for many students to decrease in the course of studies, also independent of any concrete bad experiences (Miles & Leinster 2007). These results agree with those of other investigations in which DREEM was employed. These have shown both in cross-section and longitudinally that the perception of the educational environment becomes poorer in the course of (medical) education (Till 2004, 2005; Zaini 2005; Bouhaimed et al. 2009; Riquelme et al. 2009) or at best remains the same (Jiffry et al. 2005; McKendree 2009), and improves only in isolated cases (Pimparyon et al. 2000).

In agreement with Miles and Leinster (2007), another interesting result is that students who had already completed training in another area had a more negative perception of climate than their counterparts without such experience. The possibility of drawing (negative or positive) comparisons with external experiences may possibly affect the impression of the educational climate in medical studies, leading to higher (or lower) expectations. These findings might be taken into account in choosing applicants for medical studies or in curriculum development, for example by designing areas of the curriculum suited to the needs of particular groups.

Similar to the experience of Miles and Leinster (2009), our students perceived themselves to be socially more positive than they were perceived by their teachers, whereas in contrast, the teachers evaluated themselves and their teaching more positively than did the students which could be due to a mutual self-serving bias (Lewicki 1983). In consequence of the transformation of the questionnaire in a teachers’ version, the question regarding content validity arises, as some of the items may not be really relevant to teachers or the teachers might not be in a position to give pertinent responses (e.g., “The students are encouraged to participate in class” or “The students seldom feel lonely”). In fact, some of our teachers reported difficulties in answering some of these items which might be indicative of lacking content validity. On the other hand, items relating to students’ self-perception can provide us with valuable feedback about how well teachers know the reality of students’ lives, feelings and sorrows (cf. Miles & Leinster 2009). The fact that – in contrast to the study of Miles and Leinster (2009) – our teachers assessed both the atmosphere and the entire environment significantly more positively than students did can be interpreted along these lines. Divergent assessments are indicative of the intensity of exchange and communication between these two groups. Thus, the exploration and communication of discrepancies allow both groups to reflect on group immanent self- and other-perception and thereby create chances for the improvement of the learning environment, in particular the student–teacher relationship. The promotion of mentoring programmes (Kalén et al. 2010) and the implementation of effective feedback (Archer 2010), for example, could contribute to the formation of a true community of teachers and students.

In agreement with other studies (Roff et al. 1997; Bassaw et al. 2003; de Oliveira Filho et al. 2005; Bouhaimed et al. 2009), our data confirm that the educational environment is assessed more positively by females in comparison to males, even though the effect is small. Roff (2005) reported in summary that gender differences in the perception of the teaching and learning environment depend on numerous cultural factors and are on the whole smaller when the total satisfaction is high. In contrast, we have detected significant but less relevant differences in spite of a rather lower total satisfaction. Other authors have reported exactly opposing results (Mayya & Roff 2004), i.e. female students were significantly less satisfied with the educational environment than male students in Argentina (Roff et al. 1997, 2001), or no or only minor gender differences were found (Till 2005; Miles & Leinster 2007; Carmody et al. 2009). Cultural differences could be playing a role here. Besides, socio-cultural factors seem to play a role since students whose mother tongue was not German perceived the educational climate more positively according to DREEM. In view of increasing internationalisation of courses, such findings might also be useful in developing a curriculum suited to the individual needs of particular groups.

In view of the large number of international studies in which DREEM was employed, its factorial validity as one facet of construct validity has previously only been rarely investigated. The five subscales described above are based on an explorative factor analysis by Roff et al. (1997) of an Argentinean precursor of DREEM with 58 items. Replications of the five-factorial structure were only partially successful: not only de Oliveira Filho et al. (2005) but also Wang et al. (2009) reported that, although they also found five factors by means of explorative analysis of the 50 DREEM items, these were not absolutely identical with the original five factors. Finally, O’Brien et al. (2008) also found five DREEM factors, although their results cannot be directly compared with the above-mentioned results since the shortened 32-item version of DREEM was used.

In this study, we also could not unambiguously identify the five subscales postulated by the authors. In our investigation, both in the student and the teacher samples, there was one factor that is to be considered as a methodological artefact on which mainly negatively keyed items loaded highly. Such effects occur frequently when a questionnaire contains both positively and negatively keyed items (Paulhus 1984; Paulhus & Reid 1991) and often can only be avoided by one of the following means: (1) exclusive use of positively keyed items and (2) balanced use of equal numbers of positively and negatively keyed items per scale/factor. The first solution option can be criticised in that it makes a questionnaire susceptible to acquiescence effects, i.e. when for all items, a “yes” or “I agree” answer is associated with a high manifestation of the characteristic in question or general agreement, after a certain time, the participants answer more and more frequently and automatically with “yes” or “I agree” (Hinz et al. 2007). On the other hand, the second solution can only be realised when one constructs the questionnaire oneself and not − as is the case in this investigation with DREEM − when one translates the original into another language.

Because a direct comparison with the two other studies (de Oliveira Filho et al. 2005; Wang et al. 2009) in which the five factors could also not be replicated is not possible due to insufficient details, it is not at present reasonable or it would be premature to discard the five original scales. Instead, future research efforts should be directed at verifying the factor structure in various languages and cultural environments.

With convincing psychometric properties at item and test levels, the suitability of DREEM not only for students but also for teachers to assess the educational environment has been demonstrated and it is now available in German for both groups. Both questionnaires provide the foundation for defining future improvements for faculty development. Repeated measurements will enable the assessment of progress in improving the learning environment. In future German and international investigations, a further-reaching psychometric validation of the questionnaire such as, e.g. an investigation of the factorial structure of DREEM, the relationship between the perception of climate by means of DREEM and academic performance and its retest reliability with use of an adequate retest interval would be of interest.

Declaration of interest: The authors have no conflicts of interests with respect to their authorship or the potential publication of this article. We did not receive financial support for the research and/or the authorship of this article.

Notes

1. Item discrimination is determined by correlating the sum of the respective item score (i) with the total score consisting of the sum of all item scores (t). The resulting correlation coefficient (r_it), however, is spuriously high, because the item whose discrimination is to be determined is also part of the total sum. The fewer items a test contains, the larger the resulting bias when calculating item discrimination. To obtain an unbiased item discrimination coefficient, one therefore must correlate the respective item score (e.g. item 30) with the total score minus the contribution of the item in question (i.e. total score based on the sum of items 1–29, without item 30). The resulting coefficient is called a part-whole-corrected item discrimination coefficient (cf. Davis 1958).

2. The loading matrix for the factor analysis of the DREEM items in the teacher sample is available on request from the authors.

3. A KMO value >0.60 or even better >0.75 (Kaiser 1970, 1974) as well as a significant Bartlett's (1954) test show that the variables correlate highly enough with each other to be able to perform a meaningful factor analysis.

Appendices

Appendix 1. Teachers’ version of the questionnaire.

Subscale	Item
Perception of teaching	1	The students are encouraged to participate in class
	2	The teaching is sufficiently concerned to develop the students’ confidence
	3	The teaching encourages the students to be an active learner
	4	The teaching is well focused
	5	The teaching is sufficiently concerned to develop the students’ competence
	6	The students are clear about the learning objectives of the course
	7	The teaching is often stimulating
	8	The teaching time is put to good use
	9	The teaching is student centred
	10	Long-term learning is emphasised over short term
	11	The teaching is too teacher-centred
	12	The teaching over-emphasises factual learning
Perception of teachers	13	The teachers are good at providing feedback to students
	14	The teachers have good communications skills with patients
	15	The teachers are knowledgeable
	16	The teachers give clear examples
	17	The teachers are well prepared for their classes
	18	The teachers provide constructive criticism here
	19	The teachers ridicule the students
	20	The teachers get angry in class
	21	The teachers are authoritarian
	22	The teachers are patient with patients
	23	The students irritate the teachers
Academic self-perception	24	The students are able to memorise all they need
	25	Much of what the students have to learn seems relevant to a career in medicine
	26	The students feel being well prepared for their future profession
	27	For the students, last year's work has been a good preparation for this year's work
	28	The students’ problem-solving skills are being well developed here
	29	The students are confident about passing this year
	30	The students have learned a lot about empathy in their future profession
	31	Learning strategies which worked for the students before continue to work for them now
Perception of atmosphere	32	The atmosphere is relaxed during lectures
	33	The students feel able to ask the questions they want
	34	The students feel comfortable in class socially
	35	There are opportunities for the students to develop interpersonal skills
	36	The enjoyment outweighs the stress of studying medicine
	37	The atmosphere motivates the students as a learner
	38	The students are able to concentrate well
	39	The atmosphere is relaxed during ward teaching
	40	This school is well timetabled
	41	The students find the experience disappointing
	42	Cheating is a problem in this school
	43	The atmosphere is relaxed during seminars/tutorials
Social self-perception	44	The students have good friends in this school
	45	There is a good support system for students who get stressed
	46	The students are too tired to enjoy this course
	47	The students are rarely bored on this course
	48	The students’ accommodations are pleasant
	49	The students’ social life is good
	50	The students seldom feel lonely

Note: For the preparation of the teachers’ version of the questionnaire, items formulated in the “I” form in the original DREEM, were adapted for the teachers according to the procedure of Miles and Leinster (2009).

Appendix 2. Factor analytic loadings of DREEM items in the student sample.

Item	I “Teaching and learning”	II “Self-perception”	III “Inter-personal relations”	IV “Method factor”	V “Prevailing social conditions”
26 I feel I am being well prepared for my profession	0.72
4 The teaching is well focused	0.67
9 The teaching is student centred	0.65
3 The teaching encourages me to be an active learner	0.65
7 The teaching is often stimulating	0.64
8 The teaching time is put to good use	0.61
37 The atmosphere motivates me as a learner	0.61	0.35
17 The teachers are well prepared for their classes	0.61				0.37
40 This school is well timetabled	0.60
27 Last year's work has been a good preparation for this year's work	0.60
5 The teaching is sufficiently concerned to develop my competence	0.59
2 The teaching is sufficiently concerned to develop my confidence.	0.52		0.37
6 I am clear about the learning objectives of the course	0.52
28 My problem-solving skills are being well developed here	0.51		0.45
13 The teachers are good at providing feedback to students	0.46		0.46
10 Long-term learning is emphasized over short term	0.46
16 The teachers give clear examples	0.45
25 Much of what I have to learn seems relevant to a career in medicine	0.44
47 I am rarely bored on this course	0.43				0.35
45 There is a good support system for students who get stressed	0.39
1 I am encouraged to participate in class	0.39		0.37
38 I am able to concentrate well		0.64
50 I seldom feel lonely		0.63
49 My social life is good		0.60
36 The enjoyment outweighs the stress of studying medicine		0.57		0.33
29 I am confident about passing this year		0.56
31 Learning strategies which worked for me before continue to work for me now		0.51
24 I am able to memorise all I need		0.48
46 I am too tired to enjoy this course		0.48		0.47
44 I have good friends in this school		0.47			0.39
34 I feel comfortable in class socially		0.46			0.32
48 My accomodation is pleasant		0.43			0.39
30 I have learned a lot about empathy in my profession			0.62
39 The atmosphere is relaxed during ward teaching			0.55
22 The teachers are patient with patients			0.54
35 There are opportunities for me to develop interpersonal skills			0.53
18 The teachers provide constructive criticism here	0.39		0.50
14 The teachers have good communications skills with patients	0.30		0.47
43 The atmosphere is relaxed during seminars/tutorials		0.35	0.42	0.35
33 I feel able to ask the questions I want			0.40
19 The teachers ridicule the students				0.65
20 The teachers get angry in class				0.59	0.36
21 The teachers are authoritarian				0.59
11 The teaching is too teacher-centred				0.51
12 The teaching over-emphasizes factual learning				0.49
41 I find the experience disappointing	0.33			0.46
23 The students irritate the teachers				0.35
15 The teachers are knowledgeable	0.37				0.49
32 The atmosphere is relaxed during lectures			0.37		0.39
42 Cheating is a problem in this school					0.31

Note: Loadings < 0.30 are suppressed for reasons of clarity. The items in italics are keyed negatively.