Teaching and Teacher Education 90 (2020) 103026

Contents lists available at ScienceDirect

Teaching and Teacher Education

journal homepage: www.elsevier.com/locate/tate

An international comparative study of practicum mentors: Learning

about ourselves by learning about others
Anthony Clarke a, *, Juanjo Mena b
a
Faculty of Education, University of British Columbia, 2125 Main Mall, Vancouver, British Columbia, V6T1Z4, Canada
b
Faculty of Education, University of Salamanca, P
de Canalejas, 169, P.O. Box 37008, Salamanca, Spain

h i g h l i g h t s

 The paper involves mentors from six countries: New Zealand; Thailand; China; Canada; Spain; and Australia.
 The paper represents a shift from largely idiosyncratic within-context analyses to comparative across-contexts analyses.
 The paper demonstrates the Mentoring Profile Inventory as a metric for comparative work.
 The paper highlights distinctive features of mentoring contexts that typically remain hidden within those contexts.
 The paper demonstrates that there are at least three ‘universals’ about mentoring regardless of context.

a r t i c l e i n f o a b s t r a c t

Article history: Teachers who mentor student teachers on practicum play a critical role in teacher education. However, it
Received 26 January 2019 is widely reported that these teachers are poorly prepared and lack sufficient knowledge for mentoring.
Received in revised form And what is known about mentoring is largely drawn from idiosyncratic single-context studies. In an
11 January 2020
attempt to extend the scope of this knowledge, this study draws on the Mentoring Profile Inventory
Accepted 14 January 2020
Available online xxx
(www.mentoringprofile.com) to chart how mentors conceive of their work within and across six in-
ternational contexts. The outcome is a comparative analysis that highlights what is ‘particular’ versus
what is ‘general’ across contexts that has not been previously reported.
Keywords:
Mentoring
© 2020 Elsevier Ltd. All rights reserved.
Supervision
Practicum
Teacher education
Comparative education

1. Context and background practicum mentors often rely on their own past experience as
student teachers to guide their current mentoring practices. Not
Practicum mentoring is a special form of teaching situated in the surprisingly, they often mentor as they were mentored. However,
immediacy of the action setting. Indeed, practicum mentors, often simply replicating ‘the past’ seriously limits what mentors might
called cooperating teachers, are deeply implicated in the develop- offer today’s student teachers (Sarason, 1996). And just accepting
ment of their profession or as Lave and Wenger (1991) argue: “in the status quo is of little comfort to many student teachers who
the generative process of producing their own future” (p. 57). Un- leave the profession within their first five years decrying the in-
fortunately, the literature also shows that practicum mentors often adequacy of their practicum in preparing them for a career in
receive little or no assistance in preparing for this work (Hansford, teaching (Buchanan et al., 2013).
Ehrich, & Tennant, 2004). This neglect is one of the most persistent Attrition rates for beginning teachers range from 5% in European
challenges in Teacher Education. countries, such as Germany and France, up to 30% in places like the
In the absence of professional development opportunities, United Kingdom and America (Lindqvist, Norda €nger, & Carlsson,
2014; Ronfeldt, Loeb, & Wyckoff, 2013). As Hudson (2007) argues,
we continue to squander our considerable investment in pre-
* Corresponding author.
service teacher education if we neglect one of the most important
E-mail addresses: [email protected] (A. Clarke), [email protected] elements of that experience: the practicum mentors. Underscoring
(J. Mena). the significance of this inattention is that student teachers

https://doi.org/10.1016/j.tate.2020.103026
0742-051X/© 2020 Elsevier Ltd. All rights reserved.
2 A. Clarke, J. Mena / Teaching and Teacher Education 90 (2020) 103026

universally regard their practicum mentors as critical to their suc- large, then something particular is at play in that jurisdiction, with
cess both on practicum and later within the profession (Izadinia, respect to that outcome, and therefore is worthy of closer scrutiny.
2015). To address this shortcoming at least two things must Such scrutiny might not have been obvious if that outcome had not
happen: been placed against the backdrop of the population at large. For
example, what might over time have been considered a relatively
(1) individual practicum mentors must better understand why mundane motivator for mentors in a particular jurisdiction might
they do what they do so that what motivates or challenges be revealed through a cross-context study as being highly signifi-
them in this work can be made explicit and available for cant for mentors in that jurisdiction and may have been neglected
critical review and reflection; and to the detriment of recruitment efforts in that jurisdiction. Cross-
(2) the teaching profession, collectively, must better understand context studies provide the ability to flag these outcomes within
what might be typically expected or what might be partic- contexts, something that is not possible through other means.
ularly distinctive about the motivations and challenges that However, cross-context studies of mentoring are largely con-
practicum mentors encounter in this work so that they can spicuous by their absence in the literature and with good reason.
respond accordinglydsomething which is difficult to do in These studies are inherently more difficult to undertake because
the absence of such knowledge. they require ready access to multiple jurisdictions. Also, cross-
context studies typically require some sort of ‘common metric’
The former is often taken up in the mentoring literature in terms that will allow for comparisons to be made with a degree of con-
of within-context studies that shed light on mentoring within single fidence (Alexander, 2001; Hauser, 2016). The cross-context study of
jurisdictions. For example, the literature indicates that commit- practicum mentors undertaken in this paper draws on a metric
ment to being a practicum mentor arises from at least three deliberately developed for this purpose: the Mentoring Profile In-
sources: ventory or MPI (Clarke, Collins, Triggs, & Neilsen, 2012). As such,
this study offers the potential for an important shift in the men-
 a commitment to pupils in terms of wanting to ensure the best toring literature from what has been largely idiosyncratic within-
possible teachers for students (Feiman-Nemser, 2001; Kent, context analyses to comparative cross-context analyses.
2001); The MPI is a 62-item online instrument that allows mentors to
 a commitment to the profession in terms of wanting to give back indicate what motivates or challenges them in their work with
to and ensure the continuing development of the profession student teachers. The MPI then renders participants’ responses in
(Kitchel & White, 2007; Sinclair, Dowson, & Thistleton-Martin, terms of eight motivator scales and six challenge scales:
2006); and
 a commitment to self in terms of being exposed to new ideas and Motivator #1 (M1): Renewing the Profession (4 select items)
strategies through engagement with student teachers (Clarke, Motivator #2 (M2): Improving My Own Teaching Practices (4
2006; Koskela & Ganser, 1998). select items)
Motivator #3 (M3): Student Teachers Promote Pupil Engage-
As such, these commitments represent important motivators for ment (4 select items)
being a practicum mentor and represent a starting point for Motivator #4 (M4): ‘Time Out’ to Monitor Pupil Learning (4
exploring the beliefs and assumptions that underlie their practice select items)
(Ambrose, Bridges, DiPietro, Lovett, & Norman, 2010). However, Motivator #5 (M5): Contributing to Teacher Education (4 select
very little is written about these commitments beyond the general items)
attributions outlined above and, as such, they tend to remain Motivator #6 (M6): Reminders about Career Development (4
relatively abstract concepts. select items)
The latterdto better understand what might be typical versus Motivator #7 (M7): Developing a Professional Community (4
distinctive about what mentors encounter in particular con- select items)
textsdrequires cross-context studies. Cross-context studies allow Motivator #8 (M8): Mentoring in Classroom Contexts (4 select
for the identification of issues that might otherwise remain hidden items)
or normalized from a single-context perspective. Borrowing from Challenge #1 (C1): Challenges in Guidance and Mentoring (6
Van Maanen (1995), cross-context analyses provide the opportu- select items)
nity to ‘make the strange familiar and the familiar strange’ (p. 20) in Challenge #2 (C2): Inadequate Forms and Guidelines (5 select
ways that are not possible with single-context studies. Alexander items)
(2001) suggests that cross-context studies “reveal alongside each Challenge #3 (C3): Unclear Policies and Procedures (5 select
jurisdiction’s unique mix of values, ideas and practices, powerful items)
continuities that transcend time and space” (p. 507). Further, cross- Challenge #4 (C4): Concerns about School Advising as a Sub-
context studies have the capacity to show that what might be Specialty (5 select items)
claimed as being distinctive in a one context might be less so when Challenge #5 (C5): Concerns about STs’ Pre-Practicum Prepa-
compared to other contexts. It is the juxtaposition of these ration (5 select items)
twodwhat is distinctive and what is commondthat is sought in Challenge #6 (C6): Uncertain Feedback and Communication
this study to better inform mentoring practices wherever they Practices (4 select items)
occur. Comparative analyses of this type allow individual jurisdic-
tions to compare their outcomes to the means of the ‘population at A graphic report, called an Individual MPI Profile, illustrates the
large’ for similar outcomes; in our study, that population is mentors degree to which mentors are motivated or challenged for each scale
from around the world. For example, if a particular outcome for an based on a score from 0 to 50. For those wishing to learn more
individual jurisdiction sits on or close to the mean for the popula- about the psychometric properties of the MPI, a full account is
tion at large, then that outcome is a relatively common phenome- available in Clarke et al. (2012).
non across all mentoring contexts and therefore may not warrant One important contribution of Individual MPI Profiles is that
particular attention or action. Alternatively, if the outcome for an they provide mentors with a language for talking about and
individual jurisdiction differs from the mean for the population at framing their practice (Cochran-Smith & Lytle, 1999; Scho €n, 1987).
 A. Clarke, J. Mena / Teaching and Teacher Education 90 (2020) 103026 3

For example, when asked about the impact of the MPI, one Young, O’Neill, & Mooney Simmie, 2015); in short, the intent of the
Australian mentor noted: mentoring instrument used in this study is not for the instrumen-
talization of mentoring. To this end, it is critical that conversations
I completed the inventory as part of the ‘Australian Institute for
with knowledgeable others accompany any use of the MPI,
Teaching and School Leadership’ (AITSL) professional develop-
particularly in local jurisdictions. If we accept that knowledge is
ment modules. … I found it very useful as a relatively quick and
personally constructed, socially mediated, and inherently situated
succinct means of assessing my own skills, priorities and also
(Brown, Collins, & Duguid, 1989; von Glasersfeld, 1989a; Wertsch,
direction for improvement. It did drive home the point as to
1991), then it is important that we honour these characteristics
how, when properly done, mentoring a student can actually be
within any context of mentoring if substantive learning is to take
an excellent form of professional development for the teacher
place.
and an assistance to improving teaching capacity. … It is some
years since I have supervised a pre-service teacher and am about
3. Method
to do so in the coming semester so this opportunity was very
timely indeed. (Personal communication, Thursday, July 10,
The MPI is currently available free, online, and in eight lan-
2014)
guages. To date, over 3000 mentors have taken the MPI. This
widespread use has allowed for the collective possibilities referred
As such, Individual MPI Profiles provide an external prompt, in to above. For example, when the number of MPI participants in a
addition to prompts by local jurisdictions, for thinking about one’s particular jurisdiction is large enough, it is possible to generate
mentoring practice. Aggregate MPI Profiles for a cohort of mentors in that jurisdiction.
Another important advantage of the MPI is that Aggregate MPI This study draws on six such cohorts, representing six different
Profiles can be generated for a cohort of practicum mentors from a countries, to conduct a comparative analysis of what motivates and
particular jurisdiction (e.g., a school district, a state or province, a what challenges mentors across contexts. This analysis allows us to
country, etc.). The significance of Aggregate MPI Profiles is that they better understand what might normally (or generally) be expected
allow the MPI to be used as a metric for cross-context comparative of mentors across settings and also to better discern what might be
analyses of cohorts of practicum mentors. Aggregate MPI Profiles particularly distinctive within settings. To do so, this study is
provide the opportunity to discern what might be considered guided by three research questions:
typical (or normal) versus distinctive (or significant) within and
across different contexts. RQ #1. What motivates and challenges practicum mentors
within countries?
RQ #2. How do the motivators and challenges differ across
2. The MPI in context countries?
RQ #3. Do countries group in any significant way in terms of
Mishler’s rhetorical question, “Meaning in context: is there any what motivates or challenges mentors?
other kind” (p. 1) goes to the heart of all research endeavours. With
equal poignancy, Scho € n (1987) challenges social science re- Each question has its own associated research method pathway
searchers to think about how they position themselves in relation (Table 1).
to that which they are studying: The institutional affiliation for the mentors in this study
(n ¼ 1828) is as follows: Auckland University, New Zealand
Shall he remain on the high ground where he can solve rela-
(n ¼ 178); Kasetsart University, Thailand (n ¼ 170); Northeast
tively unimportant problems according to prevailing standards
Normal University, China (n ¼ 257); University of British Columbia,
of rigor, or shall he descend to the swamp of important prob-
Canada (n ¼ 544); University of Salamanca, Spain (n ¼ 171); and
lems and non-rigorous inquiry? This dilemma has two sources,
Wollongong University, Australia (n ¼ 508). Using each of these as
first, the prevailing idea of rigorous professional knowledge,
country cohorts, a series of tests were conducted. For the sake of
based on technical rationality, and second, awareness of inde-
simplicity, the results arising from the analysis will be attributed to
terminate, swampy zones of practice that lie beyond its canons.
the country from which each practicum mentor sample is drawn.
(p 3)
However, it must be acknowledged that the results, although
designated as coming from a particular country, in effect represent
The seductive simplicity of readily codified behaviours has im- practicum mentors in the area of the country from which the
plications for the MPI. For example, the preparation of mentors sample is drawn.
would be greatly simplified if mentoring were to be viewed as First, normality and homogeneity tests were conducted to
instrumental problem solving made rigorous by the application of determine the distribution and variance for the overall and indi-
educational theory without a consideration of context. This posi- vidual country samples because the parameters of the population
tivist approach suggests that research outcomes are generalizable for all the countries was not known. The Kolmogorov-Smirnoff
across contexts and require little or no on-site interpretation normality test showed that the distribution for every country for
(Erickson, 1986). In this study, we are cognizant of being vigilant the motivator scales had values of p > .05 which indicates that the
against narrow technical-rational assumptions that cast the dataset was not different markedly from a normal distribution.
teacher-as-technician or mentor-as-mechanic. We have argued However for the challenge scales, countries such as Australia,
elsewhere that the practicum is a complex system and that learning Canada and New Zealand had values that were significant (p < .05)
is a dynamic and adaptive network of relationships and engage- and therefore not representative of a normal distribution. Since the
ments (Clarke & Collins, 2007). Further, Biesta and Miedema sample size for these countries is sufficiently large, this result is not
(2002), remind us that cultivation, not training, is the goal of any critical in terms of the overall purpose of the study. Second, a test of
educational enterprise. In this sense, this study is about the devel- homogeneity was conducted to check that all countries had the
opment of the profession not just professional development, and the same variance thus being suitable for comparative analysis. This is
MPI and should be viewed as an instrument for prompting local the first step in conducting the ANOVA to check whether the null
engagement rather than global imperatives (Gore et al., 2017; hypothesis does not differ between variances across groups
4 A. Clarke, J. Mena / Teaching and Teacher Education 90 (2020) 103026

Table 1
Three research questions and associated research method pathways.

Research Questions Hypothesis formulation Research Statistical Analyses

Approach

RQ#1
What motivates and challenges practicum Not applicable Descriptive  Descriptive statistics
mentors within countries? analysis (e.g., mean score and SD)

RQ#2
How do the motivators and challenges differ RQ#2.1 (H1 ) Teacher mentors are not equally motivated Compare  One-way ANOVA/Scheffe
across countries? What motivates mentors in to perform their practice across the six countries means post-hoc
their work with student (H0 ) Teacher mentors are equally motivated to  ROC curve
teachers? perform their mentoring across the six countries  Normalized score
RQ#2.2. (H2 ) Teacher mentors are not equally challenged Compare  One-way ANOVA/Scheffe
What challenges mentors in to perform their mentoring across the six means post-hoc
their work with student countries.  ROC curve
teachers? (H0 ) Teacher mentors are equally challenged to  Normalized score
perform their mentoring across the six countries

RQ#3
Do countries group in any significant way in Not applicable Compare  One-way ANOVA:
terms of what motivates or challenges means Homogeneous subsets,
mentors? (Bonferroni)

Note: n ¼ 1828 for all analyses.

(countries). If the variances are equal then it is advisable to proceed Once the ANOVA was conducted, a post-hoc test was used to
with the ANOVA test accordingly. Leven’s statistic for the motiva- confirm the differences between countries. For this, Scheffe’s for-
tors was 1.451 (p ¼ .203). For the challenges it was 7.654 (p ¼ .000). mula was used (where Xi  Xj is the mean scores of the samples
This was basically due to the data distribution of the New Zealand compared, ni and nj represent the sample sizes, and S2w represents
sample (mean ¼ 19.98; sd ¼ 5.001) compared to other countries variance within groups):
(mean scores above 21 and standard deviations between 6.295 and
7.529). Without this country, the rest of the country variances were
2
Fs ¼ Xi  Xj
equal. Because this assumption was violated, the Games-Howell " !#
procedure was used as it permits the use of ANOVA without
S2w 1  n1j
assuming equal variances (Games & Howell, 1976). Third, descrip- ni
tive statistical analyses were conducted to obtain information
about mean, median, maximum, minimum and standard deviation Receiver Operating Characteristic (ROC) curves and Youden
calculations for each country’s practicum mentor sample. This was index. ROC curves are typically used to determine the usefulness of
essential for conducting subsequent parametric analyses. Fourth, a a test. In this study they were calculated to demonstrate this point
comparative analysis using two processes was conducted to by graphically illustrating the differences among countries. This
determine the differences in the mentoring activity among the six analysis required the ROC calculation to determine the Area Under
countries: one-way ANOVA with Scheffe  post-hoc test and Receiver the Curve (AUC). ROC curves are basically a plot of the True Positive
Operating Characteristic (ROC) test. Each of the above analyses and Rates (TPR)dsensitivitydand the False Positive Rates (FPR)d1
their implications are described in detail below. minus specificitydfor the possible cut-off points. Each point in the
One-way ANOVA and Scheffe  post-hoc test. Statistical com- ROC curve depicts a sensitivity-specificity pair that informs a
parisons across cohorts using Aggregate MPI Profiles were done particular threshold. The Youden index was calculated to determine
with one-way ANOVAs for the 14 scales, where the p < .05 Games- the cut-off pair that shows the higher difference between a given
Howell test for post-hoc multiple comparisons was applied to country and the rest of the countries with regard to the MPI mo-
determine which specific cohort means differed from each other tivators and the challenges. The Youden index (J) is the maximum
and by how much. The Games-Howell procedure does not assume distance (e.g., vertical line) between the arc of the ROC curve and
equal variances across groups, hence corrects for unequal sample the chance line (e.g., first bisector or diagonal line). In other words,
sizes while remaining sensitive to small differences between the index represents the difference between TPR and FPR. It is
means. Throughout, a standard p < .05 alpha level was maintained calculated as: J ¼ maxc {Se (c) þ Sp (c)  1} where c stands for all
to determine the significance of differences between means or any possible values of the criterion variable (Searle, 1971; Efron & Tib-
pairs of means. Effect size was also calculated by using Etha square shirani, 1993). The resulting cut off-point for this calculation is the
(h2) since for ANOVA tests it is recommended to use this calculation optimal cut-off point (c*) where maximum differences are
instead of Cohen’s delta. Etha square (h2) was calculated as follows observed. All cut-off points (for both motivators and challenges)
(where SSeffect is the sum of squares for the effect and SStotal is the were calculated for all countries using Excel and J was identified
sum of squares for all effects): manually.
The ROC curve analysis served to illustrate the ANOVA statistical
SSeffect differences using the representation of a curve but also, and mainly,
h2 ¼ to mathematically represent the differences under non-parametric
SStotal
estimates. One of the foremost criticisms in social sciences is the
This value represents how the independent variable (i.e., use of parametric analyses with ordinal scales (i.e., it is assumed
countries) influences the dependent variable (i.e., motivators and that the answering position towards any item is the same from 1 to
challenges). It is calculated similarly to R squared and the values 2 as it is from 2 to 3, etc.). Since this is a controversial issue, and
used to interpret this coefficient are as follows: small (0.01); me- experts tend to agree that educational research should be con-
dium (0.059); large (0.138) (Cohen, 1988; Miles & Shevlin, 2001). ducted under non-parametric circumstances, ROC curves and
 A. Clarke, J. Mena / Teaching and Teacher Education 90 (2020) 103026 5

associated AUC were generated. Essentially, the AUC is a measure the final version was released in 2011 (Clarke et al., 2012). In 2018,
that shows the difference between two groups and it is calculated the MPI went through a full review (it was completely dis-
with the same values as for Mann-Whitney’s U test where pair- assembled and then reassembled) to ensure consistency and clarity
comparisons were executed between each country sample (TPR) across the entire survey.
and the remaining international sample, that is, the total sample A second important limitation is the difference in the context in
(n ¼ 1828) minus each country’s sample size (FPR). In short, the which the MPI was first developed and the subsequent contexts in
AUC ¼ U/(n1  n1 ) where AUC is the area under the ROC curve, U is which it was used in this study (Chabowski, Samiee, & Hult, 2017, p.
the statistical value from Mann-Whitney, and n1 and n1 represent 90). From the outset, an important element in the construction and
the number of participants from each sample group. An area that translation of the MPI was the involvement of international teacher
equals ‘1’ represents a perfect test or, in this study, the strongest educators from each of the six countries in this study: Australia,
difference between a country with regard the international sample, Canada, China, New Zealand, Spain, and Thailand. This interna-
while an area of 0.5 represents no difference at all (i.e., the diagonal tional effort ensured that the key concepts underpinning the survey
line in the graphical representation or the international sample’s n were clearly articulated and accurately translated so that the
value with which a given country is compared to). The standard mentors in each of the respective jurisdictions could meaningfully
error of the AUC was calculated following the method of DeLong, respond to the MPI. Nonetheless we acknowledge that mentoring is
DeLong, and Clark-Pearson (1988). not “culturally neutral and may come to symbolize a host of
Normalized scores. Current research indicates that percentiles different values driven by underlying assumptions and their
are a preferable method to normalize scores when sharing the re- meaning, use, and consequences” (Leidner & Kayworth, 2006, p.
sults of a survey with the general public. Percentiles are frequently 359). Thus, the development of international surveys such as the
used to report scoring in a test and indicate the percentage of the MPI is always an imperfect process and any claims emerging should
sample below which a certain number of values would be expected be tempered accordingly.
to fall. Subsequently, the dispersion for each scale was adjusted (8
motivators and 6 challenges) to conform to a near-normal distri- 4. Results
bution using the inverse normal probability distribution function
(in this study, Excel’s NORMINV function was used). The inverse The results detailed below are organized in terms of the three
normal probability distribution function operates with mean and research questions that guided this study:
standard deviation scores and yields values as percentages. As such,
the dispersion of the scores (0e50) for each of the scales varied RQ #1. What motivates and challenges practicum mentors
slightly (see Fig. 2). The semi-interquartile ranges (25% and 75%) are within countries?
highlighted in preference to standard deviations from the mean RQ #2. How do the motivators and challenges differ across
because most practicum mentors (and other audiences) are more countries?
intuitively familiar with percentages than standard deviations. RQ #3. Do countries group in any significant way in terms of
Nevertheless, all tests for statistically significant differences among what motivates or challenges mentors?
countries (see below) were performed using standard ANOVA
procedures.
Homogeneous subset test (Scheffe ). As a part of the ANOVA 4.1. RQ #1: what motivates and challenges practicum mentors
analysis, a homogeneous subset test was also conducted by using within countries?
multiple comparisons to see if countries clustered or grouped in
any significant way in relation to what motivates and challenges Individual MPI Profiles allow mentors to see what motivates or
practicum mentors. Assuming equal variances, the Scheffe  method challenges them in their work with student teachers, and the
was chosen because it uses F sampling distribution to conduct strength or otherwise of that potential attraction or discourage-
pairwise comparisons and for determining groups. ment. Aggregate MPI Profiles provide a similar overview for a
Reliability and factorial validity. With respect to the MPI and cohort of practicum mentors, that is each of the six countries,
its use in this study, reliability and validity was confirmed in a study involved in the comparative analysis. Over 1800 mentors
by Clarke et al. (2012). The reliability for the eight motivator factors completed the MPI as part of this study; the means and standard
and six-challenge factors resulted in scales with a-reliabilities deviations within country for the 14 scales are provided in Table 2.
ranging from 0.76 to 0.91. All scores were above what Nunnally The strength of particular motivators or challenges can be
(1978) established as recommended level: a > 0.70. Factorial val- readily observed. For example, the motivators scales for Canadian
idity was also calculated. The exploratory factor analysis executed practicum mentors range from a low of 24.18 to a high of 39.26.
in Clarke et al. (2012) arranged variables into factors using principal
components extraction. The KaisereMeyereOlkin (KMO) measure 4.2. RQ #2: how do the motivators and challenges differ across
of sampling adequacy was of 0.77 for the motivators and countries?
KMO ¼ 0.84 for 44 for the challenge items. All the estimations were
also determined by calculating a tetrachoric correlation matrix These scores can be seen in relation to the scores for the overall
(Osborne & Fitzpatrick, 2012). sample (Table 2, see ‘Overall’) where it can be seen that mentors felt
Limitations. As with all studies, there are limitations. Two are more motivated (mean score of 36.4 out of 50; that is, moderate to
particularly important in relation to this study. The first limitation significantly motivated) than challenged (mean score of 22.67 out
is the narrow band of items (n ¼ 62) that constituted the MPI of 50; that is, slightly to moderately challenged). The international
coupled with the very concise description needed to capture each sample also indicated that they were strongly motivated by
of those items using instruments of this type. These items were contributing to teacher education (M4: 40.40), renewing the pro-
reduced from an original set of 400 and the descriptors honed until fession (M1: 40.04) and providing mentoring in classroom contexts
there were only a handful of words representing each item. This is a (M8: 40.28). However, they were less motivated by the opportunity
limitation faced by anyone creating an instrument and made even offered by the student teacher’s presence to monitor their pupil
more challenging when trying to develop an online user-friendly learning (M4: 31.11) or promote pupil engagement (M3: 33.8). In
format. Nonetheless, the MPI underwent three pilot tests before short, the international sample of practicum mentors were more
6 A. Clarke, J. Mena / Teaching and Teacher Education 90 (2020) 103026

motivated to be involved in the practicum by the teaching benefits

Challenges
Combined
rendered to the student teacher and the profession at large than by
Mean and

the learning benefits rendered to their pupils by the student

SD for

(7.37)

(4.94)

(7.00)

(6.18)

(6.57)

(6.23)

(7.50)
21.41

19.87

21.12

25.00

32.09

21.28

22.67
teachers’ presence.
As for the main challenges reported, overall it was the elements
Preparation Communication
Advising Practicum Feedback and

of mentoring that were largely ‘out of the practicum mentors’

C6. Uncertain

hands’ that the international sample found most vexing: for

example, the lack of substantive pre-practicum preparation (C5:
(8.61)

(6.68)

(8.48)

(8.88)

(9.26)

(7.33)

(8.93)
21.04

21.70

21.86

24.11

31.94

21.97

22.87
24.52); unclear policies and procedures (C3: 23.18), and lack of
forms and guidelines (C2: 23.09). Here, mentors are asking for
greater support and engagement in the practicum by teacher ed-
C4. School C5. Pre-

(9.41)

(7.13)

(9.23)

(8.77)

(8.58)

(7.83)

(9.35)
23.15

21.13

24.66

25.73

32.83

21.86

24.52
ucation institutions. While the wording of these challenges might
appear quite stark, it reflects the core of a relationship that mentors
are seeking between the universities and the schools. There are no
Specialty
as Sub-

easy answers here and each jurisdiction must question how and in
(8.91)

(6.85)

(8.08)

(6.86)

(9.09)

(7.13)

(8.66)
22.25

19.65

19.32

22.66

30.76

21.95

22.01
what ways these issues are being addressed. For example, would
moving to a digital format for much of this type of communication
Procedures

enable or disable the school/university relationship? Would a dig-

Unclear
Policies

(9.54)

(7.26)

(9.75)

(8.73)

(8.51)

(8.56)

(10.1)
20.58

19.40

21.88

30.22

31.99

19.82

23.18

ital format change the sense of connection or the feeling of

and
C3.

belonging for mentors?

Guidelines
of Forms
C2. Lack

(10.11)

(10.11)

4.2.1. Single-factor analysis of variance (ANOVA) and Scheff

(7.95)

(9.85)

(8.61)

(8.44)

(8.42)

e post-
21.33

19.44

21.76

28.02

32.35

19.68

23.09
and

hoc test
Further analysis (ANOVA, ROC curves, and normalized scores)
Challenges

Challenges

Mentoring
Guidance

reveals different ways in which teachers from various countries are

(10.12)
(9.30)

(6.79)

(6.25)

(7.53)

(8.80)
20.24

18.54

18.04

20.08

32.53

22.34

20.79
(5.4)

either motivated or challenged in their work as practicum mentors.

and
C1.

in

The ANOVA highlights the statistical differences between motiva-

Motivators
Combined
Mean and

tors and challenges among countries (see Table 3).

The motivator (ALL) and challenge (ALL) scales showed signifi-
SD for

(6.03)

(5.50)

(5.64)

(5.78)

(5.84)

(5.30)

(6.00)
34.58

34.77

36.23

38.99

38.03

38.87

36.40

cant differences at the level of p < .05: motivators (ALL), F ¼ 33.35,

p ¼ .000; challenges (ALL), F ¼ 90.14, p ¼ .000. The effect size in the
Mentoring

Classroom

case of the motivators was medium (h2 ¼ 0.083) whereas for the
Contexts

(7.05)

(6.10)

(6.12)

(6.46)

(6.87)

(5.16)

(6.49)

challenges it was large (h2 ¼ 0.198). This result can be interpreted

39.26

40.04

41.02

40.10

40.31

41.85

40.28
M8.

in

as 8.3% and 19.8% of the changes (explained variance) in the mo-

Professional
Community

tivators and challenges variables, respectively, are due to the in-

Developing

dependent variable (i.e., the six participant countries). The larger

(7.69)

(8.25)

(7.23)

(7.69)

(7.25)

(7.14)

(7.61)
37.33

38.15

40.10

38.90

38.30

38.87

38.64
Means (and standard deviations) for aggregate MPI profiles (motivators and challenge) by country.

major effect sizes were found in M4 (‘Time Out’ to Monitor Pupil

M7.

Learning; h2 ¼ 0.210) and M6 (Reminders About Career Develop-

a
Development
About Career

ment; h2 ¼ 0.250). With respect to the challenges, the higher effects

Reminders

were found in C1 (Challenges in Guidance and Mentoring;

h2 ¼ 0.205) and C3 (Unclear Policies and Procedures’; h2 ¼ 0.188).
(9.43)

(9.33)

(9.58)

(7.86)

(9.44)

(8.19)

(10.5)
24.18

24.56

25.28

38.10

32.92

34.32

28.24
M6.

This means that around 20% of the explained variance in both

Contributing

challenges scales is accounted for by the countries’ differences. In

to Teacher
Education

other words, these two dependent variables are explained in high

(6.85)

(5.82)

(6.06)

(6.68)

(5.91)

(5.94)

(6.48)

proportion by just one independent variable (e.g., country).

38.75

41.07

42.10

39.42

40.96

40.81

40.40
M5.

Post hoc analyses (Scheffe) for significance showed that Spanish

challenge scores were significantly higher in relation to the inter-
M4. ‘Time

Learning

national sample of practicum mentors than any other country (see

Monitor
Out’ to

(9.57)

(9.43)

(9.77)

(7.38)

(7.47)

(7.43)

(10.1)
30.00

24.75

27.23

38.27

36.32

36.87

31.11
Pupil

Table 4).
The Spanish mean difference for challenges is above 10 (row 6,
Engagement
M3. Student

columns 2e4) when compared to Canada, Australia, and New

Teachers
Promote

Zealand; and 6.9 (row 6, column 5) when compared to China. This

(9.50)

(8.78)

(9.63)

(7.53)

(8.73)

(6.59)

(9.33)
32.36

30.03

32.40

37.74

34.82

39.54

33.80
Pupil

shows that, on average, the Spanish practicum mentors feel 20%

more challenged in their role as mentors than their international
counterparts (and around 14% more challenged than the Chinese
Improving

Teaching
My Own

Practice

practicum mentors). This is a good example of potential insight

(7.98)

(7.82)

(7.51)

(7.48)

(9.24)

(8.07)

(8.03)
37.19

40.10

40.78

39.11

38.04

35.56

38.67

gained from cross-context studies that is not always possible using

M2.

within-context studies.
Motivators

Profession
Renewing

In the case of motivators, China and Thailand (columns 4 and 7,

(8.36)

(8.21)

(7.55)

(7.19)

(6.23)

(5.77)

(7.79)
37.56

39.44

40.91

40.25

42.53

43.15

40.04

respectively) obtained the highest mean difference scores of the

M1.

the

practicum mentor sample compared to Canada (M ¼ 4.408,

SE ¼ 0.435 and M ¼ 4.294, SE ¼ 0.504, respectively) and New
Zealand
(n ¼ 544)

(n ¼ 508)

(n ¼ 257)

(n ¼ 171)

(n ¼ 170)

n ¼ 1828
Australia

Thailand
( ¼ 178)
Country

Zealand (M ¼ 4.219, SE ¼ 0.569 and M ¼ 4.105, SE ¼ 0.616,

Canada

Overall
China

Spain
Table 2

New

respectively). Broadly it could be said that Chinese and Thai prac-

ticum mentors felt almost 10% more motivated than Canadian and
 A. Clarke, J. Mena / Teaching and Teacher Education 90 (2020) 103026 7

Table 3
Source of Variation for Motivator and Challenge Scales (between groups, within groups, and overall totals).

Source of Variation SS (1) df (2) MS (3) F(4) p value (5) Effect size(h2)(6)

Motivators
MOTIVATORS (ALL)
- Between groups 5505.50 5 1101.1001 33.3500 .000 .0838
- Within groups 60155.93 1822 33.016 (medium)
- Total 65661.43 1827
MOTIVATOR #1
- Between groups 6508.217 5 1301.643 22.753 .000 0.059
- Within groups 104233.041 1822 57.208 (medium)
- Total 110741.258 1827
MOTIVATOR #2
- Between groups 5576.970 5 1115.394 17.859 .000 0.047
- Within groups 113791.557 1822 62.454 (small)
- Total 119368.528 1827
MOTIVATOR #3
- Between groups 14442.269 5 2888.454 36.437 .000 0.091
- Within groups 144434.550 1822 79.273 (medium)
- Total 158876.819 1827
MOTIVATOR #4
- Between groups 38952.498 5 7790.500 96.811 .000 0.210
- Within groups 146618.180 1822 80.471 (large)
- Total 185570.678 1827
MOTIVATOR #5
- Between groups 3354.243 5 670.849 16.650 .000 0.044
- Within groups 73408.486 1822 40.290 (small)
- Total 76762.729 1827
MOTIVATOR #6
- Between groups 50891.771 5 10178.354 121.575 .000 0.250
- Within groups 152538.791 1822 83.721 (large)
- Total 203430.563 1827
MOTIVATOR #7
- Between groups 2110.398 5 422.080 7.410 .000 0.020
- Within groups 103784.227 1822 56.962 (small)
- Total 105894.625 1827
MOTIVATOR #8
- Between groups 1287.552 5 257.510 6.195 .000 0.017
- Within groups 75737.513 1822 41.568 (small)
- Total 77025.065 1827
Challenges
CHALLENGES (ALL)
- Between groups 20131.24 5 4072.449 90.140 .000 0.1984
- Within groups 83441.03 1822 45.179 (large)
- Total 103572.28 1827
CHALLENGE#1
- Between groups 28997.291 5 5799.458 93.947 .000 0.205
- Within groups 112473.703 1822 61.731 (large)
- Total 110741.258 1827
CHALLENGE #2
- Between groups 27843.401 5 5568.680 63.840 .000 0.149
- Within groups 158931.286 1822 87.229 (large)
- Total 186774.687 1827
CHALLENGE #3
- Between groups 34985.631 5 6997.126 84.366 .000 0.188
- Within groups 151112.231 1822 82.938 (large)
- Total 186097.862 1827
CHALLENGE #4
- Between groups 17901.815 5 3580.363 54.724 .000 0.131
- Within groups 119206.077 1822 65.426 (large)
- Total 137107.893 1827
CHALLENGE #5
- Between groups 16463.263 5 3292.653 41.915 .000 0.103
- Within groups 143127.028 1822 78.555 (medium)
- Total 159590.291 1827
CHALLENGE #6
- Between groups 17194.125 5 3438.825 48.801 .000 0.118
- Within groups 128390.421 1822 70.467 (medium)
- Total 145584.546 1827

Notes: (1) Sum of squares; (2) Degrees of freedom; (3) Mean square, estimated by the equation ¼ SS/df; (4) Calculated F value according to equation F ¼ MS (between)/MS
(within; (5) p-value: if < 0.05 there is significant difference between the two data groups; and (6) Effect size. h2 (Eta squared).
8 A. Clarke, J. Mena / Teaching and Teacher Education 90 (2020) 103026

Table 4
Scheffe Post-hoc Test.

New Zealander counterparts. However, when compared to Spain,

both countries were around 2% more motivated (M ¼ 0.96; for
China vs. Spain, row 5, column 6; and M ¼ 0.84 for Thailand vs.
Spain, row 6, column 7).

4.2.2. ROC curves and the Area Under the Curve (AUC)
ROC curves were used in this study to directly compare a clas-
sification system, in this case the scores given by country on the
MPI Likert-type instrument, according to motivators and challenges
that mentors perceived in their practice. This analysis offers
particular thresholds that might be important in making decisions
about the phenomenon under study. In this study, it is possible to
decide at which point particular differences among countries are
indicative of strong response patterns. As explained earlier, AUC
scores close to 1 represent a stronger difference between a country
and the international sample, while a score of 0.5 represents no
difference. Main scores are shown in Table 5.
High values for Spain and China are worth noting. The ROC for
Spain (Fig. 1) can be summarized as follows: a true Positive value of
AUC ¼ 0.854 for challenges in the Spanish sample (Table 5: row 7,
column 8) with a confidence interval of 0.827e0.881 at 95% (the
value 0.854 is the mean score) and a SE ¼ 0.14. This highlights that Fig. 1. Spain’s ‘Receiver Operating Characteristic’ (ROC) to determine the Area Under
the Curve (AUC) for deciding True Positive Rates (TPR) and the False Positive Rates
85% of the area is under the curve (also see Fig. 1, the uppermost
(FPR).
jagged line). Therefore the differences between Spain (n ¼ 171) and

Table 5
ROC curve analysis: Differences between each country and the overall international sample.

Positive (country’s n value) Negative (international sample’s n value) Motivators Challenges

Area (AUC) J index p value SE Area (AUC) J index p value SE

Canada(1) 544 1284 0.379 0.18 **0.000 0.14 0.425 0.129 **0.000 0.15
New Zealand (2) 178 1650 0.411 0.149 **0.000 0.22 0.390 0.235 **0.000 0.28
Australia (3) 508 1320 0.483 0.058 0.265 0.15 0.420 0.134 **0.000 0.15
China (4) 257 1571 0.646 0.233 **0.000 0.18 0.630 0.247 **0.000 0.17
Spain (5) 171 1657 0.588 0.164 **0.000 0.23 0.854 0.551 **0.000 0.14
Thailand (6) 170 1658 0.636 0.206 **0.000 0.21 0.446 0.134 *0.021 0.21

Note: AUC ¼ Area Under the Curve; J index ¼ Youden Index (J): J ¼ Sensitivityc þ specificityc 1. p-value: if < 0.05 (*) or if < 0.01 (**) there is significant difference between the
two data groups: e.g. each country (labeled ‘positive’ above) with respect to the remaining total sample (labeled ‘negative’ above); SE¼ Standard error.
 A. Clarke, J. Mena / Teaching and Teacher Education 90 (2020) 103026 9

4), respectively. Therefore, the difference in motivation within

these countries can be regarded as about 25%e28% less when
compared to more highly motivated practicum mentors (e.g.,
China). In the case of challenges, the AUC values for both Canada
and New Zealand (Table 5, rows 3 and 4; column 8) are
AUCCan ¼ 0.425 and AUCNZ ¼ 0.390, respectively, meaning that
according to the model they feel challenged in a percentage of
around 40% when compared to the international sample. But when
compared to Spain (highly challenged group, AUCSp ¼ 0.854) they
perceive almost half of the difficulties, drawbacks, and limitations
in their work as practicum mentors than their Europeans coun-
terparts (e.g., AUCSp - AUCNZ ¼ 0.464).

4.2.3. Normalized scores for motivators and challenges scales by

country
Specific charts (Figs. 3 and 4) were designed to graphically
illustrate the motivator and challenges scale scores, respectively, for
each country relative to:

(1) the overall international sample of mentors (i.e., n ¼ 1828);

and
(2) the individual sample of mentors from each country.
Fig. 2. China’s ‘Receiver Operating Characteristic’ (ROC) to determine the Area Under
the Curve (AUC) for deciding True Positive Rates (TPR) and the False Positive Rates
(FPR).
The far left and right margins display the percentiles indicating
the percentage of the population that lies above and below a
particular line across the chart. In particular, there are three hori-
all the other countries (n ¼ 1657) can be accounted for in that zontal shaded lines that run from left to right across each chart. The
percentage. In other words, the area that lies under the curve line horizontal shaded line midway through the chart is the point at
represents the percentage of randomly chosen pairs from Spain which 50% of the population falls above the line and 50% falls below
that differ from the pairs chosen from all the other countries. It the line. The interquartile rangesdthe 25th percentile and 75th
shows the probability that the test distinguishes different response percentiledare highlighted by the upper and lower horizontal
patterns from Spain when compared to the other five countries. In shaded lines. For example, 10% of the population lies below the 10th
short, the Spanish practicum mentors are the ones more challenged percentile line. At the other end of the percentile range, 10% of the
in their work as mentors. In general terms, as sensitivity (true population lies above the 90th percentile line. However, it should
positives) increases (c ¼ 0.854), it can be argued that more mentors be noted that all tests of statistically significant differences among
in Spain who were challenged were identified (sensitivity) but, on countries (see below) were performed using standard ANOVA
the other hand, the accuracy of identifying those who were less procedures.
challenged was limited (specificity). The Youden Index (J) for Spain The marker (or point) that is used in this study to locate a
(sensitivityc þ specificityc 1) in challenges was 0.55 (cut-off point particular population (either the overall sample or an individual
with values of 0.789, sensitivity, and 0.238, specificity, respec- country) on a particular motivator or challenge scale is the mean
tively). Any J value above 0.5 indicates a threshold that marks score for that population on that scale. The mean score is used
strong differences between the groups and therefore can be because it best represents the collective assessment of all re-
considered relevant. spondents for that population. For the overall population sample
For motivators, China (Table 5, row 6, column 4) has the highest this point falls on the 50th percentile for all eight motivator scales
True Positive Value out of the six countries when compared with and all six challenge scales. Subsequently, each scale’s overall
the international sample (AUC ¼ 0.646 with an interval of confi- population dispersion is adjusted to conform to a near-normal
dence of 0.611e0.681 at 95%). Chinese practicum mentors are on distribution using the inverse normal probability distribution
average more motivated than any of their international counter- function. As such, the dispersion of the scores (0-50) for each of the
parts. This is a case of needing to think carefully about the context scales varies slightly in Figs. 3 and 4 depending upon the scale in
and the part it plays in explaining this outcome. Might it be the question. Please note that the line joining the scores for a particular
more communal orientation underpinning Chinese society at play country has no significance other than to assist readers to visually
here? While beyond the scope of this paper, these are issues that track the scores for an individual country in comparison to the
can, in turn, be taken up at the local level. In this case, the ROC curve overall population and other countries.
for China in Fig. 2 illustrates the differences by the noticeably Based on the mean of the overall sample on both charts (i.e., the
sharper slope (the jagged line that begins higher than the other two 50th percentile shaded horizontal line) and the spread of scores
lines at the outset) compared to the chance line (the straight di- recorded by the mentors from all six countries:
agonal line that represents the international sample to which it is
compared). The Youden index for this country shows the maximum - scores falling on the 50th percentile are regarded as standard
difference between sensitivity and specificity (Sensitivity: 0.626; motivators or challenges for mentors;
Specificity: 0.393; J ¼ 0.233). Graphically it would be represented as - scores falling 15 percentile points above or below the 50th
a vertical line from the higher part of the arc to the chance line. percentile (i.e., the 65th percentile or 35th percentile) are
Canada and New Zealand represent the lowest scores for mo- regarded as either important or, conversely, modest motivators
tivators among the countries of the sample with TPV ¼ 0.379 or challenges, respectively; and
(Table 5: row 3, column 4) and TPV ¼ 0.411 (Table 5: row 4, column - scores falling 30 percentile points above or below the 50th
percentile (i.e., the 80th percentile or 20th percentile) are
10 A. Clarke, J. Mena / Teaching and Teacher Education 90 (2020) 103026

Fig. 3. Motivator scale scores: Comparative analysis of six countries.

regarded as either exceptional or, conversely, slight motivators or 4.2.3.1. Individual country motivator and challenge scores in com-
challenges, respectively. parison to the overall sample. In the following analysis, country
scores that are regarded as noteworthy are reported.
In terms of the comparative work undertaken here, the analysis The results for motivators compared to the overall sample are as
suggests that results that fall on or near the 50th percentile (i.e., follows:
standard) are what can be reasonably expected across all contexts.
In contrast, our analysis suggests that results that fall on or near the
65th percentile or 35th percentile (i.e., important or modest), or a) Canada. Of all eight motivators, only two are noteworthy and
results that fall on or near the 80th percentile or 20th percentile both are modest motivators: M1 (Renewing the profession)
(i.e., exceptional or slight), are noteworthy because they differ and M6 (Reminders about career development). Both lie at or
markedly from what might be expected across contexts. around the 35th percentile and do not appear to be partic-
ularly strong motivators for Canadian mentors.
A. Clarke, J. Mena / Teaching and Teacher Education 90 (2020) 103026 11

Fig. 4. Challenge scale scores: Comparative analysis of six countries.

12 A. Clarke, J. Mena / Teaching and Teacher Education 90 (2020) 103026

M3 Student-Teachers Promote Pupil Engagement. The most

distinctive feature about this scale is that Chinese and Spanish
b) New Zealand. Of all eight motivators, only two are note- practicum mentors value the way student teachers promote
worthy and both because New Zealanders only indicate a pupil engagement whereas, almost to the same degree but in
slight interest in them as motivators: M4 (‘Time out’ to the opposite direction, the Australian and Canadians indicate
monitor pupil learning) and M6 (Reminders about career that it is considerably less so.
development).
M4 ‘Time out’ to Monitor Student Learning. With the excep-
c) Australia. Only one motivator of all eight is noteworthy: M4 tion of Canada (for whom the result almost maps on to the 50th
(‘Time out’ to monitor pupil learning). This motivator is only percentile), this scale splits the other five countries in a signif-
of slight interest to Australian mentors. icant way: Thailand, Spain, and China find it quite compelling;
and for Australia and New Zealand it is particularly low in terms
d) China. Three of the eight motivators are noteworthy and all of being a motivator,
three positively so. Two are important motivators: M3 (Stu-
dent teachers promote pupil engagement) and M4 (‘Time M5 Contributing to Teacher Education. This is one motivator
out’ to monitor pupil learning). One motivator is exceptional: where all six countries are relatively ‘on the same page’ and
M6 (Reminders about career development). cluster around the overall mean.

e) Spain. Two of the eight motivators are noteworthy. Both are M6 Reminders About Career Development. This is another
important motivators for Spanish mentors: M4 (‘Time out’ to motivator that splits the countries in opposite directions. This
monitor pupil learning) and M6 (Reminders about career motivator is particularly important for China, Spain, and
development). Thailand but of much less significance for practicum mentors
from Australia, Canada, and New Zealand.
f) Thailand. Four of the eight motivators are noteworthy for the
Thai mentors. Three are regarded as important: M1 M7 Developing a Professional Community. Similar to M5
(Renewing the profession), M3 (Student teachers promote above, all six countries cluster about the overall mean.
pupil engagement), and M6 (Reminders about career devel-
opment). One motivator is reported as only being of slight
M8 Mentoring in Classroom Context. Similar to M5 and M7
interest: M2 (Improving my own teaching practice).
above, all six countries cluster about the overall mean.
The results for challenges compared to the overall sample are as
The results for challenges in comparison to other countries are
follows:
as follows:
It is interesting to note that responses from two of the countries
in the comparison, New Zealand and Spain, are particularly
a) China. Two of the six challenges scales were particularly note- distinctive. Both tend to ‘flat-line’ at opposite ends of the challenge
worthy, both as important for Chinese mentors: C2 (Lack of scales. For New Zealanders, none of the six challenges are of any
forms and guidelines) and C3 (Unclear policies and procedures). great concern: all are regarded as only ‘slight’ challenges (flat-lining
around 40th percentile) within their context. On the other hand,
b) Spain. All six challenge scales were noteworthy for Spanish Spanish mentors indicate ‘exceptional’ or ‘significant’ challenges
mentors. Five scales reached the exceptional level: C1, C2, C4, C5, with all six challenge scales (flat-lining around the 85th percentile).
and C6. One scale was reported at the important level: C3 Unless otherwise reported, all other countries regard each of the
(Unclear policies and procedures). challenges as a standard concern (inside 15 points of the 50th
percentile).

C1 Challenges in Guidance and Mentoring. Compared to all

4.2.3.2. Individual country motivator and challenge scores in com- other countries, Spain is the only country to rate this challenge
parison to other countries. In this analysis, scale scores are high- as ‘critical’ (90th percentile).
lighted for individual countries that are distinctively different from C2 Lack of Forms and Guidelines. This particular challenge
other countries. For the motivator scales, there was something stands out as being exceptional in the Spanish context and as
noteworthy about all six countries. For the challenge scales, there important in the Chinese context.
was something noteworthy about two countries. C3 Unclear Policies and Procedures. Practicum mentors in both
The results for motivators in comparison to other countries are Spain and China find this to be an exceptional challenge in their
as follows: contexts.
C4 School Advising as a Sub-Specialty. This is an exceptional
challenge for Spanish practicum mentors.
M1 Renewing the Profession. This is a moderate motivator for C5 Pre-practicum Preparation. This is an exceptional challenge
mentors from all countries except Canada which falls approxi- for Spanish practicum mentors.
mately 15 percentile points below any other country. C6 Uncertain Feedback and Communication. This is as an
exceptional challenge for Spanish practicum mentors.
M2 Improving My Own Teaching Practice. In this instance, it is
the Thai mentors who do not find this to be a particular moti- 4.3. RQ #3: do countries group in any significant way in terms of
vator for working with student teachers whereas the other what motivates or challenges mentors?
countries find it more compelling but not excessively so.
Do the countries group in ways that might be anticipated or are
 A. Clarke, J. Mena / Teaching and Teacher Education 90 (2020) 103026 13

there surprising groupings that might challenge taken-for-granted  test has opened up a crack and illuminated
gets in.” The Scheffe
assumptions about those contexts? The Scheffe  homogeneous some curiosities with respect to cohort groupings that are
subset test permits the grouping of countries based on the differ- deserving of future inquiry.
ences found both in the ANOVA and normalized scores, thereby
highlighting similar response patterns (homogenous subsets) 5. Conclusion
across countries. In short, the mean scores that are listed under the
same subset are not significantly different from each other and International comparative studies allow us to learn about our-
therefore can form a statistically defined subset. This provides selves by learning about others. Central to this type of learning is a
another way of ‘making the strange familiar and the familiar common metric. In this study, the MPI played that role. As noted
strange’ and suggests more macro factors are at play (at the country earlier, it is important to be mindful of the MPI’s limitations (e.g.,
level). issues and meanings vary within contexts and can never be fully
In the following discussion, the groupings will be referred to as accounted for by a common metric). With this caveat in mind, each
high-motivated, medium-motivated, or low-motivated, or high- of the claims detailed above (in response to the three research
challenged, medium-challenged, or low-challenged. In terms of questions) contributes to our knowledge about mentoring in
motivators the Scheffe  homogeneous subset test indicates that practicum contexts at both local and global levels. For example, the
there are two significantly different groupings that were clearly results reveal that the scores for three scales for all six countries
delineated based on the harmonic mean of the group sizes map very closely onto the overall sample mean:
(calculated as 239.559 for both motivators and challenges) as being
high-motivated or high challenged, medium-motivated or - M5 (Contributing to teacher education);
medium-challenged; and low-challenged (see Table 6). - M7 (Developing a professional community); and
New Zealand, Canada, and Australia, as a cluster, emerged as - M8 (Mentoring in classroom contexts).
medium-motivated groups. Perhaps given their cultural similar-
ities, this outcome is not particularly surprising. However, some- This outcome suggests these scales potentially constitute three
what surprisingly, China, Thailand, and Spain formed another international norms for mentoring in practicum settings. This pos-
cluster: high-motivated. What might the Asian and European sibility was not known before this study. One implication of this
contexts have in common that brings about this particular finding is that all jurisdictions can use this as baseline data upon
grouping? The two groups showed an alpha of 0.05. High- which to structure mentor recruitment, preparation, sustainability,
motivated practicum mentors are classroom teachers that all con- and succession. To neglect one or more of these motivators would
texts strive for and other than speculating on the possibilities at be a mistake given their universal prominence. Likewise, other
play, a more definitive answer is beyond the scope of this paper. claims have implications that bear greater scrutiny in particular
Nonetheless, the results challenge us to think more expansively jurisdictions. In particular, this comparative analysis highlights is-
about motivation at the macro-level. sues in some countries that are not common to other countries and
As for challenges, the Scheffe test confirms three significantly therefore provide potentially useful insight. For example, this can
distinct groups (with alpha of 0.05): Canada, New Zealand, be seen in the Chinese jurisdiction where there is a very significant
Australia and Thailand are clustered as low-challenged: China, as a challenge with respect to ‘unclear policies and guidelines,’ more so
‘singleton’ is identified as medium-challenged: and, Spain as than any other challenge identified by Chinese mentors. Therefore,
another ‘singleton’ is identified as high-challenged. Again, these given its prominence, it would seem prudent to review how and in
results provoke us to think more broadly about how challenge is what way policies and guidelines are presented and communicated
constructed (and experienced) in these groups. In the case of either in this context. On the other hand, a very significant motivator in
the motivator or challenge groupings further investigation is this jurisdiction (more so than in any other jurisdiction in this
necessary to determine the macro factors at play. For example, study) is “Reminders about career development.’ Thus, in China it
what is it about the high-motivated group that might serve as a would be important to be mindful of this factor and attend to it
model for others? What is it about the low-challenged group that accordingly in seeking out teachers to serve as practicum mentors.
might similarly serve as a model to others? This prompts a further An avenue of pursuit that might be of interest to Chinese teacher
question in relation to the above results: Is Thailand the envy of all educators would be to articulate and expound upon this aspect of
the other countries? On the one hand, it groups with countries that mentoring (which is possibly taken for granted) in that setting so
are high-motivated and, on the other hand, it groups with countries that other countries might learn why it is so significant for Chinese
that are low-challenged. As the Canadian poet Leonard Cohen mentors. Whether it is the motivators or the challenges, there is
famously wrote: “There is a crack in everything, that’s how the light something in the results for each jurisdiction that is unlikely to
have been recognized as distinctive for that context prior to this
study. Each of these elements needs to be taken up in the respective
context and explored accordingly.
Table 6
).
ANOVA test: Homogeneous subsets (Scheffe
In sum, this international comparative study of practicum
mentors is important for three reasons. First, it is important for the
Level of Motivation Level of Challenge
identification of issues that might otherwise remain hidden or
Country Mean Country Mean unrecognized from a single-context perspective. Second, it is
Group 1 (High) China 38.985 Spain 32.09 important because what might be claimed as being distinctive in a
Thailand 38.871 particular context might be shown to be less so when that claim is
Spain 38.026 located against the backdrop the other mentoring contexts. Third,
Group 2 (Medium) New Zealand 34.766 China 25.00
in the case of the mentoring literature, the comparative analyses
Canada 34.577
Australia 36.227 presented here represent an important shift from what has been
Group 3 (Low) Canada 21.417 largely idiosyncratic within-context analyses to comparative cross-
Thailand 21.283 context analyses. This initial exploration of the MPI’s potential to
Australia 21.122 deepen our understanding of why and how practicum mentors
New Zealand 19.876
work with student teachers provides a different way of thinking
14 A. Clarke, J. Mena / Teaching and Teacher Education 90 (2020) 103026

about potential investigations at both the macro and micro levels of von Glasersfeld, E. (1989a). Constructivism. In T. Husen, & T. N. Postlethwaite (Eds.)
(1st ed.,, supplement Vol. 1. The international encyclopedia of education (pp.
teacher education. Examples of some countries already embarking
162e163). Oxford: Pergamon.
on these investigations include China (Lv, Wang, Ma, Clarke, & Gore, J., Lloyd, A., Smith, M., Bowe, J., Ellis, H., & Lubans, D. (2017). Effects of pro-
Collins, 2016), Australia (Nielsen et al., 2017), and Thailand fessional development on the quality of teaching: Results from a randomised
(Faikhamta & Clarke, 2018). With the MPI now in multiple lan- controlled trial of Quality Teaching Rounds. Teaching and Teacher Education, 68,
99e113. https://doi.org/10.1016/j.tate.2017.08.007.
guages, the possibility for others to tap into this potential is only a Hansford, B., Ehrich, L., & Tennent, L. (2004). Outcomes and perennial issues in
keyboard away: www.mentoringprofile.com. preservice teacher education mentoring programs. International Journal of
Practical Experiences in Professional Education, 8, 6e17.
Hauser, R. M. (2016). Comparable metrics: Some examples. Chinese Journal of So-
Acknowledgements ciology, 2(1), 3e33. https://doi.org/10.1177/2057150X15624896.
Hudson, P. (2007). Examining mentors’ practices for enhancing preservice teachers’
The authors of this study woud like to acknowledge the Social pedagogical development in mathematics and science. Mentoring & Tutoring:
Partnership in Learning, 15, 201e217.
Science and Humanties Research Council of Canada (SSHRC) for Izadinia, M. (2015). A closer look at the role of mentor teachers in shaping pre-
their support for this study. service teachers’ professional identity. Teaching and Teacher Education, 52, 1e10.
https://doi.org/10.1016/j.tate.2015.08.003.
Kent, S. I. (2001). Supervision of student teachers: Practices of cooperating teachers
Appendix A. Supplementary data prepared in a clinical supervision course. Journal of Curriculum and Supervision,
16, 228e244.
Supplementary data to this article can be found online at Kitchel, T., & White, C. (2007). Barriers and benefits to the student teacher-
cooperating teacher relationship. In G. E. Briers, & T. G. Roberts (Eds.), Vol. 34.
https://doi.org/10.1016/j.tate.2020.103026.
Proceedings of the national AAAE research conference (pp. 710e712) (Minneap-
olis, MN).
References Koskela, R., & Ganser, T. (1998). The cooperating teacher role and career develop-
ment. Education, 119, 106e125.
Alexander, R. (2001). Border crossings: Towards a comparative pedagogy. Compar- Lave, J., & Wenger, E. (1991). Situated learning: Legitimate peripheral participation.
ative Education, 37, 507e523. Cambridge: Cambridge University Press. https://doi.org/10.1017/
Ambrose, S. A., Bridges, M. W., DiPietro, M., Lovett, M. C., & Norman, M. K. (2010). CBO9780511815355.
How learning works: Seven research-based principles for smart teaching. John Leidner, D. E., & Kayworth, T. (2006). Review: A review of culture in information
Wiley & Sons. systems research: Toward a theory of information technology culture conflict.
Biesta, G. J., & Miedema, S. (2002). Instruction or pedagogy? The need for a trans- MIS Quarterly, 30, 357e399. https://doi.org/10.2307/25148735.
formative conception of education. Teaching and Teacher Education, 18(2), Lindqvist, P., Nord€ anger, U. K., & Carlsson, R. (2014). Teacher attrition the first five
173e181. https://doi.org/10.1016/S0742-051X(01)00062-2. yearseA multifaceted image. Teaching and Teacher Education, 40, 94e103.
Brown, J., Collins, A., & Duguid, P. (1989). Situated cognition and the culture of https://doi.org/10.1016/j.tate.2014.02.005.
teaching. Educational Researcher, 18(1), 32e42. https://doi.org/10.3102/ Lv, L., Wang, F., Ma, Y., Clarke, A., & Collins, J. (2016). Exploring Chinese teachers’
0013189X018001032. commitment to being a cooperating teacher in a university-government-school
Buchanan, J., Prescott, A., Schuck, S., Aubusson, P., Burke, P., & Louviere, J. (2013). initiative for rural practicum placements. Asia Pacific Journal of Education, 36(1),
Teacher retention and attrition: Views of early career teachers. Aust. J. Teacher 34e55.
Educ., 38, 112e129. Miles, J., & Shevlin, M. (2001). Applying regression and correlation: A guide for stu-
Chabowski, B. R., Samiee, S., & Hult, G. T. M. (2017). Cross-national research and dents and researchers. Thousand Oaks, CA: Sage Publications.
international business: An interdisciplinary path. International Business Review, Mishler, E. (1979). Meaning in context: Is there any other kind? Harvard Educational
26(1), 89e101. https://doi.org/10.1016/J.IBUSREV.2016.05.008. Review, 49(1), 1e19.
Nielsen, W., Mena-MarcosJuan-Jose , J., Clarke, A., O’Shea, S., Hoban, G., & Collins, J.
Clarke, A. (2006). The nature and substance of cooperating teacher reflection.
Teaching and Teacher Education, 22(7), 910e921. (2017). Australia’s Supervising Teachers: Motivators and Challenges to Inform
Clarke, A., & Collins, S. (2007). Complexity Theory and the Supervision of Student Professional Learning. Asia-Pacific Journal of Teacher Education, 45(4), 346e368.
Teachers on Practicum. Teaching and Teacher Education, 23(2), 160e172. Osborne, J. W., & Fitzpatrick, D. (2012). Replication analysis in exploratory factor
Clarke, A., Collins, J., Triggs, V., & Neilsen, W. (2012). The mentoring profile in- analysis: What it is and why it makes your analysis better. Practical Assessment,
ventory: An online professional development resource for cooperating teachers. Research and Evaluation, 17(15), 1e8. Retrieved from: http://pareonline.net/
Teaching Education, 23(2), 167e194. getvn.asp?v¼17&n¼15.
Cochran-Smith, M., & Lytle, S. (1999). Relationships of knowledge and practice: Ronfeldt, M., Loeb, S., & Wyckoff, J. (2013). How teacher turnover harms student
Teacher learning in communities. Review of Research in Education, 24, 249e305. achievement. American Educational Research Journal, 50(1), 4e36. https://
https://doi.org/10.2307/1167272. doi.org/10.3102/0002831212463813.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Sarason, S. (1996). Revisiting. In The culture of the school and the problem of change.
Hillsdale, New Jersey: Lawrence Erlbaum Associates. https://doi.org/10.1016/ New York: Teachers College Press.
Scho€n, D. A. (1987). Educating the reflective practitioner: Towards a new design for
C2013-0-10517-X.
DeLong, E., DeLong, D., & Clark-Pearson, D. (1988). Comparing the areas under two teaching and learning in the professions. San Francisco: Jossey-Bass. https://
or more correlated receiver operating characteristic curves: A nonparametric doi.org/10.1002/chp.4750090207.
approach. Biometrics, 44, 837e845. https://doi.org/10.2307/2531595. Sinclair, C., Dowson, M., & Thistleton-Martin, J. (2006). Motivations and profiles of
Erickson, F. (1986). Qualitative methods in research on teaching. In M. Wittrock cooperating teachers: Who volunteers and why? Teaching and Teacher Educa-
(Ed.), Handbook of research on teaching (3rd ed., pp. 119e161). New York: tion, 22(3), 263e279. https://doi.org/10.1016/j.tate.2005.11.008.
Macmillan https://trove.nla.gov.au/version/44840180. Van Maanen, J. (1995). An end to innocence: The ethnography of ethnography. In
Faikhamta, C., & Clarke, A. (2018). Thai Cooperating Teachers’ Motivations and J. Van Maanen (Ed.), Representation in ethnography (pp. 1e35). Thousand Oaks
Challenges in Supervising Student Teachers During Their Internship Program. CA: Sage.
Kasetsart Journal of Social Sciences. Wertsch, J. (1991). Voices of the mind: A sociocultural approach to mediated action.
Feiman-Nemser, S. (2001). Helping novices learn to teach: Lessons from an exem- Cambridge: Harvard University Press. https://doi.org/10.2307/1423207.
plary support teacher. Journal of Teacher Education, 52, 17e30. https://doi.org/ Young, A. M., O’Neill, A., & Mooney Simmie, G. (2015). Partnership in learning be-
10.1177/0022487101052001003. tween university and school: Evidence from a researcher-in-residence. Irish
Games, P. A., & Howell, J. F. (1976). Pairwise multiple comparison procedures with Educational Studies, 34(1), 25e42. https://doi.org/10.1080/
unequal n’s and/or variances: A Monte Carlo study. Journal of Educational Sta- 03323315.2014.1001203.
tistics, 1, 113e125. https://doi.org/10.2307/1164979.