Nurse Education Today 30 (2010) 44–48

Contents lists available at ScienceDirect

Nurse Education Today

journal homepage: www.elsevier.com/nedt

The self-directed learning readiness scale for nursing education revisited: A

confirmatory factor analysis
Murray J. Fisher a,*, Jennie King b,c
a
Senior Lecturer and Director Pre-registration Programs, Faculty of Nursing and Midwifery, University of Sydney, NSW 2006, Australia
b
Clinical Nurse Consultant Research, Nursing and Midwifery Directorate, Northern Sydney Central Coast Area Health Service, NSW 2250, Australia
c
Clinical Senior Lecturer, Faculty of Nursing and Midwifery, University of Sydney, NSW 2006, Australia

a r t i c l e i n f o s u m m a r y

Article history: The Self-Directed Learning Readiness Scale for Nursing Education (SDLRSNE) was initially developed as
Accepted 30 May 2009 an alternative to Guglielmino’s [Guglielmino, L.M. 1977. Development of the Self-Directed Learning
Readiness Scale. Unpublished Doctoral Dissertation, University of Georgia. Dissertation Abstracts Interna-
tional, vol. 38 (11a), p. 6467] Self-Directed Learning Readiness Scale. The aim of this study was to re-
Keywords: examine the factor structure of the subscales of the SDLRSNE and provide evidence of its validity. Data
Self-directed learning was collected using a cross-sectional survey of 227 first year undergraduate nursing students. To examine
Self-directed learning readiness
the factor structure of the SDLRSNE three one-factor congeneric models, each representing a different
Nurse education
Learning styles
subscale, were tested with maximum likelihood confirmatory factor analysis. The model fit indices of
the three one-factor congeneric models indicate that the resultant models fit the data well, providing
support for the factorial validity of the SDLRSNE. Of the 40 items, 11 items had to be removed from
the analyses as they failed to provide good fit with their subscales. Further research investigating the fac-
tor validity of the SDLRSNE is encouraged, specifically to examine the stability of the items across factors
using multi-factor models.
Ó 2009 Elsevier Ltd. All rights reserved.

Introduction in repeated attempts to replicate the eight-factor structure of the

SDLRS proposed by Guglielmino (1977) has raised concerns of its
The Self-Directed Learning Readiness Scale for Nursing Educa- construct validity (Field, 1989, 1991; Hoban et al., 2005).
tion (SDLRSNE) was first published in Nurse Education Today in The development and validation of the Self-Directed Learning
2001 (Fisher et al., 2001) and since that time there have been many Readiness Scale for Nurse Education was conducted in three
requests for its continued use in nursing education research. stages (Fisher et al., 2001). Firstly, 93 items were developed fol-
O’Shea (2003) rightly claimed that caution is required in the use lowing a comprehensive literature search. The Delphi technique
of the SDLRSNE until there is further evidence of its validity and was used to gain consensus from an expert panel of 11 nurse aca-
reliability. The aim of this study was to re-examine the factor demics and nurse educators as to whether the items reflected the
structure of the subscales of the SDLRSNE and provide evidence characteristics required for SDL. The resultant 52 items were pi-
of its validity. loted on a sample of 201 undergraduate nursing students. Princi-
Self-directed learning (SDL) is a method of instruction that can pal components analysis with varimax rotation was used to
be defined in terms of the amount of responsibility the learner ac- establish the underlying factor structure of the items in the
cepts for his or her own learning (Fisher et al., 2001). The learner’s instrument. The final SDLRSNE consist of 40 items distributed
readiness to engage in SDL is defined as ‘‘the degree the individual across three underlying factors: ‘Self-management’, ‘Desire for
possesses the attitudes, abilities and personality characteristics learning’, and ‘Self-control’.
necessary for self-directed learning” (Wiley, 1983, p. 182).
The Self-Directed Learning Readiness Scale for Nursing Educa- Background
tion was initially developed as an alternative to Guglielmino’s
(1977) Self-Directed Learning Readiness Scale. Since the develop- The internal consistency of the SDLRSNE and its subscales has
ment of the Guglielmino SDLRS, there has been significant scrutiny been reported in several studies (Newman, 2004; Bridges et al.,
of its underlying theoretical constructs and factor structure. Failure 2007; Smedley, 2007). These studies demonstrate that the
SDLRSNE is reliable and internally consistent across various sam-
* Corresponding author. Tel.: +61 2 9351 0587; fax: +61 2 9351 0654. ples. A UK randomized experimental designed study, evaluating
E-mail address: m.fi[email protected] (M.J. Fisher). the effectiveness of problem-based nursing curricula, reported

0260-6917/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.nedt.2009.05.020
 M.J. Fisher, J. King / Nurse Education Today 30 (2010) 44–48 45

the internal consistency of the SDLRSNE to be 0.86 for ‘Self-man- exploratory factor analysis. Hendry and Ginns (2009) identified a
agement’, 0.85 for ‘Desire for learning’ and 0.89 for ‘Self-control’, four factor model that fit their data. Two factors (‘Self-determina-
and the total scale Chronbach’s coefficient alpha was 0.95. tion’ and ‘Effective organization for learning’) corresponded to
In a US study of physical therapists, the SDLRSNE was used the original ‘Self-control’ and ‘Self-management’ subscales, respec-
to determine whether a propensity for self-directed learning tively. The other two factors identified, ‘Critical evaluation’ and
correlated with the use of evidence-based practice (Bridges ‘Learning self-efficacy’, did not correspond well with the original
et al., 2007). In this study the Chronbach’s coefficient alpha factors reported by Fisher et al. (2001). This suggests that further
for a shortened version of the SDLRSNE was 0.87 for ‘Self-man- research is required to establish the factorial validity for the
agement’, 0.85 for ‘Desire for learning’ and 0.80 for ‘Self-con- SDLRSNE.
trol’. This study found that self-directed learning readiness (in
particular, the ‘Desire for learning’ subscale) was a significant
predictor of the use of evidence-based practice in physical Methods
therapists.
An Australian study that evaluated the self-directed learning Data collected from a cross-sectional survey of 227 first year
readiness of undergraduate students in their first year of the Bach- undergraduate nursing students was used to examine the factor
elor of Nursing (BN) programme, supported the high internal con- structure of the SDLRSNE. The SDLRSNE, together with a partici-
sistency of the SDLRSNE (Smedley, 2007). Smedley (2007) reported pant information sheet, was distributed by the researchers to the
the Chronbach’s coefficient alpha for the subscales to be: ‘Self- convenience sample during their course orientation week. Consent
management’ 0.81, ‘Desire for learning’, 0.78, and ‘Self-control’ was considered implied on completion and return of the survey.
0.84. The total score and subscale score distributions were consis- The participants anonymously returned the survey to a centrally
tent with the Fisher et al. (2001) sample. located box. No participant names or identifiable characteristics
A further study has examined the factor structure of the were collected on the survey. Ethical Approval for this study was
SDLRSNE. Hendry and Ginns (2009) examined the factorial validity granted by the University of Sydney Human Research Ethics
of the SDLRSNE in medical students using the FACTOR method of Committee.

.57

I am self disciplined e2

.31

I am disorganised e3

.35
-. 5 .75

I set strict time frames e4

6

.29
9
.5

I have good management skills e5

.54
.12
.3 4
I am methodical e6
SELF-MANAGEMENT .45 .20

I am systematic in my learning e7
.6
2

.38
.4
3

I set specific times for my study e8

.5 0

.18
.42

I prioritise my work e10

.25
I can be trusted to persue
e11
my own learning
.18
I am confident in my ability to
e13
search out new information

Fig. 1. One-congeneric model for ‘Self-management’ with standardized regression weights.

46 M.J. Fisher, J. King / Nurse Education Today 30 (2010) 44–48

The SDLRSNE cant multivariate non-normality in the data. As a result the Bol-
len–Stine bootstrap p procedure was used to adjust model fit and
The Self-Directed Learning Readiness Scale for Nursing Educa- parameter estimates to accommodate the lack of multivariate nor-
tion consisted of a total of 40 items across three subscales: mality. Standardized residual covariances (cut-off of >4) and mod-
Self-management (13 items), Desire for learning (12 items) and ification indices were used in an iterative process to respecify each
Self-control (15 items). Four items were negatively phrased. Partic- of the models.
ipants were asked to indicate the degree each item reflected their
own characteristics using a five-point Likert scale where a score of Results
1 indicated strongly disagree and a score of 5 indicated strongly
agree. The reported Chronbach’s coefficient alpha for the total scale The participants’ age ranged from 17 to 55 years with a median
in this study was 0.87. age of 18 years. Thirty (13.3%) participants were male and 196
(86.7%) were female. One participant did not indicate their sex.
Analysis When all 13 items were included in the analysis of the ‘Self-
management’ subscale the data did not fit the specified factor
To examine the factor structure of the SDLRSNE three one-factor model (v2(65) = 160.205; p < 0.0001). It was only after the deletion
congeneric models, each representing a different subscale, were of three items (I do not manage my time well; I solve problems
tested with maximum likelihood confirmatory factor analysis using a plan; I prefer to plan my own learning) that the model
using AMOS (version 17). The separate analyses of the three one- for self-management fit the data (v2(35) = 46.654; p = 0.662).
factor congeneric models were chosen over a higher-order or Fig. 1 presents the standard regression weights (factor loadings)
multi-factor model as a result of the relative small sample size. for the resultant 10 items in the model for Self-management.
Structural Equation Modelling techniques require large sample Similarly, the original 12 items of the ‘Desire for learning’ sub-
sizes to maintain stability in the parameter estimates. Examining scale did not fit the specified factor structure (v2(54) = 162.512;
the three one-factor congeneric models separately provided a ratio p < 0.0001). After 3 iterations and the deletion of three items (I like
of number of participants to number of parameters to be estimated to gather the facts before I make a decision; I like to evaluate what I
of at least 10:1. do; I am open to new Ideas) the model for ‘Desire for learning’ fit
Influential outliers were identified by examining the standard- the data (v2(27) = 30.342; p = 0.835). Fig. 2 presents the standard
ized residuals covariance matrix and Mahalanobis distances. Out- regression weights for the resultant nine items in the model for
lier cases were deleted accordingly. The data in this study was at ‘Desire for learning’.
the ordinal level, therefore the SEM assumption of multivariate The ‘Self-control’ subscale consists of 15 items. When all 15
normality was not possible. In addition, the Mardia’s coefficients items were included in the model the data was a poor fit
for multivariate kurtosis in each model was >3, indicating signifi- (v2(90) = 312.515; p < 0.0001). After five iterations and the

.72

I want to learn new information e1

.73

I enjoy learning new information e2

.46
5
.8

5 I have a need to learn e3

.8
.35
.6 8

.6 0 I enjoy a challenge e4
.13
DESIRE FOR -.36
LEARNING .3 7 I do not enjoy studying e5
.14
.4
1
I critically evaluate new ideas e6
.4

.17
0
.27

I learn from my mistakes e10

.16

I need to know why e11

.07
When presented with a problem
I cannot resolve, I will ask for e12
assistance

Fig. 2. One-congineric model for ‘Desire for Learning’ with standardized regression weights.
 M.J. Fisher, J. King / Nurse Education Today 30 (2010) 44–48 47

removal of five items (I prefer to set my own goals; I like to make Discussion
decisions for myself; I am logical; I have high personal expecta-
tions; I prefer to set my own criteria on which to evaluate my per- The resultant models do provide reasonable evidence of con-
formance) the model had good fit with the data (v2(35) = 57.339; struct validity for the ‘Self-management’, ‘Desire for learning’ and
p = .214). Fig. 3 presents the standard regression weights for the ‘Self-control’ subscales of the SDLRSNE. However, it is clear that
resultant ten items in the model for ‘Self-control’. there is some difference in the factor structure of the instrument
The model fit indices of the three one-factor congeneric models when compared with the original exploratory factor analysis
are presented in Table 1. With multiple goodness-of-fit indices, (Fisher et al., 2001). For the majority of the items the proposed ori-
good fit is indicated by Root Mean-Square Error of Approximation ginal factor structure still hold true, however, in these analyses 11
(RMSEA) values lower than 0.05, Standardized Root Mean-Squared items have become redundant. On the surface, the items that have
Residual (SRMR) values lower than 0.05, Comparative Fit Index not loaded to their corresponding subscale do not appear to be the-
(CFI) values higher than 0.95, and Goodness-of-Fit Index (GFI) oretically essential. As a result a revised 29 item SDLRSNE could be
>0.95. The model fit indices for ‘Self-management’ and ‘Desire for considered following further examination of the factor validity.
learning’ indicate that the data has good fit with the resultant mod- A limitation in this study was the relatively small sample size. A
els. The ‘Self-control’ model has reasonable fit with borderline sample size of 227 participants limited the analyses to one-factor
RMSEA and CFI indices. congeneric models and did not allow for the examination of

.38
I am responsible for my own
e3
decisions/actions
.01

I am not in control of my life e4

.30

e5
1

I have high personal standards

.6
0

.09
-.1

5 I prefer to set my own learning goals e6

.5

.3 0 .16

I evaluate my own performance e7

.4 0
.35
SELF-CONTROL .5 9
I am responsible e9
.4 5
.20
. 52
I am able to focus on a problem e11
.5
6
.27
.6

I am aware of my own limitations e12

0

.31

I can find out information for myself e13

.36

I have high beliefs in my abilities e14

Fig. 3. One-congineric model for ‘Self-control’ with standardized regression weights.

Table 1
Goodness-of-fit indices for one-factor congeneric models.

Goodness-of-fit indices One-factor congeneric model

Self-management (10 items) Desire for learning (9 items) Self-control (10 items)
v 2 (Bollen–Stine p) (df) 46.654 (p = .662) (35) 30.342 (p = .835) (27) 57.339 (p = .214) (35)
RMSEA .039 .024 .054
SRMR .035 .032 .031
GFI .960 .971 .951
GFI–AGFI .023 .020 .028
CFI .971 .993 .930
48 M.J. Fisher, J. King / Nurse Education Today 30 (2010) 44–48

relationships between variables (items) across factors. In order to References

examine item interactions across the subscales, higher-order or
multi-factor models need to be employed. Such analyses require Bridges, P.H., Bierema, L.L., Valentine, T., 2007. The propensity to adopt evidence-
based practice among physical therapists. BMC Health Services Research 7,
much larger samples. It is possible that those items found to be 103–112.
redundant in this analysis may have significant interactions with Field, L., 1989. An investigation into the structure, validity and reliability of
other subscales. This is supported by Hendry and Ginns (2009). Guglielmino’s self-directed learning readiness scale. Adult Education Quarterly
39 (3), 125–139.
In their four-factor solution (made available on request), factor 1 Field, L., 1991. Guglielmino’s self-directed learning readiness scale: should it
(‘Critical self-evaluation’) consisted of five items, which explain continue to be used? Adult Education Quarterly 41 (2), 100–103.
the largest percent of the variation in their solution. Of the five Fisher, M., King, J., Tague, G., 2001. The development of a self-directed learning
readiness scale for nursing education. Nurse Education Today 21, 516–525.
items three items (I like to evaluate what I do; I like to gather
Guglielmino, L.M. 1977. Development of the Self-Directed Learning Readiness Scale.
the facts; I am logical) were redundant in this current study. It is Unpublished Doctoral Dissertation, University of Georgia. Dissertation
the authors recommendation that the 40 item SDLRSNE should Abstracts International, vol. 38 (11a), p. 6467.
be used until further research examines the relationships between Hendry, G.D., Ginns, P., 2009. Readiness for self-directed learning: validation of a
new scale with medical students. Medical Teacher, in press, doi:10.1080/
variables (items) across factors in different samples. 01421590802520899.
Hoban, J.D., Lawson, S.R., Mazmanian, P.E., Best, A.M., Seibel, H.R., 2005. The self-
directed learning readiness scale: a factor analysis study. Medical Education 39
Conclusion (4), 370–379.
Newman, M. 2004. Problem based learning: an exploration of the method and
The SDLRSNE was developed to measure readiness for SDL in evaluation of its effectiveness in a continuing nursing education programme.
Project on the Effectiveness of Problem Based Learning (PEPBL) Research Report,
nursing students. The validity of the scale is supported by the re-
Middlesex University, Middlesex.
sults of this study. The SDLRSNE will continue to be a useful tool O’Shea, E., 2003. Self-directed learning in nurse education: a review of the
in the diagnosis of student learning needs. Further research inves- literature. Journal of Advanced Nursing 43 (1), 62–70.
Smedley, A., 2007. The self-directed learning readiness of first year bachelor of
tigating the factor validity of the SDLRSNE is encouraged, specifi-
nursing students. Journal of Research in Nursing 12 (4), 373–385.
cally to examine the stability of the items across factors using Wiley, K., 1983. Effects of a self-directed learning project and preference for
multi-factor models. structure on self-directed learning readiness. Nursing Research 32 (3), 181–185.