Massed practice or distributed practice: Comparative study for best learning strategies

Abstract

The purpose of the study is to discover which learning methods, massed practice or
distributed practice, produces the most efficient results. The study also tested for the more
reliable testing apparatus; fill-in-the-blanks or multiple-choice questioning. Previous studies
suggest that distributed practice is more valuable regarding the retention interval. The study
tested 274 students engaged in online study. Test materials used were fill-in-the-blanks and
multiple-choice tests. Participants spent 6 hours studying in one sitting for massed practice,
and 2 hours studying on 3 consecutive days for distributed practice. Results showed that
distributed practice has an advantage over massed practice. Results also showed a better
outcome for fill-in-the-blanks than massed practice, although mass practice has better results
when multiple-choice questioning is utilised. The study supported the hypothesis that
distributed practice was a better strategy to employ overall. The study did not support the
hypothesis that fill-in-the-blanks was a superior instrument to use to assess learned material
over multiple-choice questioning. Limitations of the study indicate that, although multiple-
choice questioning produced more correct answers, it is not possible to ascertain which
questions in the multiple-choice questioning study were learned responses or guesses.
Another limitation identified is that gender is not considered as an independent variable.
 Learning is the most fundamental approach to acquiring knowledge, and so it is
beneficial to ascertain which is the most effective learning method. Through the course of
this study, the intention is to discover which of the two methods, massed practice
(information to disseminated in one session) or distributed practice (practice distributed
across multiple sessions), brings the most efficient results. As pointed out by Rohrer and
Taylor (2006, p.1), the first action is to learn a skill and then, by practice, improve. The
improvement is determined by distributed practice over the duration that the student applies
their skill. In Rohrer and Taylor’s (2006, p. 2) study, they suggest that mass practice may
have an impact upon learning retention. Alongside this, Carpenter, Cepeda, Rohrer, Kang,
and Pashler (2012, p. 4), have also discovered that an inter-session interval (ISI) of five
weeks with seven-day intervals of learning, produced better results than a retention interval
(RI) of zero days.

The study by Rohrer and Taylor (2006), tested participants over two independent
variables; study strategies (massed vs. distributed practice) and test type (fill-in-the-blanks vs.
multiple-choice test). By testing these against the dependent variable, which is their test
performance, this study’s motivation is to ascertain both the relationship between learning
outcomes for massed vs. distributed practice and the which is the more effective test
instrument; fill-in-the-blanks or multiple-choice tests.

Over several examples, the study by Carpenter et al. (2012) detects that distributed
practice brings greater rewards than the cramming nature of mass practice. The finding by
Carpenter et al. (2012, p. 4) was related to the nature of long-term retention, discovering that
students engaged in distributed practice had better retention than those of massed practice
(12.2% v 8%). However, spacing (the time spaced interval between study sessions), does not
indicate an improved long-term memory retention. Further study may identify a prime period
in which effective testing should take place in order to counter memory loss. A study by
Nazari and Ebersbach (2018, p. 11), raised concerns as to whether improvement remained if
students undertook less study time under the conditions of distributed practice. Their
findings showed that using distributed practice, under semi-self-regulated conditions, female
students are more successful (Nazari & Ebersbach, 2018, p. 10).

The studies of Rohrer and Taylor (2006), and Carpenter et al. (2012), suggest in these
studies that distributed practice is of more benefit with regard to RI. These studies have
identified differences between mass practice and distributed practice, so it would be
beneficial to consider the prime period before memory retention begins to dissipate. The
advantage of a study of this nature is to be able to set parameters for the study/learning
relationship, and so allow students optimum performance. However, Donovan and
Radosevich (1999, p. 801) indicated that, although ISI was beneficial, the effects reduced in
efficacy as the complexity of the testing increased; for example, low (rote memory), average
(recalling lists) and high (puzzle solving).

It is also necessary to consider the testing methods which were used. Rohrer and Taylor
employed multi-choice questions (MCQ) and fill-in-the-blanks (FITB). When using the
MCQ instrument, it is necessary to consider the 1:4 chance of guessing a correct result with a
four choice MCQ question, which can clearly skew the results when considering the
difference between types of learning. A study by Medawela, Ratnayake, Abeyasinghe,
Jayasinghe and Kosala (2018, p. 7), assessed dental students in their final year, showing that
FITB was a truer assessment of the students’ abilities than MCQ; MCQ only addresses the
lower levels of knowledge due, to the opportunity to guess an answer.

Hypothesis 1 It was hypothesised that distributed practise has more benefit than massed
practice in the domain of retained memory for learning.

Hypothesis 2 It was hypothesised that fill-in-the-blanks is a more reliable means of

assessing learning than MCQ.

Method

Participants

The study tested 274 students engaged in online study through the Swinburne Psychology
Research Experience Program. The demographic of the participants was

 Men 27.62%,
 Women, 70.71%
 Other or Unidentified 1.67%.

The age of the participants was 18-65 years old, the mean 21.34 and standard
deviation 3.12. Of the students enrolled in the study, 35 did not complete and so their
data was removed, leaving a final sample of 239 participants.

Materials

For the this study the materials consisted of 4 groups;

 1) material from research methods text book.

2) Test materials

i) fill-in-the-blanks test condition: a 20-item fill-in-the-blanks test.

ii) multiple-choice test condition: a 20-item multiple-choice test.

3) Instructions

i) massed practice condition: Spend 6 hours studying this material in one sitting.

ii) distributed condition: Spend 2 hours studying the material on 3 consecutive

days.

4) Demographics questions: 'Please indicate your age (in years)' and 'Please
indicate your gender.'

Procedure

Approval was given by the Swinburne University Human Research Ethics Committee
(SUHREC), with participants being informed that they could withdraw from the study at any
time without penalty. The study duration was 30 minutes. Inducement for participation was
one course credit. As the study element of the research was a component of the students’
revision, there was no remuneration for the study stage.

Sign-up was through the Research Experience Program using Qualtrics (computer
software). Part 1 of the study involved; assigning participants to study strategy conditions,
presenting the relevant instructions, and answering demographic questions. Participants of the
study were directed to log into Qualtrics three weeks after their final study session to
complete Part 2 of the study. Participants read the consent information statement and agreed
to participate in the study.

Participants studied both strategy conditions (massed vs. distributed practice), spending 6
hours studying for both tests (fill-in-the-blanks vs. multiple-choice) containing 20 items. One
half studied for 6 hours on 1 day (massed practice condition), the other half studied for 6
hours across 3 consecutive days (distributed practice condition). The tests were completed 3
weeks after their final study session. Half of the participants completed fill-in-the-blanks test
questions; the other half completed a test with multiple-choice questions.
 Participants were randomly assigned a study strategy. One half of the participants were
given the massed practice condition instructions; the other half received the distributed
practice condition instructions. The participants had 5 minutes to complete Part 1 of the study
and 25 minutes to complete the test in Part 2.

Debriefing was undertaken on completion of the study identifying the purpose of the
study, the manipulations, and the hypotheses explored by the researchers.

Results

Participants were tested over 2 independent variables for study strategies (massed vs.
distributed practice) and also test type (fill-in-the-blanks vs. multiple-choice test), with 4
different groups; massed practice with a fill-in-the blanks test; massed practice with a
multiple-choice test; distributed practice with a fill-in-the blanks test; distributed practice
with a multiple-choice test.

Massed practice results identified that correct results for fill-in-the-blanks tests show a
mean of 12.36/20. (SD 1.53). Distributed practice gives a mean result of 15.47/20 (SD 1.52).

Massed practice identifies that correct results for MCQ show a mean of 17.03/20. (SD
2.87). Distributed practice gives a mean result of 16.59/20 (SD 1.45).

Overall the results show that correct results using massed practice are 14.70/20. (SD
2.31). Distributed practice gives a mean result of 16.03/20 (SD 1.49).

The results show that overall, distributed practice brings more correct answers, when
tested over 20 questions with a result 80.15% as compared to 73.5% for massed practice.

Results for fill-in-the-blanks with massed practice show correct answers 61.8% with
distributed practice 77.35%. MCQ and massed practice show correct answers 85.15% with
distributed practice 82.95%

The results indicate that overall distributed practice has an advantage over massed
practice with a greater success rate of answering correct questions at a rate of 6.65%. In
relation to fill-in-the-blanks distributed practice shows an advantage over massed practice of
15.7%. The results for MCQ show a greater outcome of correct answers for massed practice
over distributed practice by 2.2%.
 The results show that distributed practice has an overall advantage over massed practice.
The performance when correlated with fill-in-the-blanks shows a high performance against
massed practice, although massed practice has a slightly better result when MCQ is used.

18

16

14
Correct responses /20

12

10

0
Massed Distributed

Fill in the Blanks MCQ

Figure 1: Descriptive statistics for number of correct responses as a function of study strategy
and type of test

Discussion

The intention of this study was to ascertain whether distributed practise was more
beneficial than massed practice when considering memory retention for study. The study
supported the hypothesis that distributed practice was a better strategy to employ overall. A
second hypothesis was to determine whether the FITB testing instrument was a more reliable
means of assessing learning than MCQ. The study did not support the hypothesis that FITB
was a better instrument to use to assess learned material than MCQ. The results indicate that
MCQ produced more correct answers. It is not, however, possible to ascertain with this study
instrument, which questions in the MCQ study were learned responses or guesses, as there is
a 1:4 chance of guessing a correct answer. This raises a question regarding the efficacy of
MCQ, as suggested by Medawela et al. (2018, p. 7). Their study raises a question regarding
the level of learning that is acquired. The results of this study indicate that distributed
practise shows a marked improvement over massed practice when ISI is three days. Indeed, it
would be beneficial to ascertain what length of time is ideal for maximum retention of
information. The concern expressed by Donovan and Radosevich (1999, p. 801), regarding
the level of retention that is acquired, and maintained, depending upon the complexity of the
testing, needs to be part of further studies that allows for consideration of the most
appropriate time to test after completion of study. Consideration must be given that some
testing requires less concentration and, able to accommodate MCQ more effectively. For
example, testing that requires rote learning may need less, distributed practise learning,
depending upon the nature of the data required to be recalled. Additionally, the study by
Nazari and Ebersbach (2018, p. 10), indicated that girls were better at completing distributed
practice studies than boys under certain conditions. Although gender is not considered in this
study, its inclusion in future research could lead to further insight with regard to the impact of
gender upon learning studies. Consequently, a limitation of this study was the non-inclusion
of gender statistics as the study by Nazari and Ebersbach (2018, p. 10) identified gender as an
independent variable. Gender could be identified as a factor when ascertaining the efficacy of
mass practice v distributed practice. Another limitation is not knowing the amount of massed
practice answers were guessed. This could be accounted for in future studies by asking the
participant to indicate which answers they were guessing. This would give a clearer
indication of learning acquired during the ISI period.

In conclusion, distributed practice shows a more efficient study strategy than mass
practice. Distributed practice could be employed to bring forward a more efficient means of
studying, as the retention level of learning is higher than massed practice. It can, therefore, be
employed for more complex learning strategies. We can see that a greater emphasis should
be placed on distributed practice than mass practice to allow better learning retention. With
regard to the instruments used, MCQ shows more correct answers than FITB. However,
although providing more correct answers, it is not possible to check whether responses to
MCQ have been guessed, and so the level of learning is not clear. We must be concerned that
MCQ can give a misleading representation of the learning which has been acquired. To make
these studies more rounded, it is also necessary to acknowledge that gender is an important
aspect which was not included in this study. Future research could include gender as a
component to ascertain any difference that may exist with regard to gendered learning. In
addition, if MCQ is to be continued as a test instrument, we must also create a self-
identifying response for guessing, in order to bring more reliable test results to our study.
 References

Carpenter, S. K., Cepeda, N. J., Rohrer, D., Kang, S.H.K., & Pashler, H. (2012). Using
spacing to enhance diverse forms of learning: Review of recent research and implications
for instruction. Educational Psychology Review, 24(3), 369–37. DOI: 10.1007/s10648-
012-9205-z

Donovan, J. J., & Radosevich, D. J. (1999). A meta-analytic review of the distribution of

practice effect: Now you see it, now you don't. Journal of Applied Psychology (84), 795-
805. DOI: 10.1037/0021-9010.84.5.795

Medawela, R. M. S. H. B., Ratnayake, D. R. D. L., Abeyasinghe, W. A. M. U. L., Jayasinghe,

R.D., & Kosala, N. M. (2018). Effectiveness of “fill in the blanks” over multiple choice
questions in assessing final year dental undergraduates. Educación Médica 19(2), 72-76
DOI: https://doi.org/10.1016/j.edumed.2017.03.010

Nazari, K.B., & Ebersbach, M. (2018). Distributed Practice: Rarely Realized in

SelfRegulated Mathematical Learning. Frontiers in Psychology 9, 22170.
DOI=10.3389/fpsyg.2018.02170

Rohrer, D., & Taylor, K. (2006). The effects of overlearning and distributed practice on the
retention of mathematics knowledge. Applied Cognitive Psychology, 20, 1209–1224.
DOI: 10.1002/acp.1266