Impact of Age on Capabilities to Process Recall

This study analyzes the effect of age on depth of processing effects on recall. The study uses incidental learning tasks to study the effect of age on semantic and non-semantic tasks. The results clarify the effect of age-related depletion in incidental learning.

Impact of Age on Capabilities to
Process Recall
Student Name: Student ID:
Unit Name: Unit ID:
1 | P a g e

Abstract
Incidental learning tasks were performed by two subject groups old and young. The effect of age
was studied on semantic and non-semantic tasks. Two tasks involved non-semantic analysing
power whereas other two included semantic process analysis. One control (intentional) task was
also performed. Using descriptive statistics the effect of age on in depth analysis was studied by
dividing the experiment data into 2age x 5orienting tasks. The results clarified the effect of age
related depletion in incidental learning which is evident from the ability of depth of processing
the tasks.
2 | P a g e

Table of Contents
Abstract.......................................................................................................................................................2
Table of Tables............................................................................................................................................3
1.0 Introduction...........................................................................................................................................4
2.0 Method..................................................................................................................................................6
3.0 Results...................................................................................................................................................9
4.0 Discussion............................................................................................................................................15
5.0 Reference Lists.....................................................................................................................................17
6.0 Appendix..............................................................................................................................................20
Table of Tables
Table 1: Mean number of words recalled as a function of Age Group and Recall pattern.........................13
Table 2: Age * Recall type measures.........................................................................................................13
Table 3: F values for k-clustering data......................................................................................................15
Table 4: Descriptive statistics for the entire recall pattern........................................................................20
Table 5: Multivariate Test results for the sample......................................................................................21
Table 6: Within- subjects effects and analysis...........................................................................................22
Table 7: Within- subjects contrasts analysis..............................................................................................22
Table 8: Test result between- Subjects effects..........................................................................................23
Table 9: Estimated marginal means...........................................................................................................23
Table 10: Levene's Test of Equality of Error Variancesa............................................................................24
Table 11: Pair-wise Comparisons of recall types.................................................................................24
Table 12: Multivariate Tests results...........................................................................................................25
Table 13: Normality test of the sample based on gender..........................................................................26
Table 14: Normality test of the sample based on gender..........................................................................27
3 | P a g e

Table 15: Independent sample t-test for gender.......................................................................................27
Table 16: Recall with imagery test.............................................................................................................28
Table 17: Recall with counting test............................................................................................................28
Table 18: Recall with adjective test...........................................................................................................28
Table 19: Recall with rhyming test............................................................................................................29
Table 20: Recall with control test..............................................................................................................29
Table of figures
Figure 1: A priori sample size G-Power normality curve............................................................................11
Figure 2: Box plot of the two genders.......................................................................................................12
Figure 3: Box plot of the two age groups...................................................................................................12
Figure 4: Profile plot of estimated Marginal means of two age groups.....................................................26
Figure 5: Power curve for normality test...................................................................................................27
4 | P a g e

1.0 Introduction
Learning is the process of human development from childhood through adulthood
(Botwinick 2013). Various psychologists have defined and segregated cognitive and learning
skill developmental according to age groups. Psychology based experiments conducted by well-
known scholars depicts that younger people are at more advantage in learning compared to older
adults, who are less likely to be amendable. The scope of the current study aims at analyzing the
following research question; Does age influence depth of processing effects on recall?
Significance of the Study: Craik (1968) in his experiment found that while easy learning such as
recalling color names did not have much distinction amongst age groups, there was differential
learning when tested for scrambled proverbs and texts(Eysenck 1974). Studies depict that coding
capabilities in cases of color names is highly different compared to textual materials. Craik and
Lockhart (1972) concluded that coding capabilities associated with proverbs or scrambled texts
includes greater degree of cognitive or semantic analysis. Young are capable of process of
remembering information in greater depths as compared to their older counterparts, it is assumed
to be working hypothesis with the gap being processing-deficit hypothesis (Lee 2013). From the
hypothesis taking together an assumption of retention being directly linked to processing depth, a
large number of testable deductions are arrived at.
In order to test for processing-deficit hypothesis multiple incidental-learning paradigm
can be applied. In a study conducted by Hyde and Jenkins (1973), participants were asked to
perform orienting tasks with list of words but they were not alarmed regarding recalling of t hose
words (Kausler 1965). Each of the orienting tasks differed on basis of their processing
requirements. While recalling of orienting task, there was a prerequisite regarding their
processing needs for understanding whether these words needed to be semantically processed.
5 | P a g e

Semantic tasks involved detailed processing of information hence needed more recalling
compared to other tasks. According to the hypothesis of processing-deficit, age-related recall
decreases as depth of processing of a task increases.
Variables: Age group, gender, 5 different task groups were variables of the study. Gender was
later excluded from variable list due to the fact that it did not have significant effect on the
number of words recalled by the subjects.
Hypothesis: The study consisted of the null hypothesis which was as follows:
H0: Age does not have any effect on the in depth processing for recalling information
Earlier studies aimed at analysing incidental learning at different ages in an adult’s life-
span have not been able to conclude (Alfieri 2011). Conclusions derived from these studies have
mostly been haphazard as some studies depict decreasing effects on age-based learning, whereas
insignificant deterioration was noted as age progressed. All the studies conducted earlier reflects
minor experimental control on subjects with incidental elements, therefore an effort to establish
connection between age and incidental based learning was not successful. This study attaches
great importance of natural learning with incidental variety, which is a systematic probe rather
than connecting realistic data of age-related learning (Gegenfurtner 2012).
6 | P a g e

2.0 Method
Participants: The study had followed appropriate research methods to allow it arrive at results
easily. The research design included 5 X 2 factorial analysis of variance model, where 5 is the
Orientating Task and 2 is the Age, both factors consisted of between subjects. Tasks were
assigned randomly to 10 groups of participants, where each group comprised of 5 female and 5
male participants.
Subjects
In order to conduct the study fifty college students were taken into consideration. The
participants took part in the study for exchange of either a payment or as a part of their course
needs. The participants in the study were within age of 18 years to 30 years of age along with
older participants, school teachers who were 50 in number of age 55 years to 65 years old. When
the two groups were given Mill hill Vocabulary Scale, their test results did not differ
significantly. This result reflected that sample participants selected were superior as per
educational levels. Sample selected was from urban middle class.
Ethics: The scholar for the purpose of this study has adopted ethical procedure. The scholar has
obtained permission from individual participants, prior to conducting the study. The objective of
the study and all other relevant details were discussed. The scholar did not disclose any personal
details like name, address, contact details or any other details in the study, thus maintaining the
confidentiality agreement. The scholar obtained ethics approval from ethics committee and
2017/123 as ethics number.
7 | P a g e

Materials
In order to conduct the experiment, nine categories of Battig and Montague (1969) norms was
accommodated by a 27 word list of three high-dominance members. The words given were in
English Language from one syllable and two other words rhyming with other words were used.
But within the list none of the words rhymed with each other, with the list being arranged in
random manner, apart from successive exemplars within a category, which were separated by a
minimum of three intervening words from separate categories. For clarity of information and
highlighting the same, words were printed in capital letters of 2.5 centimeters high on separate
cards measuring 10 X 20 centimeters.
Data Analysis: IV & DV
The independent variable between the groups was taken as age group whereas gender was
neglected for the study. The five different types of tasks were taken as the within the groups
dependent variables.
Tasks
The experiment made use of five orienting tasks which were self-paced and given to individual
participants for responding.
Letter Counting: The participants were asked to conduct counting and then recording number
of letters of words throughout the pack of cards (Bal 2011). In the second round through the pack
of cards, participants were made to record post counting the number of consonants present in
each word. Last time, participants had to count and then record number of letters of words before
the letter M in the words.
8 | P a g e

Rhyming: In this task participants had to figure a word that rhymed with another word in the
list. Then they had to say out the rhyming words aloud. As they were on the verge of completing
pack of cards, they were made to go through the pack again and brainstorm another rhyming
word for each word (Kausler 2012). Participants could go through the pack for another time in
case of availability of time in case they could think of another rhyming word for words provided.
Adjective: In this, task participants had to identify a modifying adjective from the list of words
and say it loud. In this task as well, participants had to go through the pack for the second time
for brainstorming a different adjective for modifying adjective within word list (Arciuli 2011). In
this task also, participants had to go through the pack again for another adjective in case time
was there.
Imagery: In this task participants had to develop an image of words present in the list. Each of
the images had to be then rated on a 5-point scale from 1 to 5, where 1 implied no image and 5
was depiction of clear and vivid normal vision. Participants upon completion of the pack had to
rate again regarding prominence of the image (Christman 2011). Participants were asked to
rerate images again in case time permitted.
Control (Intentional): Participants had to work systematically through the pack for learning as
many words they could. In this task also participants had to go through the pack for second and
third time in case time was there.
Procedure
The experiment followed by procedure that was determined prior to conducting the same. In the
various tasks that was conducted throughout the experiment, each participants selected were
tested individually (Lee 2013). Participants were advised to gradually progress on the task, but
9 | P a g e

would have to move on to the next one. It would only happen in case participant spent more than
10 seconds in a particular task. Each participant was given cards to work at his own rate. In all
totality of conditions analyzed, participants had gone through 150 seconds for conducting the
task or while working through the pack. In order to write for free recall, participants were given
10 minutes post removal of list (Melton 2014). Participants at the end were interviewed
regarding whether they could understand if they would be required to recall.
3.0 Results
Normality of the data
A normality check was performed for the given data. The female subject inclusive of
both age groups yielded a measure of 0.521 for skewness and -0.803 for kurtosis along with
0.319, 0.628 for respective standard errors. The skewness z-value for female subject was 1.633
and -1.278 which were within the limits of -1.96 to 1.96 under 5% level of significance. The
male subject similarly yielded skewness z-value of 1.45 and -0.8646 for 5% level of significance.
Similar analysis for the young (skewness z-value was0.1276 and-2.0755) and older subjects
(skewness z-value was 2.7003 and 2.0846) also showed that the data was significantly not
normal. From Shapiro-Wilk test we got significance value less than 0.05 for gender as well as
age group analysis (table 13 and 14 of Appendix).
10 | P a g e

Sample size comparison to an estimated a priori sample size
A study was performed regarding the sample size using G*Power with 5 groups of recall
types consisting of 20 subjects each. Effect size was calculated using Cohen’s-d as 1.23083
power of the test was taken as 0.95. The effect size was determined taking 5 within sample
groups and their means. The graphical plot for F-test along with the critical F (3.05557) value
plot justified the sample size of 20 for analysis. The result was used for mixed ANOVA later.
From the value of the Cohen’s-d it was evident that sample size of 20 was sufficiently large for
our purpose (figure 5 Appendix).
0
0.2
0.4
0.6
0 5 10 15 20 25
critical F =3.05557
αβ
Figure 1: A priori sample size G-Power normality curve
Gender independence of the recall data
The recall scores were analysed with the gender. An analysis of variance was used for
this purpose and it was established from the almost equal means that gender did not differ
significantly on the basis of number of recalls. The calculated test statistic value of 0.574 (t-
statistic) was well inside the acceptance region of the null hypothesis that both gender group
11 | P a g e

irrespective of age group were almost similar in semantic or non-semantic recall pattern. So
gender was not a factor in the main analysis of data.
Figure 2: Box plot of the two genders
Figure 3: Box plot of the two age groups
12 | P a g e

Mixed ANOVA for age on recall scores
The data was analyzed by arranging in 5×2 (Recall pattern X Age group) pattern. There
was a significant effect of recall type (F=47.576, with p<0.001) and interaction of age group with
recall pattern (F= 5.976, p<0.0001). The non-semantic recall pattern was almost similar for both
age groups whereas semantic recalls (adjective, imagery, intentional) were better performed by
age group of young. The age group also had an effective implication on the analysis with
F=28.996 where p<0.0001. Hence null hypothesis “Age group does not have any effect on recall
pattern” was rejected for 5% level of significance.
Table 1: Mean number of words recalled as a function of Age Group and Recall pattern
Mean S.D Mean S.D Mean S.D Mean S.D Mean S.D
young (1) 6.5 1.434 7.6 1.955 14.8 3.49 17.6 2.591 19.3 2.669
old (2) 7 1.826 6.9 2.132 11 2.494 13.4 4.502 12 3.742
Letter counting Rhyming Adjective Imagery Intentional
Age group
The mean number of words recalled for five types of tasks helped to categorize the
semantic and non semantic jobs in order of variation with respect to the age group of the
subjects. Intentional (control), imagery, adjective types had mean values greater than the letter
counting and rhyming fields for both subjects. Whereas the noticeable difference between the
numerical values of the means for semantic fields was evident. It helps to establish that older age
group subjects were lagging behind young subjects in semantic fields but not otherwise. The p-
value of the letter counting field suggested the fact that age group was not a decisive factor for
basic non semantic jobs and for this category null hypothesis was accepted. The p- value as well
as the length of the confidence interval for rhyming was also indicative of the fact that both age
groups were almost equal in efficiency of analysing.
13 | P a g e

Table 2: Age * Recall type measures
age Mean
Std.
Error
95% Confidence
Interval
Lower
Bound
Upper
Bound
1 1 6.500 .519 5.409 7.591
2 7.600 .647 6.241 8.959
3 14.800 .959 12.785 16.815
4 17.600 1.161 15.160 20.040
5 19.300 1.028 17.141 21.459
2 1 7.000 .519 5.909 8.091
2 6.900 .647 5.541 8.259
3 11.000 .959 8.985 13.015
4 13.400 1.161 10.960 15.840
5 12.000 1.028 9.841 14.159
The interaction data also exposed the old group of subjects in case of semantic recalling
with respect to the young group of subjects. The conclusions were supported by the F values
from analysis of variance of age group with five types of recalls. For letter counting and rhyming
values were noted below 1 and p values were above 0.025, hence concluded that age group had
no effect on non-semantic recall. Whereas in situations of adjective, imagery and intentional
recall F values were 7.848, 6.539 and 25.229 with p values less than 0.025. So it was concluded
that in case of semantic recall null hypotheses were rejected. The confidence intervals clearly
14 | P a g e

gave a better spread for semantic tasks such as imagery and adjective compared to the letter
counting and rhyming tasks. The interval lengths of control group were of great significance
irrespective of both age groups. It is also evident that older age subjects were better off in control
and semantic tasks but their spread of the confidence interval lengths were less compared to the
young subjects.
Categorical clustering
From the analysis of variance in categorical clustering (F=23.417, p<.0001) values clearly state
the fact that under control or intentional group the given data shows highly significant presence.
The F values of letter counting and rhyming also established the claim of being non-significant
irrespective of subject category. Hence in depth analysis was mostly clustered around the
semantic tasks control, imagery and adjective. The analysis was performed at 5% level of
significance with degrees of freedom 1/18.
Table 3: F values for k-clustering data
Cluster Error
F Sig.
Mean
Square df
Mean
Square df
letter
counting
3.333 1 2.579 18 1.293 .270
Rhymin
g
16.875 1 3.382 18 4.990 .038
adjectiv
e
61.633 1 9.787 18 6.297 .022
imagery 120.00
0
1 11.722 18 10.237 .005
control 258.13
3
1 11.023 18 23.417 .000
15 | P a g e

4.0 Discussion
The results of the experiment were highly encouraging due to the fact that they resembled
with previous knowledge of the subject, though somehow under different fields of study and
conditions (Melton 2014). The semantic tasks produced highly significant data compared to the
non-semantic jobs. The study was split into 5 groups with 2 subjects in them, which helped in
analysing the performance of both subject groups in all fields. The young subject group found
the semantic jobs more appealing than older ones, though mean scores were better in magnitude
for both the subject groups for last three categories of tasks (Nasir 2010). The control task was
the most effectively performed category for young and old subjects. This indicates that for
categorized tasks old subjects also can use their depth of analysis and performance level of
execution was raised. The young were ahead in mean scores and length of C.I also high. For the
other
Both subjects exhibited almost same amount of efficiency while performing letter
counting, rhyming tasks. But the young subjects were more efficient in control task. This result
signifies the application level of both subject groups while performing in depth analysis. The
clustering of data revealed that age group difference was having insignificant effect in non-
semantic tasks (van der Stel 2008). The control task clustering completely exposed the older
subjects but that could be due to disassociation of item wise recognition.
The study requires further analysis considering the strategy fields of subjects. Irrespective
of age subjects were able to perform the recalling task quite well. The tasks required more fields
of category in each section. This kind of cross referencing analysis should also be conducted on
repetitive time dependent data. If the subjects were interviewed with prior knowledge of the test
information then this study could have shown more decisive results. In the industry, execution of
16 | P a g e

different tasks is generally done with prior knowledge. Hence for conclusive decision more
critical analysis along with complex data set is required. The absence of item wise word list
could also have yielded highly significant data in semantic tasks (Laurence 1967). The
confidential nature of the recall test though preserved the unbiased nature of data collection for
the study.
Finally this cross referencing analysis was an indication to the fact that age differences
can only be separated on the basis of low level of uncategorized tasks. The limitation of the
experimental data and exclusion of the gender field from study produced an inconclusive though
unbiased result.
17 | P a g e

5.0 Reference Lists
Alfieri, L., Brooks, P. J., Aldrich, N. J., &Tenenbaum, H. R. (2011). Does discovery-based
instruction enhance learning?. Journal of educational psychology, 103(1), 1.
Arciuli, J., & Simpson, I. C. (2011).Statistical learning in typically developing children: the role
of age and speed of stimulus presentation. Developmental science, 14(3), 464-473.
Bal, P. M., de Lange, A. H., Ybema, J. F., Jansen, P. G., & van der Velde, M. E. (2011). Age and
trust as moderators in the relation between procedural justice and turnover: A large‐scale
longitudinal study. Applied Psychology, 60(1), 66-86.
Biesta, G. (2009). Good education in an age of measurement: On the need to reconnect with the
question of purpose in education. Educational Assessment, Evaluation and Accountability
(formerly: Journal of Personnel Evaluation in Education), 21(1), 33-46.
Botwinick, J. (2013). Cognitive processes in maturity and old age. Springer.
Christman, S. D., & Butler, M. (2011). Mixed-handedness advantages in episodic memory
obtained under conditions of intentional learning extend to incidental learning. Brain and
Cognition, 77(1), 17-22.
Eysenck, M. W. (1974). Age differences in incidental learning. Developmental
Psychology, 10(6), 936.
Gegenfurtner, A., &Vauras, M. (2012).Age-related differences in the relation between
motivation to learn and transfer of training in adult continuing education. Contemporary
Educational Psychology, 37(1), 33-46.
Guo, G., Fu, Y., Dyer, C. R., & Huang, T. S. (2008). Image-based human age estimation by
manifold learning and locally adjusted robust regression. IEEE Transactions on Image
Processing, 17(7), 1178-1188.
Kausler, D. H. (2012). Experimental psychology, cognition, and human aging.Springer Science
& Business Media.
18 | P a g e

Kausler, D. H., & Lair, C. V. (1965). RS (“backward”) paired-associate learning in elderly
subjects. Journal of gerontology, 20(1), 29-31.
Kollöffel, B. (2012). Exploring the relation between visualizer–verbalizer cognitive styles and
performance with visual or verbal learning material. Computers & Education, 58(2), 697-
706.
Kormos, J., &Csizér, K. (2008).Age‐related differences in the motivation of learning English as
a foreign language: Attitudes, selves, and motivated learning behavior. Language
learning, 58(2), 327-355.
Laurence, M. W. (1967). Memory loss with age: A test of two strategies for its
retardation. Psychonomic Science, 9(4), 209-210.
Laurence, M. W. (1967). A developmental look at the usefulness of list categorization as an aid
to free recall. Canadian Journal of Psychology/Revue canadienne de psychologie, 21(2),
153.
Lee, M., & Fortune, A. E. (2013).Patterns of field learning activities and their relation to learning
outcome. Journal of Social Work Education, 49(3), 420-438.
Marulis, L. M., &Neuman, S. B. (2010). The effects of vocabulary intervention on young
children’s word learning: A meta-analysis. Review of educational research, 80(3), 300-
335.
Mather, M., &Schoeke, A. (2011). Positive outcomes enhance incidental learning for both
younger and older adults. Frontiers in neuroscience, 5, 129.
Melton, A. W. (Ed.).(2014). Categories of human learning.Academic Press.
Merriam, S. B., Caffarella, R. S., & Baumgartner, L. M. (2012). Learning in adulthood: A
comprehensive guide.John Wiley & Sons.
Nasir, M., &Masrur, R. (2010).An exploration of emotional intelligence of the students of IIUI in
relation to gender, age and academic achievement. Bulletin of education and
research, 32(1).
19 | P a g e

van der Stel, M., &Veenman, M. V. (2008). Relation between intellectual ability and
metacognitive skillfulness as predictors of learning performance of young students
performing tasks in different domains. Learning and Individual Differences, 18(1), 128-
134.
Vygotsky, L. S. (2011). The dynamics of the schoolchild's mental development in relation to
teaching and learning. Journal of cognitive education and psychology, 10(2), 198.
Webb, S., Newton, J., & Chang, A. (2013).Incidental learning of collocation. Language
Learning, 63(1), 91-120.
20 | P a g e

6.0 Appendix
Table 4: Descriptive statistics for the entire recall pattern
Descriptive Statistics
age Mean Std. Deviation N
Recall 1
1 6.50 1.434 10
2 7.00 1.826 10
Total 6.75 1.618 20
Recall 2
1 7.60 1.955 10
2 6.90 2.132 10
Total 7.25 2.023 20
Recall 3
1 14.80 3.490 10
2 11.00 2.494 10
Total 12.90 3.538 20
Recall 4
1 17.60 2.591 10
2 13.40 4.502 10
Total 15.50 4.174 20
Recall 5
1 19.30 2.669 10
2 12.00 3.742 10
Total 15.65 4.902 20
Table 5: Multivariate Test results for the sample
Multivariate Testsa
Effect Value F Hypothesis
df
Error
df
Sig. Partial Eta
Squared
Noncent.
Parameter
Observed
Powerc
recall_type
Pillai's
Trace .922 44.108b 4.000 15.000 .000 .922 176.432 1.000
Wilks'
Lambda .078 44.108b 4.000 15.000 .000 .922 176.432 1.000
Hotelling's
Trace 11.762 44.108b 4.000 15.000 .000 .922 176.432 1.000
Roy's
Largest
Root
11.762 44.108b 4.000 15.000 .000 .922 176.432 1.000
recall_type *
AgeGroup
Pillai's
Trace
.566 4.881b 4.000 15.000 .010 .566 19.523 .873
21 | P a g e

Wilks'
Lambda .434 4.881b 4.000 15.000 .010 .566 19.523 .873
Hotelling's
Trace 1.302 4.881b 4.000 15.000 .010 .566 19.523 .873
Roy's
Largest
Root
1.302 4.881b 4.000 15.000 .010 .566 19.523 .873
a. Design: Intercept + AgeGroup
Within Subjects Design: recall_type
b. Exact statistic
c. Computed using alpha = .05
Table 6: Within- subjects effects and analysis
Source Type III
Sum of
Squares
df Mean
Square
F Sig. Partial
Eta
Squared
Noncent.
Parameter
Observed
Powera
recall_type
Sphericity
Assumed 1514.940 4 378.735 47.576 .000 .726 190.306 1.000
Greenhouse-
Geisser 1514.940 3.189 474.985 47.576 .000 .726 151.743 1.000
Huynh-Feldt 1514.940 4.000 378.735 47.576 .000 .726 190.306 1.000
Lower-bound 1514.940 1.000 1514.940 47.576 .000 .726 47.576 1.000
recall_type *
AgeGroup
Sphericity
Assumed 190.300 4 47.575 5.976 .000 .249 23.905 .980
Greenhouse-
Geisser 190.300 3.189 59.665 5.976 .001 .249 19.061 .953
Huynh-Feldt 190.300 4.000 47.575 5.976 .000 .249 23.905 .980
Lower-bound 190.300 1.000 190.300 5.976 .025 .249 5.976 .638
Error(recall_type)
Sphericity
Assumed 573.160 72 7.961
Greenhouse-
Geisser 573.160 57.41
0 9.984
Huynh-Feldt 573.160 72.00
0 7.961
Lower-bound 573.160 18.00
0 31.842
a. Computed using alpha = .05
22 | P a g e

Table 7: Within- subjects contrasts analysis
Source recall_type Type III
Sum of
Squares
df Mean
Square
F Sig. Partial
Eta
Squared
Noncent.
Parameter
Observed
Powera
recall_type
Linear 1357.205 1 1357.205 155.712 .000 .896 155.712 1.000
Quadratic 20.089 1 20.089 2.549 .128 .124 2.549 .327
Cubic 115.520 1 115.520 28.816 .000 .616 28.816 .999
Order 4 22.126 1 22.126 1.969 .178 .099 1.969 .264
recall_type *
AgeGroup
Linear 182.405 1 182.405 20.927 .000 .538 20.927 .991
Quadratic .432 1 .432 .055 .817 .003 .055 .056
Cubic .320 1 .320 .080 .781 .004 .080 .058
Order 4 7.143 1 7.143 .636 .436 .034 .636 .118
Error(recall_type)
Linear 156.890 18 8.716
Quadratic 141.836 18 7.880
Cubic 72.160 18 4.009
Order 4 202.274 18 11.237
a. Computed using alpha = .05
Table 8: Test result between- Subjects effects
Tests of Between-Subjects Effects
Measure: MEASURE_1
Transformed Variable: Average
Source Type III Sum of
Squares
df Mean
Square
F Sig. Partial Eta
Squared
Noncent.
Parameter
Observed
Powera
Intercept 13479.210 1 13479.210 1626.83
2 .000 .989 1626.832 1.000
AgeGroup 240.250 1 240.250 28.996 .000 .617 28.996 .999
Error 149.140 18 8.286
a. Computed using alpha = .05
23 | P a g e

Table 9: Estimated marginal means
1. age * recall_type
Measure: MEASURE_1
age recall_type Mean Std. Error 95% Confidence Interval
Lower Bound Upper Bound
1
1 6.500 .519 5.409 7.591
2 7.600 .647 6.241 8.959
3 14.800 .959 12.785 16.815
4 17.600 1.161 15.160 20.040
5 19.300 1.028 17.141 21.459
2
1 7.000 .519 5.909 8.091
2 6.900 .647 5.541 8.259
3 11.000 .959 8.985 13.015
4 13.400 1.161 10.960 15.840
5 12.000 1.028 9.841 14.159
Table 10: Levene's Test of Equality of Error Variancesa
F df1 df2 Sig.
Recall 1 .482 1 18 .496
Recall 2 .001 1 18 .973
Recall 3 1.455 1 18 .243
Recall 4 2.460 1 18 .134
Recall 5 .383 1 18 .544
Tests the null hypothesis that the error variance of the
dependent variable is equal across groups.
a. Design: Intercept + AgeGroup
Within Subjects Design: recall_type
24 | P a g e

Table 11: Pair-wise Comparisons of recall types
(I) recall_type (J) recall_type Mean Difference
(I-J)
Std. Error Sig.b 95% Confidence Interval for
Differenceb
Lower Bound Upper Bound
1
2 -.500 .632 1.000 -2.520 1.520
3 -6.150* .840 .000 -8.836 -3.464
4 -8.750* 1.024 .000 -12.022 -5.478
5 -8.900* .843 .000 -11.593 -6.207
2
1 .500 .632 1.000 -1.520 2.520
3 -5.650* .793 .000 -8.186 -3.114
4 -8.250* .750 .000 -10.647 -5.853
5 -8.400* .893 .000 -11.254 -5.546
3
1 6.150* .840 .000 3.464 8.836
2 5.650* .793 .000 3.114 8.186
4 -2.600 1.128 .333 -6.206 1.006
5 -2.750 .932 .085 -5.728 .228
4
1 8.750* 1.024 .000 5.478 12.022
2 8.250* .750 .000 5.853 10.647
3 2.600 1.128 .333 -1.006 6.206
5 -.150 .984 1.000 -3.295 2.995
5
1 8.900* .843 .000 6.207 11.593
2 8.400* .893 .000 5.546 11.254
3 2.750 .932 .085 -.228 5.728
4 .150 .984 1.000 -2.995 3.295
Based on estimated marginal means
*. The mean difference is significant at the .05 level.
b. Adjustment for multiple comparisons: Bonferroni.
25 | P a g e

Table 12: Multivariate Tests results
Value F Hypothesis
df
Error df Sig. Partial Eta
Squared
Noncent.
Parameter
Observed
Powerb
Pillai's trace .922 44.108a 4.000 15.000 .000 .922 176.432 1.000
Wilks' lambda .078 44.108a 4.000 15.000 .000 .922 176.432 1.000
Hotelling's
trace 11.762 44.108a 4.000 15.000 .000 .922 176.432 1.000
Roy's largest
root 11.762 44.108a 4.000 15.000 .000 .922 176.432 1.000
Each F tests the multivariate effect of recall_type. These tests are based on the linearly independent pairwise
comparisons among the estimated marginal means.
a. Exact statistic
b. Computed using alpha = .05
26 | P a g e

Figure 4: Profile plot of estimated Marginal means of two age groups
Table 13: Normality test of the sample based on gender
Sex of subject Kolmogorov-Smirnova Shapiro-Wilk
Statistic df Sig. Statistic df Sig.
No. of times of recall
female .136 56 .011 .934 56 .004
male .135 44 .042 .949 44 .049
a. Lilliefors Significance Correction
Table 14: Normality test of the sample based on gender
Age group of subject Kolmogorov-Smirnova Shapiro-Wilk
Statistic df Sig. Statistic df Sig.
No. of times of recall young .136 50 .021 .924 50 .003
older .147 50 .008 .942 50 .016
a. Lilliefors Significance Correction
27 | P a g e

Power (1-β err prob)
Total sample size
F tests - ANOVA: Repeated measures, between factors
Number of groups = 5, Number of measurements = 4,
α err prob = 0.05, Effect size f(V) = 1.23083
11
12
13
14
15
16
17
18
0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95
Figure 5: Power curve for normality test
Table 15: Independent sample t-test for gender
Independent Samples Test
Levene's Test for
Equality of Variances
t-test for Equality of Means
F Sig. t df Sig. (2-
tailed)
Mean
Differenc
e
Std. Error
Differenc
e
95% Confidence
Interval of the
Difference
Lower Upper
Recal
l
Equal
variances
assumed
.580 .448 .574 98 .567 .602 1.049 -1.480 2.685
28 | P a g e

Equal
variances
not
assumed
.579 95.28
9 .564 .602 1.040 -1.462 2.666
Table 16: Recall with imagery test
Source Type III Sum of
Squares
df Mean Square F Sig. Partial Eta
Squared
Corrected Model 88.200a 1 88.200 6.539 .020 .266
Intercept 4805.000 1 4805.000 356.219 .000 .952
AgeGroup 88.200 1 88.200 6.539 .020 .266
Error 242.800 18 13.489
Total 5136.000 20
Corrected Total 331.000 19
a. R Squared = .266 (Adjusted R Squared = .226)
Table 17: Recall with counting test
Source Type III Sum of
Squares
df Mean Square F Sig. Partial Eta
Squared
Corrected Model 1.250a 1 1.250 .464 .504 .025
Intercept 911.250 1 911.250 338.196 .000 .949
AgeGroup 1.250 1 1.250 .464 .504 .025
Error 48.500 18 2.694
Total 961.000 20
Corrected Total 49.750 19
a. R Squared = .025 (Adjusted R Squared = -.029)
Table 18: Recall with adjective test
Source Type III Sum of
Squares
df Mean Square F Sig. Partial Eta
Squared
Corrected Model 72.200a 1 72.200 7.848 .012 .304
Intercept 3328.200 1 3328.200 361.761 .000 .953
AgeGroup 72.200 1 72.200 7.848 .012 .304
Error 165.600 18 9.200
Total 3566.000 20
Corrected Total 237.800 19
a. R Squared = .304 (Adjusted R Squared = .265)
29 | P a g e

Table 19: Recall with rhyming test
Source Type III Sum of
Squares
df Mean Square F Sig. Partial Eta
Squared
Corrected Model 2.450a 1 2.450 .586 .454 .032
Intercept 1051.250 1 1051.250 251.295 .000 .933
AgeGroup 2.450 1 2.450 .586 .454 .032
Error 75.300 18 4.183
Total 1129.000 20
Corrected Total 77.750 19
a. R Squared = .032 (Adjusted R Squared = -.022)
Table 20: Recall with control test
Source Type III Sum of
Squares
df Mean Square F Sig. Partial Eta
Squared
Corrected Model 266.450a 1 266.450 25.229 .000 .584
Intercept 4898.450 1 4898.450 463.820 .000 .963
AgeGroup 266.450 1 266.450 25.229 .000 .584
Error 190.100 18 10.561
Total 5355.000 20
Corrected Total 456.550 19
a. R Squared = .584 (Adjusted R Squared = .560)
30 | P a g e