CHAPTER II

METHODOLOGY

2.1 The Sampling Techniques................................................. 81

2.2 Sample of the Study........................................................... 82
2.3 Tools ................................................................................. 83
2.4 Procedure for Data Collection ......................................... 99
2.5 Statistical Techniques........................................................ 100
 Details of the methodology adopted in the study have been presented
below under various heads, viz., the sampling technique, the sample, the tools,
procedure for data collection and statistical techniques.

2.1 The sampling technique

A population is an aggregate of all the cases that conform to some

designated set of specifications whereas ‘Stratum’ may be defined by one or
more specifications that divide a population into mutually exclusive segments. A
single member of a population is referred to as an element. When some elements
are selected with the intention of finding something about the population from
which they are taken, that group of elements is referred to as a sample.

In order to collect a good sample, the procedure should focus on collecting

as much information about the population as can be obtained at a stipulated cost,
while at the same time avoiding bias. The error, which causes a statistic to differ
from its parameter, may arise partly from random error in the selection of
individuals and partly from bias in their selection. The total of the random errors
decrease if the sample size is increased but if errors are due to bias then it does not
do so.

In the present study, stratified random sampling was used. According to

Mendenhall, Ott and Scheaffer (1971), a stratified sample is obtained by
separating the population elements into non overlapping groups, called strata, and
then selecting a simple random sample from within each stratum. This technique
is generally applied in order to obtain a representative sample. Under stratified
sampling, the population is divided into several sub populations that are
individually more homogeneous than the total population.

The following steps can be used in stratified random sampling:

 Deciding upon the relevant stratification factors, for example, gender age,
course of studies, etc.
 Dividing the entire population into sub populations based on the
stratification criteria.

81
  Setting up the units separately in each sub population.
 Selecting the requisite number of units from each sub population by using an
appropriate random selection technique
 All the sub-samples thus selected make up the main sample (Aggarwal,
1988; P.52-53).

In the present study the strata’s were based on three criteria, viz., gender
(male, female), school type (public, private, primary schools), and the
achievement levels of the students (high and low achievers). The diagram below
illustrates it further.
SAMPLE

Males Females

Private Schools Public Schools Private Schools Public Schools

High Low High Low High Low High Low

Achievers Achievers Achievers Achievers Achievers Achievers Achievers Achievers

2.2 The Sample

The sample consisted of two hundred students from both public and
private primary schools in Nairobi city (101 boys and 99 girls). The sample was
selected on the basis of the following inclusion/ exclusion criteria.

2.2.1 Inclusion Criteria

(i) Being a pupil in standard eight class for two consecutive school terms and
having sat all the required examinations for the class. Also having been graded
and ranked in class as per the academic performance by the teachers.
(ii) Having been a learner in either public or private primary schools in Nairobi
province.

82
(iii) Individuals ranging between ages 12 to 14½ , as it is the acceptable age for
being in that level of schooling in the country.

2.2.2 The Exclusion Criteria

(i) Individual who had not been in standard eight for two consecutive school
terms, and hence didn’t have marks for two terms and had also not been graded
and ranked in that class.
(ii) Individuals who were more than 14½ years of age.

2.2.3 Sample Details

The details of the sample selected on basis of the above exclusion/inclusion

criteria are presented in table 1.
Table 1. Split - up of the sample on the basis of gender, type of school, and
level academic achievement
Gender Males (N= 101) Females (n= 49) Total
Type of Private Public Private Public
school (N= 47) (N=54) (N= 53) (N=46)
Level of Low High Low High Low High Low High
Academic Achi. Ach. Achi. Ach. Achi. Ach. Achi. Ach.
Achievement
N 24 23 27 27 27 26 23 23 200

2.3 The Tools

One major problem encountered in the planning stage of the study was
selection of tools best suited for the purpose of the study. Review of literature
provided information regarding many scales that could be used to measure some
of the variables in the present study.

It was felt that the scales standardized in other cultures may not yield
valid results when used with Kenyan subjects. For a test to be reliable and valid,
the items must be suited to the contexts familiar to the subjects who respond to
them. The non-availability of adequate and valid tools to measure the variables
included in the study prompted the investigator to go in for the construction of
tests whenever it was found necessary.

83
The following are different measuring devices used in the present study.
1. Socio economic status inventory.
2. Academic Performance Blank.
3. Achievement Motivation Inventory.
4. Perceived classroom climate scale.
5. Home environment scale.
6. Standard Progressive Matrices

As noted earlier, a salient feature of the present research work is that

several tools had to be newly constructed to suit the sample under study. In fact,
all the tools, with the exception of the standard progressive matrices have been
newly constructed in connection with the present study. The test development
procedures followed Likert approach to undimensional scaling, which consisted of
three interrelated tasks viz., item construction, item scoring, and item selection
(Mclver & Carmines, 1994).

The details of the procedure involved in the development of each of the

tools are presented below with a view to help assess their adequacy and validity.

2.3.1 Socio-economic status Inventory (Muthee, 2009).

Socio-economic status of the subjects was measured using an

information schedule, which also yielded some additional information regarding
the respondent, like gender, age, family structure, etc.

Socio-economic status is an elusive construct. The indicators of socio-

economic status vary widely in different societies. This makes it difficult to
develop a standardized tool for measuring socio-economic status. Taking into
consideration the complexities involved, the present researcher decided to follow
a systematic procedure for developing a measure of socio-economic status
suitable for use with a Kenyan sample. Details of the procedure are presented
below.

A review of the related literature on socio-economic status revealed that

three components are generally considered as important indicators of socio-

84
economic status. They are educational status, occupational status, and economic
status. In addition, certain other indicators like the physical facilities available,
neighborhood, type of dwelling, etc. are also felt to be important in the case of the
Kenyan subjects under study. In accordance with these considerations, questions
were framed to tap information on the relevant aspects of socio-economic status.
These questions related to: (i) Educational level of the parent, (2) Occupation of
parents, and (3) Economic status of the parent.

Parental education was coded on a 7 point rating scale ranging from 1

(never attended school) to 7 (studied up to Ph.D level). Parental occupation,
which was marked by respondents using a long check list of available
occupations, was coded on a three point scale, viz., 1 (lowly paid job), 2 (medium
paid jobs) and 3 (highly paid jobs). Income of the household was reported on a 8
point scale, that ranged from 1 (below Ksh 5,000 per month) to 8 (Ksh 150,000
per month). Housing of the family was reported on a four point scale, viz., 1
(temporary Mabati house), 2. (semi permanent house), 3. (ordinary stone building)
and 4.(Bungalow/massonette). Possessions in the household was reported by the
respondents using an exhaustive check list, which contained items commonly
found in households of low, medium and high socio-economic status. Total
number of articles in each of the categories were weighted on a 3 point scale, viz.,
1. (low socio-economic status, 2.(middle socio-economic status) and 3.(high
socio-economic status). Employees at home was coded on a two point scale, viz.,
1(No. of employees at home ) and 2 (Yes, there are employees at home).
Frequency of purchasing a newspaper/magazine at home was coded on a four
point scale viz., 1(paper never bought at home) and 4 (papers bought on daily
basis). Average expenditure on newspapers and magazines per month in the
family was coded on a four point scale ranging from 1 (Ksh, 1500) to 4 (above
Ksh 2,500 per month). The opinion of other people about the family status was
coded on a 3 point scale viz., 1 (low class family), 2 (middle class family), and 3
(prosperous family). A draft scale with 13 items to collect information on the
various aspects of socio- economic status, as mentioned above, was constructed
and administered on to a selected sample.

85
TRY OUT: Information required in the general socio-economic status
inventory was given by students. The draft scale was tried on 40 students (20
boys and 20 girls) from standard eight class in Nairobi city. The boys and girls
were not part of the study sample. Their schools were selected by the researcher
on the basis of proximity to her residential area. Each school was personally
visited and with the help of the school head teacher and class teacher, the tests
were administered to the subjects identified. Necessary clarifications were given
by the researcher whenever the need arose. The opinions of the class teacher and
the respondents about the tools were sought. Based on the feedback given,
necessary modifications were effected in the items.

Scoring: The items which constituted the socio-economic status scale were first
weighted on a scale of 1-10 according to their importance in indicating the socio-
economic status of the respondent’s families. Each of the 13 items in the socio-
economic status scale was then weighted on a 10 point scale in accordance with
the importance of that particular item in determining the socio-economic status.
The weights given to the items were as follows: Item numbers 1, 2, 3, 4, 5 ,
6, 7, 9 and 12 were given the weight of 10 each. Item numbers 8, 11 and 12
were given a weight of 7 each and item number 13 was given a weight of 5. The
weighted scores were summed to get the total raw score. The raw scores were
converted to Z-score and then T-scores.
Reliability and Validity: The reliability of the socio-economic status scale
was established using Cronbach’s alpha which was found to be .903. The content
validity was established by obtaining opinions of experts in the field of education,
psychology and sociology on the items. The systematic method followed by the
investigator in the development and standardization of the test also helped to
ensure it’s validity. It may be re called that items for the draft scale were
meticulously prepared on the basis of an exhaustive review of related literature
which included perusal of several available measurement devices. A copy of the
final inventory is given in appendix A.
2.3.2 Academic Performance Blank (Muthee, 2009)

The academic achievement form is an information blank intended to

collect data regarding the academic performance of the students. The information

86
blanks contains columns where the total marks obtained by the students and
maximum marks on five compulsory subjects are to be entered. This information
is collected by the investigator herself from the school records and entered into the
relevant column in the blank. The percentage of marks obtained by the pupils on
each of the five subjects and the overall percentage on all the five subjects are
calculated from this information. The overall percentage provided a measure of
academic achievement of the pupils. A copy of the final form is given in
appendix B.

2.3.3 Achievement Motivation Inventory (Muthee & Thomas, 2009)

The scale is intended to assess the achievement motivation among

learners. The different steps followed in the conceptualization, item generation,
and standardization of the tool are presented below under suitable titles:

(a) Conceptualisation:
Achievement motivation is a broad concept. It is also a personality
variable that has been used to explain individual differences in a number of
contexts, including school, sports and the world of work. It is a multidimensional
concept that necessitates conceptual clarification and operational definitions
before developing a measuring instrument. Accordingly, a systematic procedure
was followed by the researcher for the development of achievement motivation
inventory. This included an exhaustive review of theoretical and empirical
literature on the subject, a review of available tests on achievement motivation,
consultation with psychologists, etc. These steps helped in identifying the
important dimensions of achievement motivation as presented below: (i)
Motivation for achievement (evidenced by competitiveness and goal orientation);
(ii) inner resources (evidenced by relaxed style, happiness, patience and self
confidence); (iii) inter personal strengths (evidenced by assertiveness, personal
diplomacy, extraversion and co-cooperativeness); and (iv) work habits (evidenced
by planning and organization, initiatives, and team spirit).

(b) Item generation;

After identifying the various dimensions involved in achievement
motivation, the next step in test construction was generation of items representing

87
these dimensions. The items were prepared in statement form taking care to
include both positively and negatively worded items in the scale. The respondent
were asked to indicate their level of agreement with each statement using a five
category response format which ranged from completely agree (5) to completely
disagree (1). This provided a more refined self description than was possible with
a two category (true/ false) format.

In order to ensure content validity of the tool, a large number of items

representing all the major domains of achievement motivation were first prepared.
This resulted in generation of about 80 items pertaining to various achievement
motivation domains. This was followed by a thorough scrutiny of the list aimed
at picking out those items that appeared to have good discriminative power and to
eliminate repetitious, overlapping, and ambiguous items. This resulted in a
shortened list with 50 items. The shortened list was again subjected to filtering
process to enlist only the best items from the original pool. Thus at the end of the
screening process, the draft scale consisted of 40 items with 24 positively worded
items and 16 negatively worded items. A copy of the draft Achievement
Motivation scale is presented as appendix (C).

Try Out: The draft tool was tried out on 40 students from standard eight classes.
They were selected from schools which were not part of the study sample. Items
that caused comprehension problems were modified to suit the language level of
the sample. Grammatical errors on a few items were also corrected.
(d) Item analysis and Selection:
To ensure the inclusion of only the best items in the final scale, item
analysis using likert method of summated ratings (Edwards, 1957) was conducted.
The purpose of item analysis was to identify those items that could ensure an
internally consistent scale and to eliminate those that do not. Accordingly, data
obtained from a sample of 200 subjects were subjected to item analysis using
Likert Method (Edwards, 1957). The response sheets were arranged in The order
of total scores so as to select the top and bottom 25% of the sample who
represented the high and low groups respectively. The power of each item to
discriminate between two groups was found using t-tests. The items that
produced statistically significant t-values were those that could discriminate

88
between two groups. The t-values obtained for the different items in the scale are
presented in table 2 below.

Table 2. The t-values relating to items in the draft achievement motivation

inventory.

Rank Serial No. in the draft scale t- value

1 20 6.955
2 31 6.916
3 37 5.943
4 10 5.814
5 21 5.803
6 26 5.464
7 40 5.431
8 39 5.399
9 29 5.161
10 38 5.154
11. 34 5.066
12. 15 4.933
13. 1 4.910
14. 19 4.854
15. 32 4.795
16. 22 4.662
17. 30 4.381
18. 24 4.381
19. 4 4.093
20. 6 3.944
21. 25 3.859
22. 18 3.800
23. 8 3.496
24. 36 3.495
25. 27 3.463
26. 7 3.260
27. 12 3.153
28. 2 3.135
29. 3 3.109
30. 23 2.880
31. 16 2.769
32. 5 2.615
33. 35 1.984
34. 11 1.951
35. 33 1.337
36. 14 1.106
37. 28 1.050
38. 9 .679
39. 13 .516
40. 17 .415

89
 From table 2, it may be seen that the first 32 items with t-values ranging
from 6.955 to 2.615 produced statistically significant t- values. All these items
were retained in the final scale. The remaining eight items were dropped from the
scale as they could not discriminate between those with low and high levels of
achievement motivation. Among the 32 items retained in the final scale, 18 were
positively worded and 14 were negatively worded. Items with positive and
negative wording were arranged randomly in the final scale. The serial numbers
of the items with positive wording are the following: 3, 4, 5, 6, 11, 13, 14, 16,
17, 20, 23, 24, 26, 28, 29, 30, 31, 32. The serial number of items with negative
wordings are: 1, 2, 7, 8, 9, 10, 12, 15, 18, 19, 21, 22, 25 and 27.
(e) Scoring: As mentioned earlier, the responses to the items were marked using
a five point Likert format. The points are, completely agree, mostly agree, agree
to some extent, mostly disagree and completely disagree. The scoring weights
given to these responses were 5, 4, 3, 2 and 1 respectively for positively worded
items and 1,2,3,4 and 5 respectively for a negatively worded item. This scoring
scheme ensured that higher scores indicate higher levels of achievement
motivation, and lower scores indicate low levels of achievement motivation.
(f) Reliability and Validity: The reliability of the scale has been computed using
Cronbach’s alpha, which was found to be 0.749. At the same time, validity for
the scale is claimed on the basis of the systematic methodology followed by the
researcher during development and standardization of the test. It may be recalled
that items for the draft scale were meticulously prepared on basis of an exhaustive
review of related literature, which included perusal of all available measurement
devices. Moreover, the items were selected on the basis of item analysis which
ensured discriminatory power of the items.
A copy of the final inventory having 32 items is given in Appendix D.

2.3.4 Classroom Climate Scale (Muthee, 2009)

The scale is intended to measure student’s perception of psychological

processes going on in the classroom. The different steps followed in the
conceptualization, item generation and standardization of the tools are presented
below under suitable titles.

90
(a) Conceptualisation: The classroom plays a very significant role in
shaping the attitudes of students towards academics and providing the necessary
academic atmosphere. The researcher realized that in order to develop a valid tool
for measuring class room climate, it is necessary to identify the various aspects of
the climate which have an impact on student’s academic activities. Accordingly,
an exhaustive review of theoretical and empirical literature on the subject was
conducted along with a review of the available classroom climate tests. Experts in
the field of educational psychology were also consulted for the purpose. These
steps helped in achieving a comprehensive understanding of the concept of a
classroom climate. The different aspects of class-climate, which the researcher
could identify as having some bearing on the academic life of students, are the
followings: (1) The physical environment that is welcoming and conducive to
learning, (2) social environment that promotes communication and interaction; (3)
An effective environment that promotes sense of belonging and a self esteem; (4)
An academic environment that promotes learning and self fulfillment.

(b) Item generation: After identification of the various dimensions involved

in the classroom climate, the next step in test construction was the generation of
items representing these dimensions. It was decided to prepare items in statement
form with four category response format. Care was taken to include both
positively and negatively worded items in the scale. A few items from the
available inventories found suitable for the present population were also adapted
in the preliminary item pool. In general, the items required the respondents to
indicate how often certain things occur in their classroom, by selecting the
appropriate choices. A total of 36 items were generated, out of which 19 were
positively worded and 17 were negatively worded. The four response anchors
provided along with the statements were ‘always’, ‘frequently’, ‘sometimes’,
and ‘never’.

(c) Tryout: The participants of the pilot study were 40 subjects in the range of
12-14½ years, who were studying in standard eight in two city primary schools.
These schools were selected by the investigator on basis of their proximity to her
residential area. Each school was personally visited and with the help of the
school head teacher and the class teacher, questionnaires were administered to the

91
subjects. Necessary clarification was given by the researcher as need arose. The
opinions of the class teacher and the participants about the items in the tool were
sought in the items by the investigator. Based on this feed back, necessary
modifications were made in the items. For example, the items that caused
problems in comprehension were reworded or totally excluded from the draft
scale. After try out the draft scale contained 40 items.

(d) Item selection: To ensure selection of only the best items in final scale,
item analysis using Likert’s method of summated ratings (Edward, 1957) was
conducted. The purpose of the item analysis was to identify those items that
could discriminate between high scores and low scores in the scale and to
eliminate those that do not. For this purpose, the score sheets were rank ordered in
terms of total scores in the draft scale. From this ranked list, the upper and lower
25% of cases were selected to represent the high and the low groups in class room
climate.

In the next step, the mean scores obtained by the high and the low group,
were compared using t-tests and the items were ranked on the basis of the t-values
obtained. The ranked t-values, along with the serial numbers of the items in the
draft scale are presented in table 3. It may be noted here that these t-values give a
measure of the extent to which a given item differentiated between the high and
the low groups.

Table 3: The t-values relating to the items in the draft Class room climate
scale.

Rank Serial No. in the draft scale t- value

1 24 9.084
2 23 7.874
3 28 7.804
4 34 7.387
5 19 7.315
6 35 7.026
7 20 6.266
8 18 6.231
9 27 6.008
10. 21 6.003
11. 22 5.728
12. 39 5.632

92
 13. 25 2.276
14. 26 5.558
15. 17 5.386
16. 15 5.346
17. 37 5.295
18. 29 5.116
19. 13 5.033
20. 11 4.748
21. 40 4.734
22. 9 4.730
23. 30 4.483
24. 10 4.409
25. 8 4.365
26. 31 4.130
27. 16 4.110
28. 6 4.058
29. 5 4.008
30. 38 3.791
31. 4 3.393
32. 12 2.975
33. 3 2.848
34. 2 2.820
35. 32 2.626
36. 14 2.620
37. 1 1.371
38. 7 1.218
39. 33 1.200
40. 36 .900
Table 3 shows that among the 40 items in the draft scale, 36 items
produced significant t-values. All the significant items were retained in the final
scale and the remaining four were dropped as they failed to discriminate between
high and low scorers in the scale. Among the 36 items which got selected in the
final scale, 19 were positively worded and 17, negatively worded. The positive
and negative items were mixed randomly and printed along with instructions in
the final scale. The serial numbers of positively worded items in the final scale
are 1, 4, 5, 6, 8, 9, 10, 12, 15, 16, 17, 18, 20, 25, 28, 31, 32, 34, 36.
Similarly, items with serial numbers 2, 3, 7, 11, 13, 14, 19, 21, 22, 23, 24,
26, 27, 28, 30, 33 and 35 were negatively worded . A copy of the final version
of the inventory is given in Appendix E, and the final version in Appendix F.

93
(e) Scoring : The scoring weights given for the four response choices mentioned
earlier for an item were 4 for 'Always', 3 for 'Frequently', 2 for 'Sometimes', and 1
for 'Never' in the case of positively worded items. The scoring weights were in
reverse order for negatively worded items.

This scoring scheme ensured that the higher the score, the better the
perceived classroom climate.

(f) Reliability and Validity: Cronbach’s alpha computed for the scale is found to
be 0.826, indicating that it has good reliability. The investigator claims validity
for the scale on the basis of the fact that a systematic methodology was followed
by her for development and standardization of the test. It may be recalled that
items for the draft scale were meticulously prepared on basis of an exhaustive
review of related literature which included perusal of all the available
measurement devices. Moreover, the items were finally selected on the basis of
item analysis which ensured ability of the items to discriminate higher scorers
from lower scorers.

2.3.5 Home Environment scale (Muthee, 2009)

The scale is intended to assess several aspects of parental involvement in

the child’s school work at home. The different steps followed in the
conceptualization, item generation and standardization of the tool are presented
below under suitable titles:

(a) Conceptualization: Home environment is a complex concept related to

the role behavior of the members of the family, customs and traditions in which
the family unit is embedded, educational and job status of the members, etc.
Home environment is reported to play an important role in determining the mood
state of family members and thereby influencing almost all of their behavior
patterns. It is a challenging task to develop a tool to assess such a complex and
multidimensional concept. In the context of the present study, academic
performance of children in the family assumed a greater significance and hence
emphasis was placed on those aspects of home environment which had some
bearing on children’s studies.

94
 As is customary in the development of tools for measuring complex
concepts, a systematic procedure was followed for development of the home
environment scale. This included an exhaustive review of relevant literature,
perusal of available tools, consultation with the experts, etc. This resulted in the
identification of the following four dimensions: (i) Values (goals), (2) Parental
involvement, (3) School Communication and (4) School involvement.

(b) Item generation: After identification of the various dimensions involved in

home environment, the next step in test construction was the generation of items
representing these dimensions. The items were prepared in statement form with
three category response format, viz., Always, Sometimes, and Not at all. Care
was taken to include both positively and negatively worded items in the scale.
Respondents were asked to indicate how often their parents/guardians engage in
certain kinds of behaviour by selecting the appropriate responses. In order to
ensure content validity of the tool, a large number of items of parental
involvement in the child’s academic work were first prepared. This resulted in the
generation of about 100 items pertaining to various parental involvement
dimensions. This was followed by a thorough scrutiny of the list aimed at
eliminating those items that appeared to be repetitions, overlapping and
ambiguous. This resulted in a shortened list with 60 items. The shortened list
was again subjected to filtering process to enlist only the best items from the
original pool. Thus, at the end of the screening process, the draft scale consisted
of 35 items out of which 27 were positively worded and 8 were negatively
worded.

(c) Try Out: The draft test was tried out on a sample of 40 standard eight
students (20 girls and 20 boys) from Nairobi province in Kenya. The age range of
the respondents ranged from 12 -14½ years. Feedback obtained from this sample
was helpful in modifying the wording of some of the items to suit the language
level of the respondents.

(d) Item selection: To ensure the inclusion of only the best items in the final
scale, item analysis using Likert’s method of summated ratings (Edwards, 1957)

95
was conducted. The purpose of item analysis was to identify those items that
form an internally consistent scale and eliminate those that do not. Accordingly,
the response sheets obtained from the sample of 200 subjects were arranged in the
order of total scores so as to select the top and bottom 25% of the sample who
represented the high and low groups respectively. The power of each item to
discriminate between the two groups was assessed using t-tests. The results
obtained are printed in table 4. It may be noted that the items that produce
statistically significant t-values are those that can discriminate between the two
groups. Table 4 shows that out of the total of 35 items in the draft scale, 29
produced significant t-values. The remaining 6 items were dropped from the scale
as they failed to discriminate between high and low scorers in the home
environment scale. A copy of the draft form of the scale is presented in Appendix
G and the final form in Appendix H.

Table 4. The t-values of items in the draft Home Environment Scale.

Rank Serial No. in the draft scale t- value

1 15 8.668
2 35 7.723
3 26 6.528
4 32 5.893
5 9 5.821
6 22 5.735
7 3 5.284
8 8 5.161
9 1 5.051
10. 29 4.950
11. 27 4.850
12. 20 4.812
13. 18 4.755
14. 25 4.630
15. 19 4.577
16. 21 3.893
17. 31 3.780
18. 7 3.710
19. 16 3.612
20. 34 3.464
21. 5 3.396
22. 24 3.281
23. 13 3.173

96
 24. 10 3.060
25. 28 2.786
26. 17 2.780
27. 33 2.676
28. 4 2.613
29. 11 2.349
30. 6 1.559
31. 23 0.892
32. 30 0.646
33. 2 .589
34. 14 -.352
35. 12 -.109

(e) Scoring: Home environment scale is a 3 point Likert type scale. The three
points are, Always, Sometimes, and not at all. The weights given for these
response categories are 3, 2, and 1 respectively for positively worded items and 1,
2, and 3 respectively for negatively worded items. Serial numbers for the items
which are positively worded are, 1, 2, 3, 4, 5, 6, 7, 8, 10,11, 12, 13, 14, 15,
16, 17, 18, 19, 21, 22, 25, 26 and 28. The negatively worded items are 20,
23, 24, 27 and 29.

(f) Reliability and Validity: Cronbach’s alpha computed for the scale is 0.715
indicating satisfactory reliability. With regard to validity, it may be noted that the
systematic procedure followed for development of the scale help ensure the
content validity and face validity. Consultation of experts during the critical phase
of test development, item analysis of the tool using a representation sample, etc.,
also help to ensure the validity of the tool.

2.3.6 Standard Progressive Matrices (SPM)

The progressive matrices, originally developed by Ravens in 1938, are a

culture free test of ability for observation and clear thinking. It measures a
person’s ability to form perceptual relations and to reason by analogy. The test
can be administered to any individuals in the age range of 6 years and above,
irrespective of language and schooling. The total score indicates an individual’s
intellectual capacity regardless of his level of education. The standard progressive
matrices (SPM) is the first and the most widely used of the three instruments,
commonly known as the Ravens progressive matrices, the other two being the

97
coloured progressive matrices (CPM) and the advanced progressive matrices
(APM). All the three tests are measures of spearman’s g.

(a) Administration: The subjects are presented with booklets containing the
problems along with answer sheets where spaces are provided against each
question number. To get the subject familiarized with the procedure, an example
is given where examiner give as much help needed as possible. The problems are
pictorial by nature and solving them involves finding a missing Link in a given
pattern. The problems demand thinking by analogy. The subjects are reminded to
answer in the appropriate spaces in the answer sheet and to find answers to as
many items as possible. No time limit was imposed for completing the test.

(b) Scoring:

The SPM consist of 60 items arranged in five sets (A,B,C,D, and E) of 12

items each. Each item contains a figure with a missing piece. Below the figure
are either six (sets A and B) or eight (sets C, D and E) alternative pieces to
complete the figure, only one of which being correct. Each set involves a
different principle or ‘theme’ for obtaining the missing piece and within a set, the
items are roughly arranged in an increasing order of difficulty. An individual’s
score is the total number of correct solutions. The correct answers are given in the
manual of the test. The raw score is typically converted to a percentile rank by
using appropriate norms. The higher the total score, the better the individual is in
problem solving.

(c) Reliability and Validity: The split half reliability of the test is reported to
range from .60 to .98, with a median of .90. Test re-test correlations are found to
range from a low of .46 for an eleven year interval to a high of .97 for a two day
interval.

Raven provided test re-test coefficients for several age groups: 0.88 (+13+
years); 0.93 (under 30 years); 0.88 (30-39 years); 0.87 (40-49 years); 0.83 (50
years and above)

Spearman considered the SPM one of the best measures of g. Several

studies conducted in western cultures, which have factor analyzed the SPM along

98
with other cognitive measures have reported that SPM has a loading higher than
.75 on a general factor. Concurrent validity coefficients between the SPM and
Stanford Binet and Wescheler Scales are reported to range between .54 and .88.

(d) Norms: Norm groups included in the manual are British children, between
ages 6 and 16 and Irish children in the age range of 6 to 12 years. Norms based on
military and Civilian subjects between the ages of 20 and 65 are also available. A
supplement includes norms from Canada, United States and Germany.

Suggested uses: Recommended uses include measurement of person’s ability to

form perceptual relations and reason by analogy in research settings.

2.4 Procedure for Data collection

Prior to the final study, a pilot study was conducted as a try out of the tools
and also to get a feel of the situation.

2.4.1 Pilot Study

The participants of the pilot study were 40 subjects (20 girls and 20 boys)
belonging to standard eight studying in the city primary schools in Nairobi,
Kenya. The selection of the schools were based on convenience sampling, the
primary consideration being proximity of the schools to the investigator’s
residential area. The draft tools were administered to the participants with the
help of the head teachers.

Necessary clarifications were given by the investigator whenever the need

arose for the same. Modifications in the wording of the items and the instructions
were made on basis of the feedback obtained from the pilot study.

2.4.2 Final study

Before visiting the schools for final data collection, a research permit was
obtained from the headquarters of the ministry of education at Nairobi, Kenya.
(Vide copy of the permit letter in Appendix I). The investigator made prior visit
to the schools to make the necessary arrangement with the school administrators

99
on how to collect data from the respondents in their schools. Once in the schools,
the investigator took special care to give adequate information regarding the
purpose of the study to school head teachers.

The researcher also met the teachers of standard eight at the staffrooms and
briefed them about the study.

The investigator left schools for future visit after making arrangements with
the head teacher and deciding on the dates of visits. The time for visit varied from
school to school as determined by school time table.

The investigator visited the schools as per pre-fixed schedule and met the
respondents in a separate room arranged by the head teacher and the concerned
class teacher. The respondents were then given a brief introduction by the
researcher which included a narration of the nature of the study and also an
assurance of confidentiality of information collected. All the test materials were
administered in a single session.

2.5 Statistical Techniques

The data collected were consolidated, coded, scored and entered into an
electronic spread sheet for statistical analysis, using software (SPSS). The
different statistical techniques employed in the present study were the following:
1. Correlation
2. t-tests
3. ANOVA and post hoc test (Duncan’s test)
4. Multiple regressions
5. Path analysis.
2.5.1 Correlation

A correlation is a concomitant variation between two variables in such away

that change in one is associated with change in the other. Correlation coefficient
is a numerical index of the degree of relationship between two variables.
Correlation coefficient range between -1 to +1. The numerical sizes of the
correlation coefficient is an expression of the strength of the relationship and sign

100
of the correlation coefficient point to the direction of the relationship. Positive
correlation indicates that a high standing on one variable is associated with high
standing on the other variable. Negative correlation indicates that high standing
on one variable is associated with low standing on the other variable.

The most common technique for computing of coefficient of correlation is

the Product Moment Method and it is used in the present study to determine the
nature and extent of the relationship of academic achievement, achievement
motivation and intelligence with socio-demographic and other familial variables.

2.5.2 t-test

The t-test is a parametric test statistic used to find out whether the
differences which occur between means of two data samples are statistically
significant. It is actually the ratio of the differences between two mean values to
it’s standard error. The statistical significance of t- depends upon it’s size and the
number of degrees of freedom.

When used in connection with determining the significance of differences

between two means, t is stated in terms of probability or P-value, which gives the
probability that differences just as large as obtained could occur by chance. By
convention, the P value should be .05 or lower for considering the mean
differences as significant. It may also be noted that as the sample size increases t
distribution approaches the normal distribution.

In the present study, the t-test was used to compute significance of

difference between mean scores obtained by the different sub groups (e.g.,
boys/girls, private/public schools) on different variables of the study.

2.5.3 ANOVA (Analysis of Variance)

Anova is employed to find out if significant differences exist among three or

more mean values. The basic principle of anova is to test differences among the
means of the sub groups, by examining the amount of variation between each of
these sub groups, relative to the amount of variation within of these sub groups.

101
Two estimates of population variance, one based on between groups variance, and
the other one based on within group’s variance, are made and then these two
estimates are used to compute the value F using a formula:

F= Estimate of population variance based on between group variance/Estimate of

population variance based on within group variance.

The F-value obtained is compared to the F limit for given degree freedom.
If the obtained F- value is equal to or exceeds the F-limit value, then it may be
assumed that there are significant differences among the sample means. One way
anova is done when the sub groups are formed on the basis of one factor (e.g.,
religion). Two way anova or multiple way anova is done when the sub groups
results from two or more factors (e.g., religion and gender). In such cases, the
anova gives an estimate of the main effects (the unique effects of each of the
independent variables, controlling for the effects of other variables) as well as the
effects due to interaction between independent variables.

In the present study, both one way and two way anova were used. One
way anova was employed to compare differences among the three sub groups of
socio-economic status (high, middle and low). Two way anova was employed to
measure the main effect due to different classificatory variables (e.g., gender,
socio-economic status, type of school, level of academic achievement, etc.) taken
two at a time, and the effects due to their interaction.

2.5.4 Post hoc test (Duncan's method)

Post hoc tests, also called multiple comparison tests, are used to determine
the significant differences between group means taken in pairs, as a follow up of
one way Anova. For example, when a significant F-value has been obtained in
one way anova, the investigator often wish to undertake further tests to determine
which particular group means differ and which do not differ from the other means.
A number of procedures, generally known as multiple comparisons techniques,
can be employed for this purpose. Such procedures aim to retain the significance
level at the required value when undertaking multiple tests. In other words, such

102
tests protect from claiming too many significant results (i.e. too many false
positives).
2.5.4 Multiple Regression

In multiple regression, a linear composite of explanatory variables is

formed in such a way that it has maximum correlation with criterion variable. The
technique is appropriate when there is a single metric criterion variable, which is
supposed to be a function of other explanatory variables. The main objective in
using this technique is to predict the variability of the dependent variable based on
its covariance with all the independent variables. Using the regression analysis
model, the scores on the dependent variable can be predicted based on the scores
on the independent variables. The analysis results in a regression equation having
the general formulai Y (the dependent variable)= B1 X1 +B2X2 +………+Bk Xk +
A, where B1, B2,…. Bk, are the regression coefficients (beta weights), X1, X2,
……, Xk , are the predictor variables, and A is a constant representing the value of
the intercept in Y the axis.

In the present study, multiple regression analysis was conducted using

achievement motivation and academic performance (marks) as dependent
variables and the other study variables as independents, with a view to find out the
extent to which achievement motivation and academic performance could be
predicted on basis of the entire set of variables considered in the present study. It
is also noteworthy here that multiple regression procedure provides opportunity
to control for multivariate interactions among the entire set of variables and
estimate the unique contributions of different predictors retained in the final
regression equation.

2.5.5. Path analysis

Path analysis is an extension of the regression model used to test fit of

correlational matrix against two or more causal models which are being compared
by the researcher. The model is usually depicted in a circle and arrow figure in
which single headed arrows indicate causal relationships. A regression is done for
each dependent variable in the model on the basis of independent variables

103
depicted. The regression weights produced by the model are compared with the
observed correlational matrix for the variables, and a goodness of fit statistic is
calculated. The best fitting of two or more models is selected by the researcher as
the best model for advancement of theory.

Path analysis requires the usual assumptions of regression. It is

particularly sensitive to model specifications because failure to include relevant
casual variables or inclusion of extraneous variables often substantially affects the
path coefficients, which are used to assess the relative importance of various
direct and indirect causal paths to the dependent variables. Interpretation of path
models should be undertaken in the context of competing alternative models, after
assessing their goodness of fit. When the variables in the model are latent
variables measured by multiple observed indicators, path analysis is termed as
structural equation modeling (SEM). Structural equation models are extremely
useful for examining the relationships among a set of variables. Specifically,
these models enable us to trace out the direct effects of one variable on another as
well as its indirect effects transmitted through other variables. Path analysis
simplifies this process by requiring that all of the variables be expressed in
standard form.
Path model: A path model is a diagram relating independent, intermediary, and
dependent variables. Single headed arrows indicate causation between exogenous
or intermediary variables and the dependent (s). Arrows also connect the error
terms with their respective endogenous variables. Double headed arrows (which
are drawn as curved arrows) indicate correlation between pairs of exogenous
variables. Sometimes, the arrows in the path model are drawn in a width which is
proportional to the absolute magnitude of the corresponding path coefficients.

Causal Paths to a given variable include: (1) the direct paths from arrows leading
to it, and (2) correlated paths from endogenous variables correlated with others
which have arrows leading to the given variable.

Exogenous and endogenous variables: Exogenous variables in a path model are

those with no explicit causes (no arrows going to them, other than the
measurement error term). If exogenous variables are correlated, this is indicated

104
by a double headed arrow connecting them. Endogenous variables, on the other
hand, are those which do have incoming arrows. Endogenous variables include
intervening causal variables and dependents. Intervening endogenous variables
have both incoming and out going causal arrows in the path diagram. The
dependent variables(s) have only incoming arrows.

Path coefficients /path weights : Path coefficient is a standardized regression

coefficient (beta weight) showing the direct effect of an independent variable on
dependent variable in the path model. Thus, when the model has two or more
causal variables, path coefficients are partial regression coefficients which
measure the extent of the effect of one variable on another in the path model,
controlling for other prior variables using standardized data of correlation matrix
as input.

Correlated exogenous variables: If exogenous variables are correlated, it is

common to label the corresponding double headed arrow between them with its
correlation coefficient.

Disturbance terms: The residual error terms, also called disturbance terms,
reflect unexplained variance (the effect of unmeasured variables) plus
measurement error. The dependent in each equation is considered as an
endogenous variable. It may also be noted that independents in each equation are
all those variables with arrows leading to the dependent.

In the present study, path analysis was employed to identify the most
important variables in the study which may serve as either dependent or
independent and also the nature and extent of the relationships among those
variables identified. It is hoped that theoretically explicable path models obtained
with the sample under study may be helpful in determining the predictors of
academic performance and motivation.

105