Sei sulla pagina 1di 12

THE DEVELOPMENT OF THE MASAKLAW NA PANUKAT NG LOOB (Mapa ng Loob) Gregorio E. H.

del Pilar University of the Philippines


An earlier version of this paper was read at the 48th Annual Convention of the Psychological Association of the Philippines, Iloilo City, lat August 18, 2011.

Like their counterpart instruments in other parts of the world, the two general Filipino multiscale personality inventories --- Enriquez' Panukat ng Ugali't Pagkatao (1985) and Carlota's Panukat ng Pagkataong Pilipino (1987) --- attempt to comprehensively measure the important personality trait constructs in their target culture. Enriquez arrived at his list of 24 traits by using as sources interviews with college students, word associations, Filipino proverbs, the personality and social science literature, and dictionaries. On the other hand, most of Carlota's nineteen trait constructs were identified from interviews with nearly 300 informants from ages 13 to 80, who had been asked to describe themselves, a person they liked, and a person they disliked. The Masaklaw na Panukat ng Loob, or Mapa ng Loob, likewise aims to comprehensively measure personality trait constructs of theoretical and practical significance in Filipino culture, whence the qualifier "masaklaw" in the instrument's name. To do this, it operationalizes the Five-Factor Model (FFM) of personality traits (McCrae and Costa, 1992; John and Srivastava, 2008), which is widely recognized as a very broadly encompassing organization of personality traits. This decision was based on the view that the model is applicable to Filipino traits, a position that will be discussed in the next section. Much as a map indicates the location of geographical elements in two dimensions, the Mapa, with its five factor-dimensions, aims to locate both trait constructs and people in its five-factor space. Doing so would allow the users of the Mapa, as a research tool and as an instrument in the applied setting, to profit directly from the rapid accumulation of systematic findings that an integrative and consensual framework such as the Five-Factor Model makes possible (Gosling et al, 2003; Markey et al, 2004; Ozer and Bennett-Martinez, 2007; De Haan et al, 2012).

The Five-Factor Model and Filipino traits The adequacy of the Five-Factor Model in organizing indigenous Filipino traits was investigated in a study by Katigbak, Church, Lapena, Carlota and Del Pilar, 2002. The Filipino traits investigated were those measured by Enriquez' PUP, Carlota's PPP, and Church and colleagues' Panukat ng mga Katangian ng Personalidad (PKP), an adjective rating scale developed for research purposes using a comprehensive lexical approach (Church, Katigbak, & Reyes, 1996, 1998). The Filipino version of the NEO PI-R, or FNEO PI-R (McCrae, Costa, Del Pilar, Rolland, and Parker, 1998) was used as the main measure of the fivefactor model. It appeared, on the basis of factor analyses on the PUP,the PPP, and the PKP, supplemented with correlational analyses involving the FNEO PI-R, that Filipino traits, generally speaking, are well-represented by the five-factor model.

The PPP The rotated factor matrix resulting from the factor analysis of the PPP scales is shown in Table 1. It will be seen, that four, rather than five factors were obtained (Katigbak et al, 2002). All of the loadings are displayed, but a variable's highest loading is indicated in boldface. Loading highest on the first
Rotated Factor Matrix for Four-Factor Solution for PPP Scales
N = 387

Factor 1 0.27 -0.23 2 3 0.22 4 0.03 0.13 -0.01 -0.20 h2 0.56 0.46 0.61 0.60 0.63 0.51 0.57 0.74 0.66 0.66 0.56 0.68

English (Filipino) Interpersonal scales Thoughtfulness (Pagkamaalahanin) Social Curiosity (Pagkamadaldal) Respectfulness (Pagkamagalang) Sensitiveness (Pagkamaramdamin) Obedience (Pagkamasunurin) Helpfulness (Pagkamatulungin) Capacity for Understanding (Pagkamaunawain) Sociability (Pagkapalakaibigan) Personal scales Orderliness (Pagkamaayos) Emotional Stability (Pagkamahinahon) Humility (Pagkamapagkumbaba) Cheerfulness (Pagkamasayahin) Honesty (Pagkatamapat) Patience (Pagkamatiyaga) Responsibleness (Pagkaresponsable) Intelligence/creativity scales Creativity (Pagkamalikhain) Risk-Taking (Pagkamapagsapalaran) Achievement Orientation (Pagkamasikap) Intelligence (Pagkamatalino)

0.66
-0.30 0.40 -0.05 0.23 0.43 0.20 0.22

0.55
-0.09 -0.17 0.03 0.33 0.18

0.67 -0.72
0.34

-0.68
-0.04 0.18 0.07 0.01 0.04 0.20 0.22

0.46 0.68
0.16 -0.01

0.81
0.06 0.10 -0.12

0.81
0.04 0.46 0.06

0.80 0.57
0.27

0.75
-0.22
0.10 -0.07 0.13 0.28 0.10 0.30

0.44

0.54
0.33

0.27
0.16 0.15

0.61
0.57 0.66 0.64 0.48 0.57 0.66

0.65
0.22 0.31 0.31 0.32 0.28

0.76
0.31 0.16

0.64 0.52
0.38

0.56
0.27

0.65

Table 1. Rotated factor matrix for four-factor solution of the PPP scales. The hightest loading of each variable is indicated in boldface.

factor, which appears to be a blend of A and N, are the scales for Pagkamagalang (Respectfulness), Pagkamatulungin (Helpfulness), Pagkamaunawain (Capacity for Understanding), Pagkamapagkumbaba (Humility), and Pagkapasensyoso (Patience); and those for Pagkamaramdamin (Sensitiveness) and Pagkamahinahon (Emotional Stability). To confirm whether the interpretation of this factor as a blend of A and N was correct, factor scores were computed for Factor 1, and correlated with the domain scores of the Filipino NEO PI-R (henceforth to be referred to as FNEO PI-R.) Indeed, it turned out that the factor scores on Factor 1 correlated most with A (r = .53), and with N (r = -.57). 2

Three of the five scales which load highest on the second factor are clearly related to Conscientiousness, namely, Pagkamaayos (Orderliness), Pagkaresponsable (Responsibleness), and Pagkamasikap (Achievement Orientation). The other two,Pagkamaalalahanin (Thoughtfulness) and Pagkamatapat (Honesty), are made up of items that deal with culturally-valued behaviors as might be taught at home or at school (examples from the Thoughtfulness items are: "Kapag napuyat ang mga kasambahay ko nang nakaraang gabi at ako ay maagang nagising kinaumagahan, ingat na ingat akong gumalaw upang hindi sila maingayan o magising," "Kung oras na nang pagkain at wala pa ang ibang kasambahay ay ipinagbubukod ko sila ng pagkain nang hindi sila maubusan," "Kung gagabihin ako ng pag-uwi, ipinaalam ko ito sa aking mga kasambahay nang hindi sila mag-alala." Examples from the Pagkamatapat scale are: "xxx," "xxx," and "xxx." ) Thus, it is not surprising that these two scales load as well on what was interpreted as the C factor. Factor scores on this factor correlate most highly (r = .57) with the C domain score of the FNEO PI-R, supporting the interpretation of the second factor as the C factor. The third factor, on which the Pagkamadaldal (Social Curiosity), Pagkapalakaibigan (Sociability), and Pagkamasayahin (Cheerfulness) scales have their highest loading, is evidently the E factor. Indeed, the factor scores on Factor 3 correlate .71 with the E domain scores of the FNEO PI-R. Finally, the fourth factor, on which the Pagkamalikhain (Creativity), Pagkamatalino (Intelligence), and Pagkamapagsapalaran (Risk-taking) scales have their highest loading, seems interpretable as the O factor. The high negative loading of the Pagkamasunurin (Obedience) scale on this factor appears to reinforce this interpretation, since high O individuals, as is well-known, have a tendency not to conform. Indeed, the factor scores on Factor 4 correlate highest with the O domain score of the FNEO PI-R (r = .45). The fusion of the expected A and N factors into one in the PPP is hypothesized to be attributable to the underrepresentation of the N domain in the PPP (Katigbak et al, 2002). As had been mentioned, only the Pagkamaramdamin and the Pagkamahinahon scales are related to N, while five scales appear to be related to A.

The PKP The Panukat ng mga Katangian ng Personalidad began with 6,900 person-descriptive terms culled from a Filipino dictionary and reduced to 1,297 personality-related adjectives as agreed upon by Filipino judges and samples of college students (Katigbak, et al, 2002). Versions of the instrument containing from 281 to 861 adjectives were factor analyzed in three samples, yielding seven factors that replicated fairly consistently. Six of these had substantial overlaps with the five factors. In the study by Church, Reyes, Katigbak, and Grimm (1997), which made use of 281 adjectives, the correlations reported were with scores on the NEO Five-Factor Inventory. The second column of Table 2 shows the highest correlation of each of the obtained lexical factors with the five FFI scores. One sees that the lexical factors correlate highest with the FFI scale that one would expect it to. In the same study, a second set of scores on the Big Five was computed on the basis of random samples of Filipino adjectives from the 3

Filipino Factor

Five-Factor Inventory Scales and correlation

Filipino adjectives classified using Goldberg's Big Five taxonomy of adjectives


A, r = .81 C, r = .77 N, r = -.58 C, r = .- 54; A = -.45, N, r = .40 O, r = .56 E, r = .66

Makakapwa vs Makasarili (Concern for others vs Egotism) Disiplinado (Conscientiousness) Tiwala sa Sarili (Self-Assurance

A, r = .54 C, r = .33 N, r = -.33

Sumpungin (Temperamentalness)
Matalino (Intellect) Kalog (Gregariousness)

N, r = .22
O, r = .34 E, r = .44

Table 2. Factors from the 281-item version of the Panukat ng mga Katangiang Pilipino (PKP), showing each one's correlation with the Five-Factor Inventory scale (Column 2), and the Goldberg 1990 factor taxonomy adjective scale (Column 3) with which each correlates the most.

pool of 281 adjectives, consensually classified by 9 judges according to Goldberg's 1990 five-factor taxonomy of adjectives. The correlations are shown in the third column of Table 2, showing once again the Big Five trait correlating highest with each Filipino factor. It will be noted that the ad-hoc scales with which the Filipino factors correlate the most are identical to the FFI scales with which the Filipino factors were most correlated. An exception is the Sumpungin or Temperamentalness factor, which correlates most with C, then with A, but still highly with N. As a footnote, the seventh Filipino factor was labeled Mga Katangiang di Kanais-nais (Negative valence or Infrequency), made up of highly evaluative negative terms that are infrequently endorsed (e.g., troublesome, useless, drunkard, and stupid), and whose substantive status remains uncertain (xxx, 20xx). Lastly, when additional components are extracted, a religiosity factor sometimes appears. The PUP The PUP was factor anlyzed at the item level because many of its scales were too short to have adequate internal consistency reliability. Six factors were found to be interpretable, four of which resembled A, C, N, and E. These interpretations were confirmed by correlations between factor scores for these four factors and the corresponding FNEO PI-R domain scores, with correlations ranging from .53 to .60. The fifth factor produced factor scores whose highest correlation was with O (r = .28). However, since it was defined by a set of items that had very high rates of partcipant agreement or disagreement, and additionally, its correlation with O was quite modest and in fact counter-intuitive, it appears to be best interpreted as Communality, and therefore, nonsubstantive as well. Finally, the last 4

factor, defined by items indicating autonomy and will, was also correlated with the A domain (r = -.36), and more specifically with the Compliance facet (r = -.40). The above results appear to show that the five-factor model can well represent Filipino traits as they have so far been measured by the three existing Filipino inventories. In more technical terms, the five-factor model appears to capture a sufficient amount of variance in Filipino self-reports of typical behavior, as represented by traits. It thus becomes meaningful to have clear measures of these five factors using Filipino traits.1

The development of the Mapa ng Loob The initial work on the Mapa ng Loob was undertaken by a graduate class in Personality Scale Construction and two undergraduate classes on Psychological Measurement taught by the author at the University of the Philippines. The first field test resulted in a preliminary instrument made up of fifteen facet scales, three for each of the five domains. Each facet scale was made up of fifteen items that were to be reduced to eight or ten in a manner aimed at recovering a clear five-factor structure. Subsequent modifications were undertaken by a team headed by the author and made up of advanced graduate students in personality, most of whom were members of the graduate class just mentioned. Undergraduate students in Psychological Measurement taught by the author continued to participate by way of contributing items for field testing. A fourth facet was later added to each of the domains, which has brought the total number of facet scales to twenty. After the first field test, four more were undertaken in successive semesters, whose main objectives were the improvement of the reliability, content valdity (content coverage), and keying balance of each of the facet scales; and the improvement of the factor structure of the facet scales taken all together. Selecting the target constructs The inclusion of a trait construct in the Mapa ng Loob depended on 1) its theoretical and practical significance in Filipino culture; and 2) how good a marker it was of each of the five factors. The inventory was initially meant to be based on the Panukat ng Pagkataong Pilipino (PPP, Carlota, 1987). The way the trait constructs were selected for the PPP(see p. 1) argued for their theoretical importance in Filipino culture. As the graduate class of 6 was made up of students from a variety of backgrounds2, it could also keep well in sight, throughout the test construction process, the
1

This conclusion does not close the door to the inclusion, in later versions of the Mapa, of other Filipino traits that lie outside the factor space defined by the Five-Factor Model. Indeed, the developers of the Mapa ng Loob look forward to investigating candidates for such traits, such as religiosity.
2

The class was made up of the head of students affairs of a college in UP, a practicing counselor, a member of the teaching staff of the business college of UP, a staff member of the testing office of a tertiary institution in Manila, a teacher at a tertiary institution in Manila, and a member of the teaching staff at UP who worked in industry for more than ten years.

practical importance of the traits in a variety of settings (among students, in the counseling situation, in industry). Thus seven of the PPP scales, whichwere considered to be good markers of the five factors, were selected: Pagkamaramdamin for Neuroticism, Pagkamasayahin and Pagkapalakaibigan for Extraversion, Pagkamaunawain for Agreeableness, and Pagkamasikap, Pagkamaayos, and Pagkaresponsable for Conscietiousness3 (see Table 3 below). The selection of the remaining traits was likewise based, generally speaking, on the two criteria mentioned above. A notable exception is the fourth extraversion facet (Pagkamadaldal), whose inclusion is justified only by the second criterion, that it, its being a good marker for its domain (Extraversion). The definition for each of the trait constructs is given in Appendix A.
Basis for selection as part of the Mapa N1 Hina ng Loob Vulenerability to Stress Originally from PPP Pagkamahinahon N2 Pagkamaramdamin Oversensitiveness PPP N3 Pagkamapag-alala Worrying anxiety New; good marker N4 Pagkasumpungin Temperamentalness New; good marker, peculiarly Fil; PUP? E1 Pagkamasayahin Joyousness PPP E2 Pagkapalakaibigan Friendliness PPP E3 Pagkamasigla Energy New; good marker E4 Pagkamadaldal Loquaciousness New; good marker O1Kakaibang Pag-iisip Originality Originally from PPP Pagkamalikhain O2 Hilig sa Bagong Kaalaman Intellectual Curiosity New; good marker O3 Pagkamakasining Aesthetic Sensitivity New; good marker O4 Pagkamaharaya Openness to Fantasy New; good marker A1 Pagkadimayabang Modesty Originally from PPP Pagkamapagkumbaba A2 Pagkamapagtiwala Trust New; good marker A3 Pagkamaunawain Capacity for Understanding PPP A4 Pagkamapagparaya Compliance for Harmony New; good marker C1 Pagkamasikap Achievement-striving PPP C2 Pagkamaayos Order PPP C3 Pagkaresponsable Responsibleness PPP C4 Pagkamaingat Carefulness New; good marker Table 3. The scales of the Mapa ng Loob and the basis for each's inclusion in the core of the inventory. Scale

The initial selection also included Pagkamahinahon (N), Pagkamalikhain (O), and Pagkamapagkumbaba (A). Each of these PPP scales was significantly redefined and renamed. In the case of the Pagkamahinahon scale, the field test of the Neuroticism domain, of which it was a facet, included items being field tested for the Agreeableness domain. Correlational analysis confirmed misgivings that the anger-related items contained in the Pagkamahinahon scale correlated both with the items selected to represent the N domain, and those selected for the A domain. Consequently, these items were removed, and the scale renamed Hina ng Loob. In the case of Pagkamalikhain, it was judged to be at a higher level than the other constructs in the inventory, and was thus reduced to one of its facets, namely, original thinking, and thus renamed Kakaibang Pag-iisip. Concerning Pagkamapagkumbaba, it was judged to be too psychologically complex and reconceputalized and renamed as Pagkadimayabang.

Generation of the initial item pool and the first field test As previously mentioned, a graduate class in Personality Scale Construction and 2 undergraduate classes in Psychological Measurement taught by the author at the University of the Philippines contributed items for the Mapa scales. In particular, each undergraduate student was assigned two facet scales, and asked to write ten items in all, that is, 5 for each scale. Items from the PPP, plus the new items written by the graduate and undergaduate students, and the author, numbered approximately 650. After elimination of items on the basis of similarity, the surviving items at that point were rated by the graduate students, on a scale from 1 to 4, on the extent to which they were prototypical of the behaviors manifesting each facet trait. Only those with ratings of 3 and 4 were retained for the first field test. The smallest number of items for a facet scale was 19, for the Pagkamaramdamin facet, while the largest was 39, for the Lawak ng Isip facet (subsequently replaced by the Hilig sa Pagkamasining facet). For the domains, Neuroticsim had the smallest number, at 59; while Openness to Experience had the largest, at 90. Five separate questionnaires were constructed, one for each of the domains of the Five-Factor Model. The questionnaires were administered to classes at the University of the Philippines and at the Colegio de San Juan de Letran. To control for environmental testing factors, the questionnaires were arranged so that they would be given out in the fixed sequence N, E, O, A, C throughout all of the classes, so that each class would contribute approximately equal numbers of participants for each of the questionnaires. The plan was to have approximately 500 students from each institution, a hundred for each of the five questionnaires, so that combined, each questionnaire would be answered by 200 participants. In the end, our participants numbered 963, approximately half of whom were from either institution. Both samples represented a variety of majors. The sex distribution, for each domain, is as shown in the table, with females ranging from 58 (Extraversion) to 68 (Neuroticism) percent. Using the NEO PI-R criterion of 16% blank responses, no participant was dropped from the analyses. Reliability analysis was performed on each of the scales, values on the trial scales were systematically higher for the UP sample, but were satisfactory as well for the Letran sample.

Sampling plan for the first field test


Institution

N
100

E
100

O
100

A
100

C
100

Total

UP

500

Letran
Total

100
200

100
200

100
200

100
200

100
200

500
1000

Frequency Majors
Percentage of female participants

Actual participant data N E O A 197 188 191 191 Various 68 58 66 60

C 196 63

Total 963

Table 4. Sampling plan for the first field test. Item analysis was performed on each domain scale by randomly dividing each group into two subgroups. Separate reliability analyses using SPSS were run on each subgroup, with approximately 100 participants for each subgroup. For an item to be retained, it had to be correlated with its total at alpha = .05 for both subgroups. The two subgroups were then combined, and item-total correlations recomputed for the surviving items. In selecting the fifteen items that were to make up each of the facet scales, both the item-total correlation and the content of the item was taken into consideration. Specifically, we made sure that the fifteen items would cover all the identified subfacets of each facet.4 The reliability of each of the facet scales was computed and was found to range from .74 to .87, with a mean of .81.

Second field test

The use of the SPSS allows the identification of the set of items from a pool that has the highest value of Cronbach's alpha. It does this by indicating what the resulting Cronbach's alpha value would be if a particular item were to be deleted from the pool. It is thus possible to undertake an iterative procedure that removes, at each cycle, the "worst" item, that is, the item with the lowest item-total correlation. The removal of such an item would increase the value of Cronbach's alpha by the greatest amount. When this procedure is continued until Cronbach's alpha no longer increases with the removal of one of the remaining items, it will leave a core of mutually intercorrelated items with the highest possible alpha for the pool. However, this method could also remove all but that set of items that make up the largest facet of the target construct.

Preparatory to the second field test, a second review was conducted with the aim of ridding the scales of whatever couplets or triplets that remained. This resulted in some changes in some of the scales. The second field test sampled from the University of the Philippines in Diliman (N = 489) and St. Michael's College in Binan, Laguna (N = 316). Both samples were similar in terms the of the heterogeneity of majors represented, and the sex distribution (in terms of percentage of females, both UP and Binan had 62 percent). In terms of age, the UP sample was older and more heterogeneous, with a mean of 19.1 and a standard deviation of 2.58; while the Binan sample had a mean of 17.6 and a standard deviation of 1.64. As in the first field test, the data was screened for quality. One participant who had 59 blanks was eliminated, while the rest had twelve blanks or less. Again, similar to the results from the first field test, the reliability of the scales were noticeably higher in the UP sample, but were nevertheless satisfactory for the Binan sample. This was the case on all but one of the scales. The sole exception was the O3 Lawak ng Isip scale, for which the reliability of the Binan sample was a mere .30. This presumably is due to range restriction, since the reliability of the combined samples (UP and St. Michael's) remains at .60, the UP value. Nevertheless, it will be noted that this facet has the lowest reliabilty among the 15 scales, with the next lowest, the third A facet Pagkamaunawain, at.71.

UP Binan Total Sample N = 489 N= 315 N = 804 N1 Hina ng Loob 75 66 72 N2 Pagkamaramdamin 80 80 80 N3 Pagkamapag-alala 81 73 78 E1 Pagkamasayahin 83 78 81 E2 Pagkapalakaibigan 91 85 89 E3 Pagkamasigla 76 66 72 O1 Kakaibang Pag-iisip 85 68 80 O2 Hilig sa Bagong Kaalaman 81 74 79 O3 Lawak ng Isip 60 30 60 A1 Pagkadimayabang 81 75 79 A2 Pagkamapagtiwala 77 70 75 A3 Pagkamaunawain 75 62 71 C1 Pagkamasikap 83 73 79 C2 Pagkamaayos 87 79 85 C3 Pagkaresponsable 86 77 83 Table 5. Cronbach's alpha reliability of the Mapa facet scales in two samples and in the total sample.

Scale

Results of the factor analysis Two analyses were done on the data, a principal components analysis, and a principal axis factoring, both of which were followed by varimax rotation. Both analyses indicated five factors using the eigenvalue greater than one criterion. For both analyses, the minimum loading for display was slightly lowered to .35, from the usual .40, to make certain expected clusters appear more clearly. Following is the result of the Principal Components Analysis, with the loadings of the three facets that load highest on the factor highlighted. As can be seen, the first factor, defined primarily and clearly by the C facets, loads the Pagkamasigla facet of the E domain, and the first two facets of the O domain, Kakaibang Pag-iisip and Hilig sa Bagong Kaalaman. All three are primary loadings.

Component 1 N1Hina N2Dam N3Ala E1Saya E2Kai E3Sig O1Iba O2Bago O3Lawak A1Dimayab A2Tiwa A3Unawa C1Sik C2Org C3Res .800 .828 .845 -.367 -.478 .552 .638 .697 .551 .430 .619 2 .826 .802 .890 .854 .798 .461 -.365 .390 .425 .855 3 4 5

Table 6. Principal Component Analysis of the Mapa facet scales, with Varimax with Kaiser Normalization Rotation.

We looked at the items in these three scales to find the probable sources of these unintended loadings. In the Pagkamasigla scale, it might have been items that made use of the term "gawain" or "ginagawa" in a way that implied a task, rather than a mere activity. Examples are: "Mabilis kong tinatapos ang anumang gawain;" "Gustong-gusto ko ang laging may ginagawa," and "Kapag tapos na 'ko sa isang gawain, agad-agad akong naghahanap ng bagong gagawin." Consequently, task-orientation, a feature of C, likely influenced responses to these items. In the Kakaibang Pag-iisip facet, the loading on the C factor might have been due to such items as the following: "Magaling akong mag-isip ng kakaibang solusyon para sa isang problema," and "Masasabi kong kakaiba akong dumiskarte." These items probably tap self-perceptions of general ability, or competence, and therefore, conscientiousness. In the facet Hilig sa Bagong Kaalaman, the culprit items might be those such as "Nagtatanong ako para makakuha ng sapat na impormasyon sa mga nais kong malaman," "Pakiramdam ko'y kulang pa ang nalalaman ko kung kaya't sinisikap kong makakuha pa ng mas maraming impormasyon sa kahit anong paraan," and, "Nakakayamot ang mag-aral 10

ng mga bagong bagay" This last one is negatively keyed for Hilig sa Bagong Kaalaman, but possibly for Conscientiousness as well. As for the N, E, A and O factors, these are quite clear. It will be noted that on the E factor, the Agreeableness Pagkadimayabang facet has a secondary loading slightly higher than that of the Pagkamasigla facet, although this makes sense (that is, those high on E tend to be Mayabang, or appear to be Mayabang, ie, they like being praised in front of others, they like showing off their talents, and so forth). The O facets, despite the primary loadings of the first two on the C factor, also form a rather clear O factor that loads the O3Lawak ng Isip facet highly. The two other facets, O1Kakaibang Pag-iisip, and O2Hilig sa Bagong Kaalaman, have low and moderate loadings on this factor, respectively. The Principal Axis Factoring yielded the rotated factor matrix shown in Table 7 below. The structure is quite similar to that obtained in the preceding analysis, with the C factor loading E and O facets, and the N, E, and A facets grouping pretty much as expected. The most notable difference between the two structures is the O factor. In PAF, the first two O facets define the O factor, while the O3 facet, which now loads on A, has only a low secondary loading on the O factor. This shift is the result of using communalitiy estimates in the prinicipal diagonal of the correlation matrix in PAF, while unity, of course, is used in PCA. Since, as was mentioned, O3 has the lowest reliability, and consequently, also the lowest communality, the effect on the O structure is quite pronounced. The question therefore remains as to the correct definition of the O factor as measured by the current Mapa scales, whether it is closer to Lawak ng Isip, on the one hand, or to Kakaibang Pag-iisp and Hilig sa Bagong Kaalaman, on the other.

Factor 1 N1Hina N2Dam N3Ala E1Saya E2Kai E3Sig O1Iba O2Bago O3Lawak A1Dimayab A2Tiwa A3Unawa C1Sik C2Org C3Res .706 .820 .787 .433 .418 .436 .596 .437 .412 2 .748 .697 .829 .783 .731 .379 .613 .503 (.296) 3 4 5

Table 7. Principal Axis Factoring of the Mapa facet scales.

11

Third field test The third field test had as its prinicpal aims the reduction of the fifteen-item facet scales to eight items, and the improvement of the factor structure of the instrument by removing the unintended loadings of the facets shown in Table 7 (A1Dimayabang on the E factor, and the E3Pagkamamasigla, the O1Kakaibang Isip, and the O2Hilig sa Bagong Kaalaman laodings on the C factor). In addition, it was decided to replace the problematic Lawak ng Isip scale with an aesthetic sensitivity scale.

12

Potrebbero piacerti anche