A study of developing attitude scale for piano teaching

In this study, the development phases of a measurement instrument, which can be used for measuring the piano teaching attitudes of piano teachers, are investigated. Validity and reliability studies of the scale, which was developed in Turkey’s circumstances, were carried out with 196 piano teachers giving piano lectures in different districts of Turkey in the year 2015. Content validity, principal component analysis and confirmatory factor analysis together with construct validity and total correlation of items of piano teaching attitude scale, which was created by 25 positive, 10 negative, total 35 items including fivefold Likert type alternatives, were examined via Cronbach Alpha, lower and upper 27% t test. The results show that the validity of the scale, which is developed by the validity and reliability studies conducted and ended up with 3 factors and 19 items, is high and that it can be used as a scale for measuring the teaching attitudes in piano education.


INTRODUCTION
Musical communication realized through piano, which is acknowledged as the most universal and most fundamental instrument in terms of acquiring-understanding the skills of playing-listening-reading music, creating musical knowledge and providing a basis for other musical studies, is significantly influential for creating desired individual behaviours/behavioural changes of individuals (as cited in Kasap, 2004from Buchanan, 1964;Vernezza;MEB, 2006from Uçan, 1988).However, despite cultural differences, students receiving piano lectures all over the world may display four different "problematic learning" models such as possessing limited technical skills despite musical sensitivity, lacking musicality despite overcoming technical problems, progressing slowly due to going into details and overlooking details despite possessing intuition-musicality. Though, a student, who is ready to constant change and who can believe the importance of changeforgettingdiscountenance -unleashing (abandoning) -pushing the limits of thought and the necessity of intellectual and action freedom, would not exhibit such problems (Ercan, 2008;Baştuğ, 2009).In this context, practising of a curriculum while taking individual differences into consideration and using different methods/techniques while staggering them pedagogically by the piano teacher, who are the most fundamental element of an effective piano education, matters in terms of students to play piano well, to enjoy it completely, to be motivated, not to be disappointed and not to get bored; because, education, which is an activity of affection and reaches utopia, is kind of a fantasy world of the wishes for the future and means to be able to overcome all difficulties and to untangle problematic issues meticulously (Gökbudak, 2013(Gökbudak, , 2005;;Yazıcı, 2013;Say, 2005).
Because real education occurs with the "activities practiced in the classroom for the students" after the teacher closes the classroom door, educational practices draw the limits of education as a significantly determinant factor in terms of students' ability to acquire/not to acquire gains (Fidan, 2012;Bandura, 1995).In this sense, due to the possibility of a teacher with negative attitudes to rapidly and profoundly inflict damage on education, the concept of "attitude" has a great importance with regard to influencing and changing thoughts and behaviours of the teacher and determining the quality of teaching practices (as cited in Özmenteş andÖzmenteş, 2009 from Mullins, 1984;Yazıcı and Kılıç, 2015).Hence, when the fact that an educator can enable students to acquire positive/negative behaviours through his/her excitement, personality, behaviours and knowledge is considered, the piano teaching attitudes of piano teachers come to the forefront; because, educator, thanks to his/her extraordinary difficult-complicated expert/technician role in creating human behaviour, performs an important function related with learning/ teaching processes such as accelerating and correcting a behaviour, which can originally be acquired slowly/falsely (Semerci and Semerci, 2004;Alkan and Kurt, 2014).
Attitude, which forms one of the main subjects of social psychology due to its effects on social perceptions and behaviours, has an important influence on learning /teaching process covering acquiring and changing knowledge, skills, strategy, belief and behaviours.Thereby, an educator possessing positive attitudes in teaching processes will try to make the lecture more fun, always provide feedback, enrich the personal development of the students, include all students in the learning process and teach the lesson with the help of rich methods (Tolan, 1983;Schunk, 2009;Yılmaz, 2008).Thus, "full understanding", in which a planned education service sensitive to the students is provided, suitable and timely support is given to those facing learning difficulties and almost all students develop high-level learning skill, will be ensured (Bloom, 2012).On the other hand, negative attitudes may drag teachers to disapproval, refusal, dislike of students, failure to improve himself/herself, being unexcited and unwilling in their professional life and consequently may negatively affect exhaustion level of the teachers.A piano teacher suffering from emotional exhaustion will have difficulty in focusing on his/her job, feel strained and frustration, experience unwillingness for education and his/her per-sonal achievements will deteriorate.As a result, quality of education will decrease and students will be adversely affected in terms of potential gains they may acquire.Yet, if education service is of high quality, students will experience a quality educational activity.However, due to negative attitudes, learning processes are encumbered with "errors" reaching huge amounts, whereas, an effective and highly satisfactory learning by most of the students is only possible with an educational atmosphere with "minimum errors" (Semerci and Semerci, 2004;Kılıç and Yazıcı, YAZICI 359 2012;Duman, 2008;Bloom, 2012).Therefore, determination of piano teaching attitudes of piano teachers is of particular importance for increasing the quality of education and level of learning of students.Though there is no sample of a scale for determining the attitudes of piano teaching in the literature in Turkey; for this purpose, in this research it is aimed to develop a Likert-type scale for determining attitudes of piano teaching.

METHODOLOGY
In this research, which is an effort aimed to develop Likert-type scale, successive functional steps, such as defining attitude (attribute) to be measured, engineering test scale, applying the test and analysing acquired data, are as followed (Tezbaşaran, 2008).

Participants
The research conducted with randomly selected 196 piano teachers teaching piano in the different provinces of Turkey in 2015.61.7% of the participants are females (n=121) and 38.3% are males (n=75).

Defining the attitude (attribute) to be measured
In this step, the literature is reviewed; the situations to be observed in relation to the attitude to be measured are projected and worded in accordance with the scaling technique used.

Engineering the test scale and content validity
Firstly, the scale material is prepared and projected so as to be used by the researcher and reproduced.In line with the comments of two academicians specialized in psychological counselling and guidance, the scale prepared by the researcher in the wake of necessary simplification was scaled down to 35 items and in order to realize the content validity questions and size of the scale together with the assessment forms were sent to two academicians specialized in piano teaching.It was reported that they agree that items are suitable for the field and the wording is adequate in terms of comprehensibleness.The last version of the questionnaire was published online and the links of the questionnaire were emailed to piano teachers.Application lasted for ten days, data analysis only started after 196 participants, which is fivefold of the item number of the scale, were reached (Kline, 1994;Harrington, 2009).

Analysis of the data acquired via test scale
This step, in which SPSS and Lisrel programs are used, consists of functional steps such as scoring the answers given to the items, calculating the raw scores of the individuals in the scale, the characteristics of the distribution of raw scores, the characteristics of the distribution of item scores and item analysis.Scale is a measuring instrument consisting of fivefold Likert-type (1=Never, 5=Always) 35 questions planned to be grouped under 3 sub-dimensions.The items of the scale and the dimensions they belong are shown in Table 1.7 items (1,3,20,21,32,33,35) in the scale form the sub-dimension of "The importance given to teaching piano", 14 items (2, 4,5,6,8,9,11,14,15,16,17,18,19,26) constitute the sub-dimension of "The compassion for teaching piano", 14 items (7,10,12,13,22,23,24,25,27,28,29,30,31,34) comprise the sub-dimension of "The professional approach In order to measure construct validity, explanatory factor analysis (EFA) and confirmatory factor analysis (CFA) were carried out.Within the scope of the reliability study, item analysis methods consisting of total correlation of items, Cronbach Alpha, lower and upper 27% quartile t tests were applied.

Explanatory factor analysis
Within the explanatory factor analysis, which is multivariate statistic aiming to find and explore limited number of unrelated and conceptually meaningful new variables (factors, dimensions) by gathering a number of related variables, an operation for finding factors based on the relations between variables is conducted.In the factor analysis, the factor weight values are high (while being 0.45 or more is a good measure, it can be decreased to 0.30) in selecting the items not measuring the same construct and items are observed to have high weight value for a single factor, while to have lower weight value for other factors (the factor weight value difference between a factor for which an item has the highest factor weight value and other remaining factors should be minimum 0.10) (Büyüköztürk, 2011).
The KMO in the explanatory factor analysis was measured as 0.857, while significance level in the Bartlett's Sphericity Test is measured as p<0.01.It is observed that it is appropriate to make an explanatory factor analysis with the data acquired by the application carried out with 196-person-wide sample (Table 3).The results gathered form the first Varimax rotation are shown in Table 3.
When Table 3 is examined, 35-item scale forms a 9factor structure which has a value higher than 1.However, because the current situation is not in conformity with the 3-factor structure identified in the content validity, scree plot graphic was examined for determining the most important factor number.When the scree plot graphic was examined, the slop significantly disappears after the third factor, and three factors are important (Figure 1).
For the conformity of factor structure, the weight values of the items in the factors they belong, the weight values in other factors and the difference between weight values in more than one factor were observed.When Table 3 was examined, it was observed that 15 items (33, 35, 2, 4, 5, 6, 8, 9, 26, 7, 10, 12, 13, 24 and 34) are not placed in conformable dimensions and have higher weight values in the dimensions they do not belong to.
Because of the unconformable factor structure of the scale, non-observant items were incrementally eliminated and the results shown in the Table 4 were acquired.A total of 16 items, including all of the 15 items that are not in conformity with the factor structure in the first Varimax rotation and 3 rd item, were removed from the scale.When the structure acquired by the latest Varimax rotation was examined, it was observed that there were 3 factors whose values are higher than 1 and that there are 19 items falling into 3 sub-dimensions.The contributions of the 3 sub-dimensions to the variance are respectively 21.50% -21.50% -12.50% and the total variance for the scale in general was found to be 55.06%.It was also observed that the weight values of the items in the factors they belong varied between 0.56 and 0.83 and that every single item has the highest weight value in their own factor.

Confirmatory factor analysis
In confirmatory factor analysis and structural equation modelling, Ki-square statistic (X 2 ), the ratio of ki-square statistic to degree of freedom (X 2 /df), statistical significance of individual parameter estimations (t value), (SRMR, GFI) "based on the inclusions", (NFI, NNFI, CFI) "based on the independent models" and root mean square error of approximation, which are classified as conformality indices, are used for the assessment of   (Bayram, 2010).Acceptance and good conformality values of the conformality indices are given in Table 5 (Bayram, 2010;Sümer, 2000;Raykov, 1997).

Item No
According to Table 5, it was observed that the 9-factor structure of the scale and its conformality indices could not reach adequate levels.When Table 6 and Figure 2 were analysed, it was also observed that there are quite a number of items standardized factor weight of which are below 0.50.
When the acquired findings and the results of the explanatory factor analysis were assessed simultaneously, it was precipitated that there are some items requiring to be removed.Therefore, starting to eliminate the items with the lowest factor weight, it was determined that the scale reaches its best structure with 3 factors and 19 items (Table 7).
Findings show that the conformality indices turn out to be in acceptable and very good levels in general after modification connections, that the t values for each item is significant at 0.01 level and that the model displays a decent conformality (Figure 1     Factor weights standardized in accordance with the results of confirmatory factor analysis varies between 0.55-0.70 for "The importance given to teaching piano", 0.54-0.77for "The Compassion for teaching piano" and 0.55-0.74for "The professional approach towards teaching piano" (Table 8; Figure 3).

Reliability of the scale
Item-total correlation, as one of the item analysis methods applied as part of reliability study, explains the relationship between scores gathered from the test items and the total score of the test.Positive and higher values for item-total correlation indicate that the items illustrate similar behaviours and that the test has a high level of internal consistency.In general it is possible to claim that 0.30 and higher values for item-total correlation mean that item can successfully differentiate individuals while items with values between 0.20-0.30can be included in the test only when it is necessary.Another method used for item analysis is lower and upper quartile t-test.A result of significant difference between groups, which is gathered by examining the difference between average item scores of the lower 27% and upper 27% quartile according to the total test score by using unrelated t-test )(Çokluk et al., 2010;Meydan

Figure 1 .
Figure 1.Scree plot graphic gathered by the first varimax rotation of the scale.

Table 1 .
Planned factor structure of the attitude scale for piano teaching.

Table 2 .
Descriptive statistics of attitude scale for piano teaching.

Table 3 .
The results of the explanatory factor analysis (first varimax).

Table 6 .
DFA Results of the Scale (first results).