A Study of Developing an Attidude Scale for Using Scientific Process Steps

Calismanin amaci ortaokul 6, 7 ve 8. sinif ogrencilerinin fen bilimleri dersi kapsaminda bilimsel surece yonelik tutumlarini olcen bir olcek gelistirmektir. Literaturde bilime ve fen dersine yonelik tutum olcegi gelistirme calismalari yaygin olarak bulunmaktadir. Bu olcek gelistirme calismasinin digerlerinden farki bilime karsi tutumlari olcmek yerine bilimsel surec basamaklarini kullanmaya yonelik tutumlari olcmeyi hedeflemis olmasidir. Olcek maddeleri uzman gorusleri dikkate alinarak hazirlanmistir. Gelistirilen olcek uc farkli ortaokulun 6, 7 ve 8. Siniflarinda okuyan toplam 321 ogrenciye uygulanmistir. Yapilan faktor analizi sonuclarina gore, tek boyutlu, 5’li likert tipi 33 maddelik bir tutum olcegi gelistirilmistir. Olcegin Kaiser-Meyer-Olkin (KMO) degeri .94, Barlett testi degeri 4386.48, Cronbach alpha degeri .94 ve Test-tekrar-test guvenirligi .91 olarak bulunmustur.


Introduction
One of the main aims of science education today is to enable students to acquire scientific thinking skills that are to teach how to use scientific process (Erdoğan, 2007;Tan & Temiz, 2003).The core elements of science such as investigating, questioning and experimenting are the fundamentals of scientific process.While the concepts of scientific thinking or critical thinking concepts have been viewed as methods of science previously, today scientific method is defined with some distinct skills.In general, this method is considered under the title of scientific process skills (Özgelen, 2012).
Enabling students gain scientific process skills is one of the main aims of current science education programs (MEB, 2013).Raising individuals who investigate, question, can link the science subjects to daily life, use scientific method to solve problems, approach the phenomena with a scientist's point of view, constitutes the fundamentals of science teaching (Kaptan, 1999).To possess these characteristics mentioned, it is essential for the individuals to learn research methods and practices.
Scientific process skills are thinking skills that are utilised to structure information determine the problem and reach to conclusion (Bybee & DeBoer, 1993).Scientific process has been defined by various experts dividing it into different steps.In general, the process is divided into two parts as basic and combined skills (Padilla, Okey & Garrard, 1984).Basic skills are defined as observing, measuring, classifying, predicting, projecting and scientific communication (Duran, 2008;Karar, 2011;Kılıç, 2003;Rezba, Sprague, McDonnough & Matkins, 2007;Serin, 2009).Combined scientific process skills are, in general, considered under the titles of determining the variables and verification, constructing hypothesis, interpreting the data, experimenting and modelling (Duran, 2008;Kaptan, 1999;Rezba et. al., 2007).
Science education, which aims to enable students gain an understanding of the nature of science and ability to approach the problems scientifically, tries to make students develop positive attitudes towards science as well (MEB, 2013).Developing scientific process skills of the students, who are going to use the steps of the process, is going to be affected by the students' attitudes towards the process.Scientific skills can be effective when students develop positive attitudes to the process.Allport (1935, p.810) defines attitude as "a mental and neural state of readiness, organised through experience, exerting a directive or dynamic influence upon an individual's response to all objects and situations with which it is related".Eagly and Chaiken (1993, p.1) presents another more recent definition as "A psychological tendency that is expressed by evaluating a particular entity with some degree of favor or disfavor".Beliefs are the building stones of the attitudes (Özgüven, 1994).Whereas an individual saying "Mathematics is a matter of talent, I am not talented" expresses his/her belief regarding mathematics, not liking mathematics lesson is an indicator of attitude.Attitude has three dimensions which involve direction, degree and intensity (Allport, 1935;Özgüven, 1994).The dimension of 'Direction' regards attitude attribution as liking or disliking whereas 'Degree' refers to the emotional tone of this aspect.On the other hand, 'Intensity' is related to how the attitude is felt by the individual and to what extent it is reflected to (Eagly & Chaiken, 1993).
Measurement of attitudes during learning and teaching process is beneficial in terms of several issues such as predicting individuals' future behaviors by determining their attitudes to specific situations, changing their attitudes or revealing their actual emotions to develop new attitudes (Anastasi & Urbina, 1997;Demirbaş & Yağbasan, 2006;Hançer, Uludağ & Yılmaz, 2007).By this means, causes of individuals' behaviours can be determined and education environments can be reorganized according to the needs.This would enable an opportunity to provide higher quality education offered to all individuals.
Previous studies have shown that individuals' attitudes have an effect on their acquisition of learning and skills.For instance, students with positive attitudes exhibit more positive behaviours towards the course which leads them to be more motivated to complete the given tasks and to put more effort to learn (Braten & Stromso, 2006;Duarte, 2007;Özden, 2009).
In this regard, this study aims to develop a scale which reveals students' attitudes towards using scientific process steps.The study differs from the previous studies in that it focuses on developing a scale for students' attitudes towards using scientific process steps, rather than towards science or science lesson.Thus, the possible relationship between students' attitudes towards using scientific process steps and their level of gaining scientific process skills can be identified.

Participants
Universe of the study involves grades 6, 7 and 8 lower secondary school students in Turkey during 2013-2014 terms.321 students studying at three different schools in the North-West region town of Turkey, constituted the sample group.The distribution of the participant students according to the grades is as follows: At the first school 34, 30 and 36 students from 6th, 7th and 8th grades involved whereas at the second school 55, 59 and 61 students from 6th, 7th and 8th grades took part and 15, 18 and 13 students from 6th, 7th and 8th grades participated from the third school in the study.The distribution of the participants regarding gender is presented in Table 1 below.

Process of Developing the Attitude Scale for Using Scientific Process Steps
The planned scale had 33 Likert type items and the answers were given in five degrees as: "I definitely do not agree: 1", I do not agree: 2", "I'm not sure: 3", "I agree: 4" and "I definitely agree: 5".All items were prepared as positive statements.In the process of item formation, basic scientific process skills have been taken into account.Integrated scientific process skills, which require high cognitive level, have not been considered since the participant students' levels have already been taken into consideration.
The stages of developing the scale have been planned as follows:  Composing the attitude items,  Enquiring expert opinions,  Conducting a Pilot study,  Factor analysis of items,  Calculation of reliability.

Composing the Attitude Items
Prior to composing the attitude items, a comprehensive literature review was done covering scientific process and application steps of this process.Then, previous scales measuring students' attitudes towards the process were investigated.In addition to these, science education programme of secondary school students, whom the scale is going to address, was taken into consideration.By examining these and documents of psychological tests regarding developing attitude tests, a 35question-item pool was composed.

Enquiring Expert Opinions
Following completion of the items pool, the items were presented to some professionals for corrections and some corrections were done.These professionals were two science teaching experts, three science teachers, three educational sciences experts, two psychologists and two Turkish language teachers.Besides these, three students' opinions (one student from each grade 6th, 7th and 8th) were asked about the clarity of the items.After the experts' evaluation, 2 items out of the original 35 were excluded from the scale and 12 statements in the scale were revised.Consequently, a 33-item 5 degree likert type scale was developed.

Pilot Study
The attitude scale for scientific process was applied to 321 students (157 girls, 164 boys) from 6th, 7th and 8th grades at three different schools.Equality in number of the participants involved has been taken into consideration for each school and grade.According to the literature in scale development, the number of participants is required to be at least five times bigger than the number of items for high validity (Bryman & Cramer, 2005).This notion was taken into account and sufficient number of participants were provided.The scale was administered to the students during psychology and rehabilitation lecture and students were asked to complete the scale in 30 minutes.The aim of the study was explained to the students and they were confirmed not to write their names whether they did not want their names known.

Factor Analysis of the Items
Following the expert opinions, content validity of the scale was provided and application was done.Scores from the application of the scale were put through factor analysis to reveal the construct validity.Construct validity is important for determining to what extent of accuracy a measurement tool measures the target values (Klinie, 2005;Tavşancıl, 2002).Through factor analysis, whether numerous variables can be expressed with a few variables, is revealed and if so, it is revealed what these factors are.There are two kinds of factor analysis methods, explanatory and confirmatory (Brayman & Cramer, 2005).While explanatory factor analysis is used to develop a new test, confirmatory is used to examine reliability and validity of a pre-existing test (Erkuş, 2012).
Since our aim is to develop a new scale, explanatory factor analysis has been chosen in this study.Each items' factor weighing, possible factor number in the scale, and variance percentage of factors were discovered through using principal component analysis in SPSS programme.Besides these, Kaiser-Meyer-Olkin (KMO) test and Barlett test results which are useful to check conformity of the scale were calculated.Kaiser-Meyer-Olkin (KMO) test shows the sufficiency of the data, the factor analysis is performed on, to measure the structure whereas Barlett test gives information on whether the items can compose a texture.Once these values are adequate, factor analysis can proceed.Kaiser (1974) says that if value of KMO is higher than .60,a scale is suitable to subject to factor analysis.Also Barlett test result must be significant with a p<.00 degree.KMO and Barlet test values in this study are presented in Table 2. Table 2 shows that KMO value is higher than .60(.94) thus internal consistency between items is high.Barlett test value is significant with a p=.000 value; distribution of universe is found normal.These results indicate that the scale is suitable for factor analysis.
Consequently to the analysis, for deciding the item and factor numbers to be included in the scale matters should be considered as follows.Degree of the relationship of items with factor is determined by the factor weighing value.For the items to be included in a certain factor, the factor weighing value should be at least .30.The factors to which the items belong are determined by looking into the weighing values and an item is included in the factor with higher weighing value with at least .10difference.
The factor number of the model is as the same number of factors with an eigenvalue higher than one.Scree plot shows the possible factors of the model more clearly.The second criteria of determining the factor number is the amount of variance clarified by the factors.Each factor must clarify at least 5% of the variance and in total, factors must clarify 35% of total variance (Klinie, 2005;Büyüköztürk, 2002).

Scree Plot
Component Number  Through these analyses, items were accumulated under 5 factors revealing an eigen higher than one (Chart 1).Variance clarified by 5 factors was 51.19%.For each individual factor the clarified variances were: 35.88%, 4.59%, 4.10%, 3.42%, 3.19% respectively.As seen from these values, there was a big gap between the first factor and the others.First factor has a sufficient clarity degree of total variance as higher than 35.00% (Table 3).Moreover, the other factors could not reach the required variance level of at least 5.00%.According to these results, it was decided that the scale can be one factor test and there was no need for rotation.
On the other hand, factor weighing values of the scale items were found to be between .42 and .65 (Table 4).Since the weighing values were higher than .30,all 33 items were evaluated and none were excluded at this step.Hence, the scale was accepted as one factor and 33 items.

Calculation of Reliability
After the factor analysis, and final revisions, Cronbach Alpha internal consistency score of the scale was calculated as α=.94.For reliability internal consistency coefficient is required to be at least α=.70 (Bryman & Cramer, 2005).Hence, the scale was shown to have high internal consistency.Furthermore, reliability of the scale items was calculated through test-retest correlation analysis.For this, the emerged structure was reapplied to the same group two weeks later and consistencies of the answers were compared.
However, the second application could cover 225 of the 321 students that had taken the first test.Pearson product-moment correlation was analysed through SPSS and found as .91 which indicate sufficient and rather high reliability (Mayer, 2013).The values arising from all these analyses and are summarized in Table 5, together with the basic information regarding the scale.

Discussion & Conclusion
Attitudes of students towards science and towards carrier choices in science have been object of curiosity since 1960s (Osborne et al., 2003).Among the prominent reasons, the decrease in students' preference of science fields and deterioration in academic success can be mentioned.Hence, measurement of student attitudes that can be associated with success in science and comparing those with the success has become a necessity.
Looking into the relevant literature, several scales measuring the attitudes towards science and science course are available, which evaluate the attitudes under different factors.For example, Kind et al. (2007) developed a multi-factor scale measuring in-school and out-school attitudes towards science, content of science, importance of science, doing practice in science and working in science in future.In another study, Cheung (2009) developed a scale measuring the attitudes of students towards chemistry course in four dimensions namely liking theoretical and practical chemistry courses, believes regarding functionality of the chemistry course and tendency to learning chemistry.In the scale they developed, Francis and Greer (1999) focused on attitudes towards importance of science, the science course and science as a carrier choice.Attitude scale of Pell and Jarvis (2001) included subtitles of interest in science, independent research tendency, liking the school, social content of science and science as a difficult subject.Finally, Nuhoğlu (2008) addressed the attitudes towards the science and technology course as attitudes towards the course and towards the activities performed.
As it emerges from these examples, the present attitude scales do not measure the attitudes towards using the scientific process steps, which are inherent to the teaching of science.With the purpose of filling this gap in the literature, this study aimed to develop a scale which can measure attitudes of lower-secondary level students towards using scientific process steps.By this means, investigating the effects of students' attitudes towards using scientific process steps over their gaining of scientific process skills is conducted.
In brief, Attidude Scale for Using Scientific Process Steps is a 5 point Likert type scale measuring the degree of attitudes as from "I definitely do not agree: 1" to "I definitely agree: 5" points (For the last version of the scale, see the appendix 1).For each item, the lowest point 1 indicates negative attitude.As the points increase the attitudes turns to positive and 5 points indicate the highest degree of positive attitude.As a result of factor analysis, the scale was concluded as a one dimension test covering 33 items.The points that can be scored range from the lowest 33 to the highest 165.
Internal Coherence analysis of the total scores resulted in a Cronbach Alpha coefficient of .94indicating sufficient reliability.Furthermore, test-retest analysis revealed high correlation (.91) between the first and second applications of the scale.Consequently, development of the scale was concluded and a valid and reliable attitude scale was achieved towards using of scientific process steps.

Chart 1 .
Eigen chart of the factors.

Table 1 .
Gender Distribution of the Sample Group.

Table 3 .
State of Factors in Clarifying the Total Variance.

Table 4 .
Factor Weighing Values of The Single Factor Scale.

Table 5 .
Data regarding the Scale.