THE IMPORTANT ATTRIBUTES OF LOCAL AND IMPORTED APPLE : A FACTOR ANALYSIS APPLICATION Atribut Penting

This study aims to fi nd the important attributes of local and imported apple. A consumer survey was conducted in Java Island representated by four big cities: Malang, Surabaya, Yogyakarta, and Bandung. Three categories of purchasing place at each city are supermarket, fruit stall and traditional market. Accidental sampling was done to choose the samples of consumer. The total respondents interviewed was 240 consumers consisting of 120 for each attribute. The data of apple attributes and purchasing place were analyzed by exploratory factor analysis supported by SPSS software. The result suggests that the important quality attributes of local and imported apples are grouped in each two latent factors namely quality and cultivar. Meanwhile, purchasing place is grouped in one latent factor. All quality attributes of local and imported apples have positive correlation with the apple quality, but there is a different fi nding in which apple fl avor of local and imported apples are not the most important attribute that ranked fi rst based on consumer assessment. The factor of apple cultivar always associates with the apple size, both on local and imported apples. Rome Beauty (local apple) and Fuji (imported apple) have greater average size than other cultivars. Thus, these attributes are already attached one another. The factor of purchasing place becomes an important part which accompanies the apple’s quality attributes and is considered by consumers in purchasing apple.


INTRODUCTION
suggested a new approach to consumer theory.According to him, it each good has intrinsic properties that are relevant to the theory of consumer.
Traditional consumption theory is unable to explain consumer reactions to a new commodity and quality's variation.Lancaster (1966) developed an assumption that consumption is an activity that uses the goods as inputs and a collection of characteristics as the output.
The Lancaster theory has become an important base in developing hedonic demand analysis so called hedonic price.
It is defi ned as the implicit price of the attributes and disclosed to represent the economic value of the observed prices within differentiated products and the specifi c number of related characteristics (Rosen, 1974).Price, p(z) = p (z1, z2, ..., zn) is defined as each point on a plane and guide the locational choice of consumers and producers with regards to the characteristic of package which is purchased and sold (Rosen, 1974).

Figure 1. Consumers' Equilibrium on
Hedonic Price (Rosen, 1974) fruits, as shown in Table 1.This study added and modifi ed some attributes.Firstly, this study modifi es the apple's size to weight, since it is more measurable.The apple's weight is measured from the volume of apples bought by consumers divided by the number of apples.Secondly, this study does not use price as the attribute because the price is used as the dependent variable of a hedonic price on the next study.Thirdly, this study uses origin (local and imported) as the compared object.Fourhtly, we add juiciness attribute.Fifthly, the factor analysis includes purchasing place because it will infl uence the apple price.

Table 1. Apple Attributes Investigated by Previous Studies
Researcher & year Attributes Manalo (1990) size, flavor, price, color, crispness Péneau, et al (2006) size, flavor, appearance, aroma, freshness, cultivar, organic Rahayu, et al (2012b) size, price, color, texture, sponge nets packaging Skreli & Imami (2012) size, varieties/cultivar, apple origin (local and imported) Widyadana, et al (2013) size, flavor, appearance, color, crispness, aroma, hygiene Moor et al. (2014) flavor, price, appearance, health benefi ts This study s i z e ( w e i g h t ) , f l a v o r, a p p e a r a n c e , c u l t i v a r, texture, aroma, freshness, juiciness.
We include purchasing place in the analysis Various previous studies on hedonic price meassure the attributes directly without explained by the classical economic theory but also affected by its attributes.increases the maximum utility that coould be reached.Therefore, if p(z) convect and regular enough at any places, the higherincome consumers are expected to buy more number of all characteristics (Rosen, 1974).
It is the time that agribusiness actor pays attention to the importance of the attribute values according to consumer's assessment.One of the benefi ts of attributes value is for analyzing hedonic price, an implicit price that is affected by the product's attributes.
Although searching of previous studies has been done, it is possible that the attributes referenced are not relevant to the condition of the investigated products.This research uses factor analyis to the attributes elaborated from previous research and fi eld of studies.
The reduction result from factor analysis is expected to fi nd the important attributes that will be implemented in the next step, hedonic price analysis on local and imported apple.
Statistical analysis that can be used to reduce many attributes is few.The important attributes are factor analysis and principal component analysis (Field & Miles, 2010).Factor analysis has been applied to some previous studies to analyze the quality attributes of fruits and food.Widyadana, et al (2013)  Bandung represents a big city in Western Java.In addition, local apple is delivered to these four cities regularly in large quantities.
Therefore, apple marketing chain can be traced coherently according to the rules of snowball sampling.
In each city, we conducted a consumer survey at three categories of purchasing place: supermarkets, fruit stall, and traditional markets.At each point of analyzed separately.Formulae of a factor is mentioned below (Field & Miles, 2010): 2) extraction to get factors with eigenvalues more than one, and 3) rotation to draw up a factor matrix and variables in the diagonal form.Technically, factor analysis is done by using SPSS software.
In the whole study, all attributes in the significant factors are then used as the attribute on hedonic price analysis.
However, this paper specifi cally discusses the important factors in which apple's quality attributes grouped.Measurement of attributes (except the size and price) is done with Likert scale ranges from 1 to 5 with general details as follows: Score 1: strongly disagree Score 2: disagree purchasing place, the number of samples were predetermined into ten consumers of both local apples and imported apples.
Consumer at each purchasing place was chosen by applying accidental sampling.
The total number of consumers interviewed in each city were 30 for each type of apple consumer.Thus, total number of respondents in the four cities were 120 local apples and 120 imported apple consumers.Those number has over the minimum sampling requirement.Based on the attributes analyzed by factor analysis (9 attributes), the minimum respondent numbers are 90 per apple consumer type.Recommended number of respondents to factor analysis is between 5 to 10 participants per variable up to a maximum of 300 participants, so that the test parameters tend to be stable (Kass andTinsley, 1979 in Field &Miles (2010).
Accidental sampling or also called convenience sampling is a technique in which samples were taken as part of the population that is easily found (Bhattacherjee, 2012).In this study, accidental sampling was used in determining consumer sample who purchased the apple, since there is no information about their population data, it is not possible to do with probability sampling.question with the total value of all questions (Sanusi, 2015).
We also do reliability test.Reliability is a consistency of measurement in refl ecting the measured construct (Field & Miles, 2010).This research uses Cronbach Lower alpha shows that the sample items have a lower ability in capturing the construct measured.In the opposite, higher alpha shows that the k item is closely correlated with the right value (Churchill, 1979).It is generally agreed that the alpha value > 0.7 indicated that the data is reliable.

RESULTS AND DISCUSSION
This section describes the consumers of local and imported apple, reliability and validity of data, the result of factor analysis, and discussion.We can also describe consumers in the various city based on apple cultivars boughed (Figure 1 and Figure 2).We found that there is a different consumers' preference, especially on the apple cultivar.
On local apple, generally, consumers prefer cultivar of Manalagi, except them in Bandung (Figure 1).(Field & Miles, 2010).Higher reliability values are indicated by the alpha approaching one, it is generally agreed that reliability is considered satisfactory if alpha > 0.7 (Churchill, 1979).

The Result of Factor Analysis
Factor analysis was performed to determine the important attributes of apple quality.Some apple attributes that will be used in hedonic price analysis are fi rstly analyzed by factor analysis.The result of Kaiser-Meyer-Olkin (KMO) test showed homogeneity of the variables studied.
The higher the value of KMO means the variables are more homogen.
The KMO values of factor analysis on local and imported apple are 0.740 and 0.691 respectively (Table 6).Those values indicate that the factor analyses have suffi cient sample so that it can proceed.
The range of KMO value is between 0 and 1, with a size limit on 0.5 (Kaiser, 1974in Field & Miles (2010).
Table 7 explains the total variance of the three latent factors that is obtained from the orthogonal rotation (varimax method).Both local and imported apples, each gained three signifi cant factors with the eigenvalue greater than one (Field & Miles, 2010).Percentage of variance shows the ability of a factor as an important factor.
The higher percentage of variance shows the stronger ability.While the cumulative percentage of variance shows the ability of 0,000 0,000 the three factors found.Thus, the ability of the local apple factors to explain the quality attributes is 67.92%, while the same thing on imported apple is 61.31%.
Table 7 shows the loading factors after rotation.Loading factor on each factor is arranged sequentially from the largest to the lowest with a minimum limit of 0.5.place supports the previous study that the discount store or supercenter infl uence the price of tomato (Huang & Lin, 2007).

Loading factor indicates
All attributes of quality factor on both local and imported apples have positive loading factor.This means that all attributes are correlated with the same direction with the latent factors of apple quality.In more detail, local apple quality is getting better if the attributes are increasingly juicy, smooth, fresh, smell nice and sweet.
While imported apple quality is better if the quality attributes are increasingly smooth, crisp, smell nice, juicy and fresh.
The different fi ndings of this study are on apple's juiciness of local apple and appearance of the imported apple.They are signifi cant and rank on the fi rst order of the fi rst factor.Juiciness attribute has not been found in the abovementioned studies (Table 1).Meanwhile, the attribute of appearance in the previous studies was  9).This fact supports the explanation of this analysis result.
This study measures the attribute of appearance from the consumer 's assessment to the apples have just been bought.Moreover, the previous study only used cultivar in the description of consumers' number and the percentage who buy certain cultivars (Widyadana et al, 2013).This fi nding also different with the previous study, in which consumers' perception on the freshness of apple was asked before they purchase apple (Péneau et al, 2006).Meanwhile, this study asked consumers' assessment on freshness after purchasing apple.All attributes of apple quality in this study are measured from the existing condition.Thus, this study has a contribution in giving information about the existing quality attributes base on consumers' assessment.with the lowest loading factor.This fi nding is not in line with the previous studies that fl avor is ranked the highest of ten attributes of apple quality (Jones, 1996).This result is also somewhat different from Widyadana et al (2013)  Using size of apples with the average weight per fruit differs from previous studies that used diameter [Manalo (1990); (Jones, 1996)) ;Carew, (2000); Péneau, et al (2006); Harker, et al (2008) place.However, the data shows that in general, consumers assess apples sold in a supermarket with a higher score than those in fruit stall and traditional market (Figure 3).Thus, the presence of texture is closely associated with purchasing place, so that they are located at the same latent factor.
Consumers in supermarket give the highest score for all quality attributes on local apple (Figure 3).

CONCLUSION AND SUGGESTION
The important quality attributes of

Figure 1
Figure 1 describes consumers' equilibrium on hedonic price.The consumer's willingness to pay for z at the utility index and fi xed income is θ(z; u, y) while p(z) is a minimum price that should be paid in the market.The utility will be maximum when θ (z* ; u* , y) = p(z*) and θ zi (z* ; u* , y) = p i (z*), i = 1, … , n, where z* and u* are optimum.An optimum location on z happens when p(z) and θ (z* ; u* , y) are the connected tangent of both variables.One dimension of consumer's equilibrium is illustrated on fi gure 1.One family of indifference curve, where there is only one member (on u*), is defi ned as θ (z 1 , z* 2 , … , z* n ; u, y).Two different buyers are shown in the fi gure, i.e. θ 1 and θ 2 .The last buyer buys one merk offering more z 1 .The increasing income always Important attributes on local and imported apples are found by exploratory factor analysis.The two category of apple origin, namely local and imported scale needs validity test of the research instrument.Validity is the consistency of the measurement result.However, the only consistency is not enough (Churchill, 1979); the valid instrument is able to measure what is being measured.Pearson correlation could be used to test the validity of the instrument, by using the correlation coefficient between each value on one

α
to test the data reliability.The Cronbach α value is counted with formulae (2)................. (2)Source:Field & Miles (2010) the strength of the relationship between the attributes with the factors.The attributes in Factor 1 indicate the quality attributes of apple.Meanwhile, attributes of factor 2 show cultivars and size of apples.Lastly, factor 3 indicates purchasing place of apples.For the purpose of seeking hedonic price, especially, all attributes are raised in the analysis, not merged into a single factor.However, in this paper, there is a discussion of the factors that contain a group of attributes.The result of factor analysis suggests that there is a little difference in composition between the attributes of local and imported apple.On local apples, fi rst latent factor includes all quality attributes of apple, second latent factor includes attributes of cultivars and size, and the third latent factor includes the purchasing place and attribute of texture.On imported apple, the first latent factor also includes quality attribute, second latent factor includes cultivar and size, while the third latent factor includes the only purchasing place.However, the overall results indicate that the attributes tested are recommended to be analyzed further on hedonic price model.In particular, this paper discusses more detail about the results of the factor analysis.The fi rst latent factor on local apple is closely related positively to the quality attribute of apple, namely juiciness, appearance, freshness, aroma, and taste.Among the fi ve quality attributes, juiciness has the greatest loading factor, which means that this attribute has the strongest correlation with the latent factors.Nevertheless, the loading factors on all attributes of factor one are only a little bit differ, and all of them are positive.
If we compare the second latent factor between local and imported apples, there are some similarities.Firstly, based on loading factor, the attributes order on local apple are RB, Manalagi and size; whereas on the imported one are Fuji, Washington, sizes and flavor.There is a similarity between local and imported apples in terms of cultivar and size.However, fl avor attribute on imported apples entered in the last sequence of the second latent factor.Second, there is a similarity on negative loading factor, i.e. on the cultivar of Rome Beauty and Washington.This means that the correlation between those cultivars with the second latent factor is strong enough but has the opposite relationship.The last, the third latent factor is compared.The order of attributes based on loading factor on local apple is a fruit stall, supermarket and texture attribute, whereas the order on imported apples are fruit stall and then the supermarket.Loading factors on the fruit stall are negative, both on local and imported apples.It means that the correlation between fruit stall with the third latent factor is strong enough but it has the opposite relationship.DiscussionThe results show that the important factors in the purchasing apples are the attributes of quality, cultivar and purchasing place.There are some similarities between local and imported apples.These results generally support previous studies published in some journals that the quality attributes of apple affect consumer preferences {Manalo (1990);Péneau, et al (2006);Harker, et al (2008), andWidyadana, et al (2013).Furthermore, apple cultivars also affect consumer preferences, in accordance with previous studies {Péneau, et al (2006); Carew, (2000) dan Skreli & Imami, 2012b)}.Likewise, a signifi cant purchasing measured from the consumer's preference when they are choosing an apple {Péneau, et al (2006); Widyadana, et al (2013); and Moor et al. (2014)}.In addition, consumers assess imported apple better than the local one, where the biggest gap happens on the attributes of juiciness and appearance (Table who obtained three top attributes of size, color, and fl avor with almost the same consumer ratings.This research found that fl avor attribute is important in determining quality attributes.However, it is not in the fi rst order of the consumers' preference.A further analysis is needed to proof this fi nding, namely to analyze the infl uence on the apple price and consumers demand on apple.In the factor of cultivar, there are attributes with negative loading factor, i.e.Manalagi cultivar on local apple and Washington cultivar on imported apple.It means that those cultivars have a negative correlation with the latent factors.The cultivar attribute is a dummy variable, namely the cultivars purchased when the survey was conducted.The negative loading shows a negative correlation between the cultivars with the latent factor.Size is grouped on the factor of cultivars factors, both at local and imported apples.This means that size of apple closely correlated with factors of cultivars.The explanation can be given is that on local apple, the cultivar of Rome Beauty has larger average size than Manalagi and Anna.The fi eld data shows the average weight (ounce per fruit) of Manalagi, RB and Anna are 1.38, 1.56 and 1.14, respectively.While on imported apples, Fuji cultivar also has larger average size than Washington cultivar.Field data shows the average weight (ounce per fruit) of Fuji, Washington and others are 2.05, 1.67, and 1.75 respectively.
; and Skreli & Imami (2012)].Measuring apple by using weight per item is technically easier to do.It is done by dividing the weight of apples purchased by the number of apples.Local apple traders also perform grading by counting the number of apple fruit per Kg, it is ranging from 5-6 fruits, 7-8 fruits, 9-10 fruits, 11-12 fruits, and so on.Using weight in measuring apple in this research is closer to the practical way done by the trader in Malang Raya who grade apple by using the number of fruit per kilogram.Thus, this research gives a contribution in measuring technique that more adapted in Indonesian situation.On imported apples, there are different results with the local one, namely attribute of fl avor is grouped on the factor of cultivar.On average, consumers give a greater score to Fuji and other cultivars compared with Washington.Thus fl avor attribute is closely attached with the attribute of cultivar so that they presence in the same factor.The factor of purchasing place has similarities between local and imported apples, namely fruit stall has negative loading factor.It means that fruit stall has a negative correlation with the factor of purchasing place.Generally, the fruit stalls sell local and imported apples with the quality attributes assessed lower than supermarkets but higher than traditional markets (Figure 3 and 4).Since the comparison is dummy variable of the supermarket, so that fruit stall has a lower score, thus fruit stall has a negative correlation with the factor of purchasing place.Texture attributes of local apple are grouped in the factor of purchasing place, with a positive sign.It is hard to explain the relation of texture and purchasing local and imported apples are grouped in each two latent factors are quality and cultivar and purchasing place.All quality attributes have a positive correlation with the apple quality, but there is a different fi nding.Flavor is not the most important attribute that ranked fi rst based on consumers' assessment.The factor of apple cultivar always associates with the apple size, both on local and imported apples.Rome Beauty on local apple and Fuji on imported apple have greater average size than other cultivars, so these attributes are already attached one each other.The factor of purchasing place becomes an important part which accompanies the apple's quality attributes and is considered by consumers in purchasing apples.The importance of this factor is related to the selection of apple quality by retailers, as well as comfort and service given, will also infl uence consumer assessment on the quality attributes of apple.

Figure 3 .Figure 4 .
Figure 3.The Average Score of Consumers' Assessment of The Attributes of Local Apples at Three Categories of Purchasing Place

Table 2 .
The Attributes of Local and Imported Apple Juiciness Very small up to very much Consumers' assessment on juiciness Freshness Very unfresh up to very fresh Consumers' assessment on freshness Texture Very hard up to very crispy Consumers' assessment on texture Size (weight)* in the scale of ounce Apple's weight per fruit (ounce) Appearance Very un-smooth up to very smooth Consumers' assessment on appearance *Apple size is measured in weight per fruit, it is counted from total weight of apple purchased divided by its number.etal., 2013).In the imported apple, the proportion of consumer is not related to the production number, however, it is

Table 3 .
The Number of Consumers Who Buy Apple Cultivars In Various Categories of Purchasing Place * Percentage is counted for each category of local and imported apple Figure 2a.Local apple cultivar based on city Figure 2b.Impoted applecultivar based

Table 4 .
Pearson Correlation For Validity Test on Apple Attributes

Table 5 .
Reliabilities of Local and Imported Apple Consumers

Table 6 .
The Result ofKMO and Barlett Test

Table 7 .
Total Variance Explained Meanwhile, the last attribute of local apple is fl avor, whereas the last order on imported apples is freshness.

Table 8 .
Rotated Component Matrix of Local and Imported Apple

Table 9 .
Consumers' Assessment on The Apple's Attribbutes All attributes are shown in the avarage skore of Likert scale, except size* that is measured with average weight of apple (sounce).