HOME

This study relies on the use of copy testing. Copy testing is an advertising research tool used to evaluate and understand consumer’s attitude towards specific advertisements.   This report contains copy testing for three brands of protein bars, Atkins, Myoplex and Snickers.

The copy testing was conducted through an online survey, where consumers were asked to rate three brands of protein bars.  Then they were exposed to corresponding advertisements, which were followed up by further questions used to determine the effectiveness of the ads.

The purpose of this study is to examine brand preference among the three types of protein bars as it corresponds to print advertisement.   The online survey was used to determine which brand consumers felt the most favor to and to what extent consumers brand attitude was swayed by advertisements.

The final report contains methodology of the online survey, analysis, final conclusions as well as a sample survey.  

 

HOME

BASIC STATISTICS REGRESSION ANALYSIS DISCRIMINANT ANALYSIS FACTOR ANALYSIS

 

BASIC STATISTICS

1. Paired t-tests for Brand Index Score

Table 1.1

sample size = 62

The Brand Index score for each brand was calculated by summing ten Likert items pertaining to each brand.  The highest Brand Index Score possible is 50, while the lowest is 10.  Each of the three brands tested have a Brand Index Score in the middle, meaning respondents were neutral to each brand.

Table 1.2

Paired Samples	t-ratio
Atkins-Myoplex	3.22*
Myoplex-Snickers	.564
Snickers-Atkins	3.28*

*p<=.15

The paired samples of Atkins and Myoplex and Snickers and Atkins are statistically significant, while the paired sample of Myoplex and Snickers is not.  This means that in 85 or more samples drawn from the same population as this sample of 62 people, it would be expected that the difference between the brand index mean scores for Atkins and Myoplex, and Snickers and Atkins would be about the same as they are in this sample.

 

2. Between-Group t-tests

Table 2.1

Change Score	Number of Respondents	Brand Index Score Mean	Standard Deviation	t-ratio
Up Movers	29	27.4	3.8	.72
Down Movers	29	26.7	4.6

 

Table 2.2

Change Score	Number of Respondents	Ad Index *Score Mean	Standard Deviation	t-ratio
Up Movers	29	7.6	2.5	.89
Down Movers	29	7.1	2

*Ad Index Score was calculated by summing the positive responses to section 7 and 8 of the brand preference survey.

There are 29 individuals, 48.3% of the sample, whose attitude toward Atkins became more favorable following exposure to its advertisement.   These individuals had a mean Brand Index Score of 27.4, and a mean Ad Index Score of 7.6.  The results found in this sample are not statistically significant, meaning they could not be applied to a larger population.
At the same time 29 individuals’, 48.3% of the sample, attitude toward Atkins became less favorable after exposure to its advertisement. These individuals had a mean Brand Index Score of 26.7, and a mean Ad Index Score of 7.1.  The results found in this sample are not statistically significant, meaning they could not be applied to a larger population.

 

3. Chi Squared Significance Test

Atkins Change Score		Brand Index Score Above Median	Brand Index Score Below Median
Up Movers	Count	15	17
	Row %	46.4	53.6
	Column %	48.1	51.7
	Total %	23.2	26.8
Down Movers	Count	13	13
	Row %	50.0	50.0
	Column %	48.1	44.8
	Total %	23.2	23.2
Same	Count	1	1
	Row %	50.0	50.0
	Column %	3.7	3.4
	Total %	1.8	3.4

Chi-Squared=.072 *, Total Cell Count: 62
*p<=.15

The median Brand Index Score for Atkins is 28.0.  Out of the 62 respondents 50% felt more favorably to Akins after viewing the advertisement, while 3.6% stayed the same, and 46.4% felt less favorable.  Of the up movers 23.3% had a higher Brand Index Score than the mean for the sample, while 26.8% had a lower one.  The down movers were spilt in comparison to the median Brand Index Score.

The Chi-Squared test is .072 meaning it is not statistically significant, and cannot be carried on to other samples.

4. Post vs. Pre Test Frequencies

Change Score	Atkins	Myoplex	Snickers
Up Movers	29	27	28
Same	6	4	11
Down Movers	29	29	23

There are 29 individuals, 48.3% of the sample, whose attitude toward Atkins became more favorable following exposure to its advertisement.   While, there were 48.3 % who felt less favorable.

There are 27 individuals, 43.5% of the sample, whose attitude toward Myoplex became more favorable following exposure to its advertisement.   While, there were 46.8 % who felt less favorable.

There are 28 individuals, 45.2% of the sample, whose attitude toward Snickers became more favorable following exposure to its advertisement.   While, there were 37.1 % who felt less favorable.

Any increase in favorability for each of the three brands was generally offset by an approximately equal group of respondents who found the brand less favorable.  The effectiveness of the brand is inconclusive using pre vs. post testing numbers.

5. Frequency Counting Comparing Atkins and Myoplex

Out of the 62 respondents, 25 individuals or 40.3% rated Atkins as having a higher Brand Index Score than that of Myoplex.  While 37 individuals or 59.7%, rated Myoplex as having a higher Brand Index Score than Atkins.  This show that more respondents had a more favorable attitude toward Myoplex than Atkins.

6. Simple Correlation Coefficient

A simple correlation was calculated, comparing the Brand Index Score of Atkins to Myoplex.  The coefficient correlation is .422, which means the Brand Index Score of Atkins and Myoplex are moderately correlated.  This relationship is also statistically significant at an alpha level of .15.  These means in 85 or more samples out of every 100 samples drawn from the same population as this sample of 62 people, it is expected that the correlation between the Brand Index Score  for Atkins and Myoplex would be about the same as found in this sample. Therefore, we can project the results of the survey to the population. 

7. Mean Comparison

Martial Status	Mean Brand Index Score of Atkins
Single	28.5
Married	26.9
Divorced	21.8

Those respondents who said they were single had a mean Brand Index Score of 28.5 in regards to the Atkins brand.  Married respondents had a mean of 26.9, while divorced people reported a mean of 21.8.  While the differences in means is negligible single people seem to respond best to the Atkins advertisement.

TOP

 

REGRESSION ANALYSIS

This focuses on a regression analysis of the copy-testing results of three print brands for three different protein bars.  The analysis was conducted using the pre-post ad exposure change score as the dependent variable, while the independent variables are the ten individuals Likert item responses.  The sample size is 62.

	Multiple Regression (R)	Coefficient of Multiple Determinant (R2)	Standard Error	F-ratio
Atkins	.4	12.4%	3.4	.70
Myoplex	.5	21%	3.2	1.27
Snickers	.3	11%	4.05	.58

 

	Atkins				Myoplex				Snickers
Likert Item (Brand Attitude)	b	beta	t-ratio		b	beta	t-ratio		b	beta	t-ratio
Constant	0.6		0.16		2.0		0.79		1.1		0.29
It is a good protein bar.	-0.3	-0.1	-0.85		-0.7	-0.3	-1.90		-0.6	-0.2	-1.47
It gives me energy.	-0.1	0.0	-.017		0.2	0.1	0.37		-0.01	0.0	-0.11
It is a filling snack.	0.1	0.0	.030		-0.5	-0.2	-1.14		0.2	0.1	0.44
I prefer it to other protein bars.	-0.1	0.0	-0.14		0.7	0.3	2.01		0-.4	-0.1	-0.81
It helps me stay physically fit.	0.6	0.3	1.32		-0.2	-0.1	-0.57		0.0	0.0	0.05
It helps me lose weight.	0.1	0.0	0.16		.05	0.2	1.09		-0.1	0.0	-0.23
I do not like the taste of it	-0.4	-0.2	-0.85		-0.1	0.0	-0.25		0.3	0.1	0.67
I would purchase it	0.6	0.3	1.40		-0.1	-0.1	-0.43		-0.3	-0.1	-0.55
I believe athletes eat it.	-0.7	-0.3	-1.50		-0.2	-0.1	-0.56		0.0	0.0	-0.09
It tastes good.	-0.4	-0.2	-0.98		-0.2	-0.1	-0.67		0.4	0.1	0.98

Beta’s considered important => the highest beta/2

Atkins Regression

The multiple regression analysis for Atkins shows a low coefficient of multiple determination or R2.  This means there is almost no relationship between how respondents rated the Atkins brand based on the ten-likert items and their brand attitude as indicated by the post-pre change score.  Moreover, the standard error is 3.4, which means respondents were on average 3.4 units up or down from the regression line.  This data is not significant, and therefore cannot be scaled to a larger population.

The betas that are considered important to this analysis are those related to five Likert items: It helps me stay fit, I do not like the taste of it, I would purchase it, I believe athletes eat it and it tastes good.

The un-standardized coefficients (b) show the relationship between the brand attitudes as understood by the Likert items, and how respondents favored the ads.  They can be used to determine the equation of the regression line.  The equation for Atkins regression line is: pre-post ad exposure change score = 0.6 - 0.3(its good) - 0.1(energy) + 0.1(filling) - 0.1(prefer) + 0.6(healthy) + 0.1(weight loss) – 0.4(bad taste) + 0.6(purchase) – 0.7(athletes) - 0.4(taste good).

Myoplex Regression

The multiple regression analysis for Myoplex shows a coefficient of multiple determination of 21%.  This indicates that there is a small relationship between how respondents rated the Myoplex brand based on the ten Likert items and their brand attitude as understood by the post-pre change score.  Furthermore, the standard error is 3.2, which means respondents vary from 3.2 above the regression line to 3.2 below. Moreover, the data is not statistically significant, therefore it cannot be applied to a larger population.

The betas that are considered important are: It’s a good protein bar, it is a filling snack, I prefer it to other protein bars, and it helps me lose weight.

Snickers Regression

The multiple regression analysis for Snickers shows a coefficient of multiple determination of 11%.  This indicates a very weak relationship between the Likert items and the change score. Likewise, the snickers standard error is the highest of the three equaling 4.05.  This means the actual results vary quite a bit from the regression line. Moreover, the data is not statistically significant, therefore it cannot be applied to a larger population.

The betas that are considered important are: It’s a good protein bar, it is a filling snack, and I prefer it to other protein bars.

Overall the analysis reflects poor relationships.  The ten Likert items are poor predictors of the post-pre change exposure score. 

TOP

 

DISCRIMINANT  ANALYSIS

The discriminant analysis compares the ten Likert items and the pre-post change score for the Atkins brand.  The independent variables are the ten Likert items, and the dependent variable is the change score. 

Table 1: Discriminant Analysis

Likert Items	Up movers		Down Movers		Standardized Discriminant Coefficient	Un standardized Discriminant Coefficient
	Mean	Standard Deviation	Mean	Standard Deviation
It is a good protein bar.	2.4	1.5	1.0	0.0	-.9	-.6
It gives me energy.	1.8	1.3	1.5	.7	-.4	-.3
It is a filling snack.	3.0	1.3	3.5	.7	-.4	-.3
I prefer it to other protein bars.	3.3	1.1	3.5	.7	-.1	-.1
It helps me stay physically fit.	2.8	1.8	3.0	2.8	1.0	.5
It helps me lose weight.	2.8	1.1	2.5	.7	-.3	-.3
I do not like the taste of it	2.4	1.2	2.0	1.4	0.0	0.0
I would purchase it	3.1	1.6	3.5	2.1	1.0	.7
I believe athletes eat it.	2.6	1.5	3.0	1.4	.1	.1
It tastes good.	3.2	1.4	4.0	1.4	.1	0.0

 

Table 1 shows that the mean Likert score for the Up Movers as compared to the Down Movers is similar.  The highest mean for the Up Movers is 3.3, while the highest for the Down Movers is 4.0.  This helps further illustrate the parity between the two groups.  However the standard deviation for the Down Movers appears to be smaller than that of the Up Movers.  This means there is less variability among the Down Movers. 
According the standard discriminant coefficient there are only three variables that are considered important.  They are: Atkins is a good protein bar, it helps me stay physically fit, and I would purchase it.  From a logical standpoint, the importance of these variables makes sense, since taste, and how healthy the protein bar is ultimately effect the decision to buy.

 

Table 2: Centroid Analysis

Group	Score
Up Movers	-.1
Down Movers	1.2

Table 3: Wilks Lambda

Wilks’ Lambda	Chi-Square	Degrees of Freedom
.91	2.26	10

As indicated by Table 2, the group centroids of the up movers and down are not close to one another.  This contributes to a Wilks’ Lambda of .91 and Chi-Square of 2.26.  Both of these metrics are not statistically significant at the alpha level of .15; thus they cannot be applied to a larger sample.

 

Table 4: Confusion Matrix

	Predicted up movers	Predicted down movers
Actual Up Movers	29	0
Actual Down Movers	2	0

T = (.935 – 0.5) / sqroot(((.935*.065) + (.5*.5))/ 31) = 4.34

As indicated by Table 4, 93.5% of actual respondents were classified correctly.   At an alpha of .15 this data is statistically significant, therefore these findings can be applied to 85 or more samples drawn from the same population as this sample.

TOP

 

FACTOR ANALYSIS

The analysis provided below consists of a factor analysis, where the relationship for ten Likert items is examined.  This analysis is conducted for all three brands of protein bars, Atkins, Myoplex, and Snickers.  The sample size for this analysis consists of 60 respondents.

Table 1:  Total Variance Explained

Factor	Atkins			Myoplex			Snickers
	Eigenvalues	Percent variance	Cumulative variance	Eigenvalues	Percent variance	Cumulative variance	Eigenvalues	Percent variance	Cumulative variance
I	3.2	32.2	32.2	2.3	22.6	22.6	1.9	18.7	18.7
II	1.6	17.0	49.1	1.6	16.0	38.6	1.6	16.1	34.8
III	1.2	12.1	61.2	1.4	13.5	52.2	1.4	13.6	48.4
IV	1.1	10.6	71.8	1.1	11.1	63.3	1.3	12.9	61.3
V	.8	7.9	79.7	1.1	10.7	74.0	0.9	9.5	70.8
VI	.6	6.4	86.1	0.8	7.9	81.9	0.8	8.3	79.2
VII	.5	5.1	91.3	0.6	6.2	88.2	0.7	6.6	85.8
VIII	.4	3.8	95.1	0.5	5.3	93.5	0.6	5.5	91.3
IX.	.3	3.0	98.1	0.4	3.8	97.3	0.5	4.6	95.9
X	.2	1.9	100	0.3	2.7	100.0	0.4	4.1	100.0

 

 

 

 

 

 

 

 

Four of ten factors for Atkins and Snickers have EigenValues above 1.  This means that they explain the variance of a Likert item.   For example factor III of Atkins explains about a 60% variance for the ten Likert items, leaving approximately 40% unexplained.   While Myoplex has five factors that help explain the Likert items.  Factor III of Myoplex has a cumulative variance of about 52%, meaning 48% of variance was left unexplained.      

 

Table 2:  Factor Varimax Rotated Matrix (Atkins)

Likert	Communalities	Factor matrix				Varimax rotated matrix
		I	II	III	IV	I	II		III	IV
It is a good protein bar.	0.7	-0.3	-0.5	0	0.6	0	0		-0.1	0.8
It gives me energy.	0.7	-0.1	0.7	0.2	0.4	0.1	0.7		0.3	0
It is a filling snack.	0.5	0.5	0.3	0.1	-0.4	0.3	0		0.1	-0.6
I prefer it to other protein bars.	0.7	-0.3	-0.5	0.6	0	0.1	0.1		-0.7	0.3
It helps me stay physically fit.	0.8	0.8	-0.4	-0.2	0.1	0.6	-0.6	0.2		0
It helps me lose weight.	0.7	0.5	0.4	-0.3	0.5	0.3	0		0.8	0.1
I do not like the taste of it	0.6	-0.7	0.2	-0.2	0	-0.7	0.2		0	0.1
I would purchase it	0.8	-0.5	0.5	0.5	0.1	-0.1	0.8		-0.2	0
I believe athletes eat it.	0.7	0.7	0	0.4	0.2	0.8	0		0.1	-0.1
It tastes good.	0.6	0.7	0.1	0.4	0	0.7	0		0.1	-0.2

For the above table a Factor Matrix was used for further analysis because it has the least number of ambiguous rows.  Contributing factors have values higher or equal to the absolute value of 0.5.  The above data is heavily dispersed, but the majority of the Likert items are loaded in factor 1.

Table 3:  Factor Varimax Rotated Matrix (Myoplex)

Likert	Communalities	Factor matrix					Varimax rotated matrix
		I	II	III	IV	V	I	II	III	IV	V
It is a good protein bar.	0.8	-0.6	-0.1	0.2	0.3	0.3	-0.3	0.3	0	0.6	0
It gives me energy.	0.7	-0.4	0.4	0.5	-0.4	0.4	0.1	0.9	0.2	0.1	0.1
It is a filling snack.	0.5	0.7	0.2	0.2	-0.3	0.1	0.7	0	-0.1	-0.4	0.2
I prefer it to other protein bars.	0.2	-0.2	-0.3	0.2	0.6	0.4	0.1	0	0	0.8	0
It helps me stay physically fit.	0.8	0.4	-0.2	0.7	0	-0.2	0.5	0.1	-0.7	0	-0.1
It helps me lose weight.	0.8	0.6	0.5	0.3	0.3	0.2	0.8	-0.1	0.1	0	-0.3
I do not like the taste of it	0.6	-0.1	0.8	-0.2	0.1	0.3	0.3	0.2	0.8	0	-0.1
I would purchase it	0.2	0	-0.4	-0.3	-0.5	0.6	-0.1	0	0	0	0.9
I believe athletes eat it.	0.3	0.6	0.1	-0.5	0.2	0.2	0.3	-0.6	0.3	-0.1	0.2
It tastes good.	0.5	0.6	-0.5	0.1	0.1	0.3	0.4	-0.5	-0.4	0.2	0.4

The above table depicts both the factor matrix and varimax rotated matrix for Myoplex.  It shows five factors.  For further analysis the factor matrix was used.  This is due to the fact that it has the most Likert items loaded into factor 1.  However similarly to Atkins the data is fairly spread out.

Table 3:  Factor Varimax Rotated Matrix (Snickers)

	Communalities	Factor matrix				Varimax rotated matrix
		I	II	III	IV	I	II	III	IV
It is a good protein bar.	0.8	0	0.4	-0.1	-0.7	-0.3	0	0.1	-0.7
It gives me energy.	0.9	0.6	-0.4	-0.2	-0.2	0.8	0.1	-0.2	-0.2
It is a filling snack.	0.5	0.4	-0.6	0	0.3	0.7	0	-0.1	0.3
I prefer it to other protein bars.	0.6	-0.5	-0.4	0.4	-0.2	0	-0.4	0.7	0.1
It helps me stay physically fit.	0.8	-0.2	0.7	-0.3	0.3	-0.7	0	-0.5	0
It helps me lose weight.	0.5	-0.6	-0.2	-0.5	0	-0.1	-0.8	-0.2	0.1
I do not like the taste of it	0.4	-0.3	-0.1	0.3	0.6	-0.2	0	0.1	0.7
I would purchase it	0.5	0.6	0.2	0	-0.1	0.2	0.5	-0.3	-0.3
I believe athletes eat it.	0.5	0.4	0.4	0.4	0.3	-0.1	0.7	-0.1	0.2
It tastes good.	0.3	-0.2	0.1	0.7	-0.3	-0.2	0.2	0.8	0

The factor analysis for Snickers yielded less number of Likert items that are loaded into a particular factor.  However, much like Atkins and Myoplex there are several factors causing the data to be spread out.  Nonetheless, the factor matrix was used for further comparison because it has the least number of ambiguous rows.

Table 5 and 6:  Attitude and Paired T-test

	Mean	N	Std. Deviation
ATKINS	2.7	60.0	0.4
SNICKERS	2.5	60.0	0.4
MYOPLEX	2.5	60.0	0.5

	MEAN DIFFERENCE	STANDARD	T
ATKINS - MYOPLEX	0.2	0.5	3.1*
ATKINS - SNICKERS	0.2	0.6	2.3*
MYOPLEX - SNICKERS	0.0	0.5	-0.4

*p < .15

Taking the average score of those Likert items that had a factor score of 0.5 or higher derived the above tables.  The means for all the three brands are very close.  Atkins had a mean brand score of 2.7, while both Snickers and Myoplex had brand scores of 2.5.  Though Atkins had the highest score it was by a very narrow margin.  However, the paired t-tests comparing Atkins with Myoplex and Atkins with Snickers are statistically significant at an alpha less than or equal to .15.  Thus it can be concluded that in 85 or more samples out of every 100 samples taken from the same population as these 60 respondents, we would expect to find a difference of mean attitude scores of the same magnitude for each pair as shown in this sample.  

TOP   

 

 

HOME

BASIC STATISTICS REGRESSION ANALYSIS DISCRIMINANT ANALYSIS FACTOR ANALYSIS

 

 

 

 

BASIC STATISTICS

Paired t-test
Each brand share a similar brand index score, with Atkins being the highest by a slim margin.  The paired t-test showed that the paired samples  of Atkins and Myoplex, and Snickers and Myoplex are statistically significant.

Between-Group t-tests
This test showed that there was no relationship between the exposure of the ad and the attitude of a brand.  Respondents who moved up or down post-ad had a very similar brand index score.

Chi Square
The data collected is not statistically significant.  However it did show that out of the 62 respondents 50% felt more favorably to Akins after viewing the advertisement, while 3.6% stayed the same, and 46.4% felt less favorable. 

Frequency
For all three brands there were about the same number of up movers as down movers.  These numbers do not give any clear indication of the effectiveness of the ad.

Correlation
There is a weak correlation between the brand index scores of Atkins and Myoplex.  It is also statistically significant at an alpha of .15.

Mean Comparison
Those respondents who are single on average had a higher brand index score for Atkins.   This may be due to the name recognition Atkins has as a healthy alternative.

 

 

REGRESSION ANALYSIS

Atkins Regression
This analysis shows that there is almost no relationship between how respondents rated the Atkins brand based on the ten-Likert items, and their brand attitude score as indicated by the post-pre change score.  The five Likert items considered important are: “It helps me stay fit, I do not like the taste of it, I would purchase it, I believe athletes eat it, and it tastes good.”

Myoplex Regression
Myoplex has a small relationship between how respondents rated the Myoplex brand based on the ten Likert items and their brand attitude as understood by the post-pre change score.  The important Likert items are:  It’s a good protein bar, it is a filling snack, I prefer it to other protein bars, and it helps me lose weight.

Snickers
This analysis indicates a very weak relationship between the Likert items and the change score.  The most important Likert items as indicated by beta are: It’s a good protein bar, it is a filling snack, and I prefer it to other protein bars.

TOP

 

DISCRIMINANT ANALYSIS

The important Discrimant variables used to categorize up movers and down movers are: Atkins is a good protein bar, it helps me stay physically fit, and I would purchase it.  The more respondents said Atkins was a good protein, the less they liked the ad.  The more respondent that said Atkins helps them stay physically fit and they would purchase it, the higher they rated the ad.  Furthermore, this model did a very good job of correctly classifying respondents.  93.5% of actual respondents were classified correctly, which means this data is statistically significant and can be projected to a larger sample.

TOP

 

 

FACTOR ANALYSIS

Atkins
Four of the ten factors created for Atkins have Eigenvaluse greater than one.  Together these four factors explain about 74% of the variance of at least one Likert item.  Eight variables loaded positively into the first factor.  These are: it is a filling snack, I prefer it to other protein bars, it helps me stay physically fit, it helps me lose weight, I don’t like the taste of it, I would purchase it, I believe athletes eat it and it tastes good.

Myoplex
Five of the ten factors created for Atkins have Eigenvaluse greater than one.  Together these five factors explain about 71% of the variance of at least one Likert item.  Six variables loaded positively into the first factor.  These are: it is a filling snack, it is good, it helps me lose weight, I believe athletes eat it and it tastes good.

Snickers
Five of the ten factors created for Atkins have Eigenvaluse greater than one.  Together these five factors explain about 67% of the variance of at least one Likert item.  Four variables loaded positively into the first factor.  These are: it gives me energy, I prefer it to other protein bars, it helps me lose weight, and I would purchase it.

TOP

 

 

HOME

The purpose of this study is to evaluate the brand preference of three different brands of protein bars, as well as the effectiveness of print ads on these preferences.  To achieve this purpose copy testing was used.  Variables were gathered that would help to explain why respondents felt a certain way.  An initial brand score was compared to a brand score collected after respondents were shown the ads.  This comparison is the heart of the analysis. 

This study collected 60 online surveys through the use of snowball sampling.  The data collected was run through SPSS to further analyze the information.  Models such as t-tests, chi-squared, factor analysis and discriminant analysis were relied on. 

The study yielded no real conclusive evidence.  Instead it showed that respondents had about the same brand attitude towards each brand both before and after the ads.  No brand was stronger than the other.  Because of this, no real conclusion can be drawn about the effectiveness of the ads.

 

 

HOME

Please follow this like to view the online survey complete with the results.

To view the original survey please continue to Appendix B.

 

HOME

View Original Survey

HOME