R Markdown Kente Project
Albert Mensah Kusi
9/11/2022
Project Topic: Assessing Public (Users and Non-Users) Values on Conservation of Kente Weaving and Interpretation of Symbols in Ghana
Introduction
Kente is a traditional hand-loom-woven designer cloth that is only found in Ghana and closely linked with royalty. Despite this association with royalty, kente is now used by all social classes. Kente is used for both its beauty and symbolic significance. There are more than 300 different kente designs and each has its name and meaning. The names and meanings are derived from historical events, individual achievements, proverbs, philosophical concepts, oral literature, moral values and codes of conduct, among others.
The disappearance of intangible cultural heritage together with associated symbols and meanings in sub-Saharan Africa undercuts 2003 UN Convention for Safeguarding of the Intangible Cultural Heritage. To address this worrying pattern, this research project evaluate value of conserving traditional kente weaving and interpretation of kente symbols as an intangible cultural heritage by establishing national demonstration centers in Ghana.
First, this project examine statistical relationship between individuals who have kente cloth and their place of residence. Secondly, this study assess whether the population distributions are identical for perceived knowledge on Kente cloth and Kente weaving. Moreover, it estimate the differences among means of willingness to pay to conserve Kente and its factors Furthermore, this project evaluate the relationship between willingness to pay to conserve Kente and its demographic variables. Also, assessing the likely factors that affect respondents willingness to pay to conserve Kente and year of birth.
Materials and Methods
Beneath are the study areas and the sampling methods. Also, methods of data analysis and variables definition
Study Areas and Sampling Methods
The data used in this study were obtained from surveys in which willingness-to-pay questions questions were posed to samples drawn from the public. The WTP CV surveys were conducted in Bonwire and Kumasi in Ashanti Region, Accra in Greater Accra Region and Ho and Agotime Kpetoe of the Volta Region of Ghana. Bonwire and Agotime-Kpetoe were selected since both towns are associated with kente weaving and we would like to test whether WTP values elicited from these towns are different from those elicited from other parts of Ghana. The study interviewed about 50 respondents from each of these two small, kente weaving towns, and 200 respondents from each of the nearby cities of Ho and Kumasi. Moreover, in order to test for distance decay in WTP, we interviewed 200 respondents from Accra. In total, we had an overall sample of about 722 respondents in the survey.
The sample consists of users and non-users of kente cloth. For the first step, the metropolises were purposively sampled from each of the three regions and the three metropolises are Accra, Kumasi, and Ho. In the second step, convenience samples in suburbs in these metropolitan areas were selected to represent low-, middle-, and high income areas.
Method of Data Analysis
- Means, Median, standard deviation, minimum and maximum.
- Pearson’s Chi-squared Test for Count Data.
- Wilcoxon Rank Sum and Signed Rank Tests.
- ANOVA & MANOVA - (Multivariate) ANalysis of VAriance.
- Boxplot.
- Multi scatterplot in groups.
- Table of means.
- Tukey multiple comparisons of means.
- MANOVA (multivariate analysis of variance).
- Multivariate multiple linear regression.
Variables Definition
- WTP_Va:Willingness to pay values to establish national centers across the country to preserve Kente weaving and interpretation of its symbols
- Household_size: number of people above 18 years living in the household
- Place:study areas respondents interviewed
- Year_birth:year respondents were born
- have_kente:respondents having Kente cloth(or a clothing that is completely made of Kente)
- Gender: male and female respondents
- Knowledge_Cloth:How respondents are knowledgeable of Kente cloth in general
- Knowledge_Weaving:How respondents are knowledgeable of Kente weaving
- Knowledge_symbol:How respondents are knowledgeable of Kente symbols
Results and Discussion
# Set working directory:
setwd('C:/Users/Lenovo-PC/Desktop/sch appy/Marine/interview/proposal/Course Materials/Econometric Modelling 2/Survey_Data_Analysis')
getwd()## [1] "C:/Users/Lenovo-PC/Desktop/sch appy/Marine/interview/proposal/Course Materials/Econometric Modelling 2/Survey_Data_Analysis"# read datasets
data <- read.csv('DATASETS/Kente_public18.csv')- Descriptive statistics of some of the variables studied
summary(data)## respno place bonwire kpetoe
## Min. : 1.0 Length:722 Min. :0.00000 Min. :0.00000
## 1st Qu.:181.2 Class :character 1st Qu.:0.00000 1st Qu.:0.00000
## Median :361.5 Mode :character Median :0.00000 Median :0.00000
## Mean :361.9 Mean :0.07064 Mean :0.06787
## 3rd Qu.:542.8 3rd Qu.:0.00000 3rd Qu.:0.00000
## Max. :723.0 Max. :1.00000 Max. :1.00000
##
## weaving_town have_kente parts_kente use_everyday
## Min. :0.0000 Min. :0.0000 Min. :0.000 Min. :0.000000
## 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.000 1st Qu.:0.000000
## Median :0.0000 Median :1.0000 Median :1.000 Median :0.000000
## Mean :0.1385 Mean :0.6357 Mean :0.554 Mean :0.001385
## 3rd Qu.:0.0000 3rd Qu.:1.0000 3rd Qu.:1.000 3rd Qu.:0.000000
## Max. :1.0000 Max. :1.0000 Max. :1.000 Max. :1.000000
##
## use_week use_month use_year1 use_year2
## Min. :0.0000 Min. :0.0000 Min. :0.0000 Min. :0.0000
## 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000
## Median :0.0000 Median :0.0000 Median :0.0000 Median :0.0000
## Mean :0.2078 Mean :0.7147 Mean :0.5097 Mean :0.8587
## 3rd Qu.:0.0000 3rd Qu.:0.0000 3rd Qu.:0.0000 3rd Qu.:0.0000
## Max. :2.0000 Max. :3.0000 Max. :4.0000 Max. :5.0000
##
## use_year3 use_year4 use_notatall use_kente
## Min. :0.0000 Min. :0.0000 Min. :0.000 Min. :0.000
## 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.000 1st Qu.:3.000
## Median :0.0000 Median :0.0000 Median :0.000 Median :5.000
## Mean :0.5817 Mean :0.2618 Mean :1.706 Mean :4.842
## 3rd Qu.:0.0000 3rd Qu.:0.0000 3rd Qu.:0.000 3rd Qu.:6.750
## Max. :6.0000 Max. :7.0000 Max. :8.000 Max. :8.000
##
## X1.4 X1.5a X1.5b X1.6
## Length:722 Length:722 Length:722 Length:722
## Class :character Class :character Class :character Class :character
## Mode :character Mode :character Mode :character Mode :character
##
##
##
##
## Knowledge_Cloth Knowledge_Weaving Knowledge_symbol X1.10.1
## Min. :1.000 Min. :1.000 Min. :1.000 Length:722
## 1st Qu.:2.000 1st Qu.:1.000 1st Qu.:1.000 Class :character
## Median :3.000 Median :2.000 Median :3.000 Mode :character
## Mean :3.307 Mean :2.711 Mean :3.076
## 3rd Qu.:4.000 3rd Qu.:4.000 3rd Qu.:4.000
## Max. :6.000 Max. :6.000 Max. :6.000
##
## X1.10.2 X1.10.3 X1.10.4 X1.10.5
## Length:722 Length:722 Length:722 Length:722
## Class :character Class :character Class :character Class :character
## Mode :character Mode :character Mode :character Mode :character
##
##
##
##
## X1.10.6 X1.10.7 X1.10.8 X1.10.9
## Length:722 Length:722 Length:722 Length:722
## Class :character Class :character Class :character Class :character
## Mode :character Mode :character Mode :character Mode :character
##
##
##
##
## X1.10.10 X1.10.11 X1.10.12 X1.10.13
## Length:722 Length:722 Length:722 Length:722
## Class :character Class :character Class :character Class :character
## Mode :character Mode :character Mode :character Mode :character
##
##
##
##
## X1.10.14 X1.10.15 X1.11.1 X1.11.2
## Length:722 Length:722 Length:722 Length:722
## Class :character Class :character Class :character Class :character
## Mode :character Mode :character Mode :character Mode :character
##
##
##
##
## X1.11.3 X1.11.4 X1.11.5 X1.11.6
## Length:722 Length:722 Length:722 Length:722
## Class :character Class :character Class :character Class :character
## Mode :character Mode :character Mode :character Mode :character
##
##
##
##
## X1.11.7 X1.11.8 X1.11.9 X1.11.10
## Length:722 Length:722 Length:722 Length:722
## Class :character Class :character Class :character Class :character
## Mode :character Mode :character Mode :character Mode :character
##
##
##
##
## X1.11.11 X1.11.12 X1.11.13 X1.11.14
## Length:722 Length:722 Length:722 Length:722
## Class :character Class :character Class :character Class :character
## Mode :character Mode :character Mode :character Mode :character
##
##
##
##
## X1.11.15 X1.12.1 X1.12.2 X1.12.3
## Length:722 Length:722 Length:722 Length:722
## Class :character Class :character Class :character Class :character
## Mode :character Mode :character Mode :character Mode :character
##
##
##
##
## X1.12.4 X1.12.5 X1.12.6 X1.12.7
## Length:722 Length:722 Length:722 Length:722
## Class :character Class :character Class :character Class :character
## Mode :character Mode :character Mode :character Mode :character
##
##
##
##
## X1.12.8 X1.12.9 X1.12.10 X1.13.1
## Length:722 Length:722 Mode:logical Length:722
## Class :character Class :character NA's:722 Class :character
## Mode :character Mode :character Mode :character
##
##
##
##
## X1.13.2 X1.13.3 X1.13.4 X1.13.5
## Length:722 Length:722 Length:722 Length:722
## Class :character Class :character Class :character Class :character
## Mode :character Mode :character Mode :character Mode :character
##
##
##
##
## X1.13.6 X1.13.7 X1.13.8 X1.13.9
## Length:722 Length:722 Length:722 Length:722
## Class :character Class :character Class :character Class :character
## Mode :character Mode :character Mode :character Mode :character
##
##
##
##
## X1.13.10 X1.13.1a X1.13.1b X1.13.1c
## Mode:logical Length:722 Length:722 Length:722
## NA's:722 Class :character Class :character Class :character
## Mode :character Mode :character Mode :character
##
##
##
##
## X1.13.1d X1.13.1e X1.13.1f X1.13.1g
## Length:722 Length:722 Length:722 Length:722
## Class :character Class :character Class :character Class :character
## Mode :character Mode :character Mode :character Mode :character
##
##
##
##
## X1.13.1h X1.13.1i X1.13.1j WTP_Va
## Length:722 Length:722 Length:722 Min. : 0.00
## Class :character Class :character Class :character 1st Qu.: 0.00
## Mode :character Mode :character Mode :character Median : 5.00
## Mean : 18.72
## 3rd Qu.: 15.00
## Max. :500.00
##
## X2.1.1b X2.1.2a X2.1.2b X2.1.2c1
## Length:722 Length:722 Length:722 Mode:logical
## Class :character Class :character Class :character NA's:722
## Mode :character Mode :character Mode :character
##
##
##
##
## X2.1.2c2 X2.1.3 X2.1.3name NationalCentre_Visit
## Mode:logical Length:722 Length:722 Length:722
## NA's:722 Class :character Class :character Class :character
## Mode :character Mode :character Mode :character
##
##
##
##
## Year_birth Hometown Gender EDUCATION
## Min. :1928 Length:722 Length:722 Length:722
## 1st Qu.:1973 Class :character Class :character Class :character
## Median :1983 Mode :character Mode :character Mode :character
## Mean :1980
## 3rd Qu.:1990
## Max. :2002
##
## OCCUPATION Kente_business Household_size X3.7.2
## Length:722 Length:722 Min. : 0.000 Min. : 0.000
## Class :character Class :character 1st Qu.: 2.000 1st Qu.: 1.000
## Mode :character Mode :character Median : 3.000 Median : 2.000
## Mean : 3.693 Mean : 2.223
## 3rd Qu.: 5.000 3rd Qu.: 3.000
## Max. :20.000 Max. :20.000
## NA's :77
## X3.8 X3.9 X
## Length:722 Length:722 Length:722
## Class :character Class :character Class :character
## Mode :character Mode :character Mode :character
##
##
##
## That is, 60% mean(0.6357) have kente cloth (or a clothing that is completely made of kente). Also, 50% mean(0.554) of respondents have parts of some of clothing made of kente cloth.How often do respondents use clothing made of Kente. Greater percentage of respondents shown once a month, 70% mean (0.7) once a year 50% mean(0.5), once 2 year, 80% mean(0.8), once 3 year, 50% mean(0.5). This means, respondents seldom use clothing made of Kente. This is, this is normally worn on occasions.
The perceived knowledge on Kente cloth, weaving of kente and meaning of kente symbols were assessed during the survey using a 1-to-6 Likert scale from “not knowledgeable at all” (1) to “very knowledgeable” (6). Respondents expressed an average level of perceived knowledge, mean (3.3)median(3.0) of Kente cloth in general. Again, respondents indicated a bit less than average perceived knowledgeable level mean(2.7) median(2.0) of Kente weaving. In addition, respondents shown average perceived knowledgeable level mean(3.07), median (3.0) of Kente symbols.Furthermore, about 30% mean(3.7) median(3.0) number of people above 18 years living in the household studied and 20% mean(2.2) median(2.0) number of people who are 18 years or younger living in the household studied.
- Examining the statistical relationship between individuals who have Kente cloth and their place of residence
Using chi-squared contingency table tests - independence / dependence of events
two variables included
H0: proportion of having Kente cloth is independent of their place of residence
H1: proportion of having Kente cloth is associated with their place of residence
table(data$have_kente, data$place)##
## Accra Bonwire Ho Kpetoe Kumasi
## 0 50 11 84 14 104
## 1 158 40 123 35 103chisq.test(table(data$have_kente, data$place))##
## Pearson's Chi-squared test
##
## data: table(data$have_kente, data$place)
## X-squared = 38.553, df = 4, p-value = 8.615e-08addmargins(table(data$have_kente, data$place))##
## Accra Bonwire Ho Kpetoe Kumasi Sum
## 0 50 11 84 14 104 263
## 1 158 40 123 35 103 459
## Sum 208 51 207 49 207 722H0 rejected, H1 accepted - proportions of having Kente cloth depend on study area (place). Thus, there is a significant association between the categories of the two variables. In other words, the row and the column variables are statistically significantly associated (p-value = 0). Accra as the country’s capital recorded higher proportion of respondents who have kente cloth, followed by Kumasi and Ho respectively, although the these major cities have equal sample sizes.The Kente weaving towns, that is Bonwire and Kpetoe recorded similar proportion of respondents having Kente cloth.
- Assessing whether the population distributions for this study are identical for perceived knowledge on Kente cloth and Kente weaving
Applying Wilcoxon rank sum test for the above two variables (ordinal data)
H0: median of sample A is equal median of sample B
H1: median of sample A is not equal median of sample B
wilcox.test(data$Knowledge_Cloth, data$Knowledge_Weaving)##
## Wilcoxon rank sum test with continuity correction
##
## data: data$Knowledge_Cloth and data$Knowledge_Weaving
## W = 315835, p-value = 1.173e-12
## alternative hypothesis: true location shift is not equal to 0wilcox.test(data$Knowledge_Cloth, data$Knowledge_Weaving, paired=TRUE)##
## Wilcoxon signed rank test with continuity correction
##
## data: data$Knowledge_Cloth and data$Knowledge_Weaving
## V = 48894, p-value < 2.2e-16
## alternative hypothesis: true location shift is not equal to 0Alternative hypothesis: true location shift is not equal to 0. At .05 significance level, we conclude that perceived knowledge of Kente cloth and Kente weaving in Kente data are nonidentical populations.
- Assessing the differences among means of willingness to pay to establish national centers and its factors
That is, the factor impact (Gender) on the dependent variable (willingness to pay establish national centers)
Also, the factor impacts (Gender and having Kente cloth) on the dependent variable (willingness to pay to establish national centers)
To understand if there is an interaction between Gender and having Kente cloth on willingness to pay value
Applying One-way and Two-way analysis of variance (ANOVA) approach respectively
Descriptive statistics in groups
library(psych)## Warning: package 'psych' was built under R version 4.1.3describeBy(data$WTP_Va, data$Gender)##
## Descriptive statistics by group
## group: female
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 401 13.75 33.54 5 6.94 7.41 0 500 500 8.71 111.4 1.67
## ------------------------------------------------------------
## group: male
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 321 24.92 67.63 10 10.45 14.83 0 500 500 5.67 35.59 3.77Statistics in groups show that the average values (willingness to pay values among gender are different) are different. Also, average mean for female respondents is 13.8 GHS and average mean for male respondents is 24.9 GHS.
One-way ANOVA approach
Testing groups to see if there’s a difference between them. That is, testing different groups
fit<-stats::aov(WTP_Va ~ Gender, data)
summary(fit)## Df Sum Sq Mean Sq F value Pr(>F)
## Gender 1 22260 22260 8.376 0.00392 **
## Residuals 720 1913409 2658
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1In this case p-value<0.05 so H1 accepted & H0 rejected. Variances between group is different than within group - factor matters. Again, this means there is a statistically difference between the means of the different level of the gender variable.
- Two-way anova - adding additional factor
fit<-stats::aov(WTP_Va~Gender+have_kente, data)
summary(fit)## Df Sum Sq Mean Sq F value Pr(>F)
## Gender 1 22260 22260 8.420 0.00382 **
## have_kente 1 12581 12581 4.759 0.02947 *
## Residuals 719 1900828 2644
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1In this case both p-values<0.05 so H1 accepted & H0 rejected. Variances between groups are different than within group - both factors matter. Thus, there is an interaction between the independent variables (Gender and have_Kente) on the dependent variable (WTP_Va).
Two-way ANOVA with interaction-variables
Testing to see if an important interaction is present
fit<-aov(WTP_Va~Gender+have_kente+Gender:have_kente, data)
fit<-aov(WTP_Va~Gender*have_kente, data)
summary(fit)## Df Sum Sq Mean Sq F value Pr(>F)
## Gender 1 22260 22260 8.536 0.00359 **
## have_kente 1 12581 12581 4.824 0.02838 *
## Gender:have_kente 1 28446 28446 10.908 0.00100 **
## Residuals 718 1872382 2608
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1In this case all three p-values<0.05 so H1 accepted & H0 rejected. Therefore, the variances between groups are different than within group - both factors matter individually and their interaction also matters.
- Applying Analysis of Covariance Approach
fit<-aov(WTP_Va~Gender+place, data)
summary(fit)## Df Sum Sq Mean Sq F value Pr(>F)
## Gender 1 22260 22260 8.643 0.00339 **
## place 4 69373 17343 6.734 2.53e-05 ***
## Residuals 716 1844035 2575
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1drop1(fit, ~., test="F")## Single term deletions
##
## Model:
## WTP_Va ~ Gender + place
## Df Sum of Sq RSS AIC F value Pr(>F)
## <none> 1844035 5676.4
## Gender 1 19652 1863687 5682.1 7.6304 0.005886 **
## place 4 69373 1913409 5695.1 6.7340 2.534e-05 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1fit<-aov(WTP_Va~place+Gender, data)
summary(fit)## Df Sum Sq Mean Sq F value Pr(>F)
## place 4 71981 17995 6.987 1.61e-05 ***
## Gender 1 19652 19652 7.630 0.00589 **
## Residuals 716 1844035 2575
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1drop1(fit, ~., test="F")## Single term deletions
##
## Model:
## WTP_Va ~ place + Gender
## Df Sum of Sq RSS AIC F value Pr(>F)
## <none> 1844035 5676.4
## place 4 69373 1913409 5695.1 6.7340 2.534e-05 ***
## Gender 1 19652 1863687 5682.1 7.6304 0.005886 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Inclusion of study areas (place) is highly significant
Graphics and Statistics of the relationship between willingness to pay to establish national centers (WTP_Va) and its demographic variables.
That is, have_kente/place, Gender, Household size, Year of birth and study area (place).
Boxplot of interacting variables.
boxplot(WTP_Va~Gender*have_kente, data, frame = FALSE, col=c("#00AFBB", "#E7B800", "#FC4E07"), cex.axis=0.5)
interaction.plot(factor(data$have_kente), factor(data$Gender), data$WTP_Va, type="b")
interaction.plot(factor(data$Gender), factor(data$have_kente), data$WTP_Va, type="b")
It can be seen on figures above the average values of response variable (WTP_Va) in groups by first factor and second factor (that is, Gender and have_Kente). That is, difference between both figures is in factors.
library(gplots)## Warning: package 'gplots' was built under R version 4.1.3##
## Attaching package: 'gplots'## The following object is masked from 'package:stats':
##
## lowessplotmeans(data$WTP_Va~data$Gender)
plotmeans(data$WTP_Va~data$have_kente)
It can seen on figures above the average values of response variable (WTP_Va) in groups by one of the factor (Gender/have_kente). That is, difference between both figures is in factors - which of them is on x.
- Additional ways of summarizing outputs - table of means.
fit<-aov(WTP_Va~Gender*place, data)
summary(fit)## Df Sum Sq Mean Sq F value Pr(>F)
## Gender 1 22260 22260 8.760 0.00318 **
## place 4 69373 17343 6.825 2.15e-05 ***
## Gender:place 4 34857 8714 3.429 0.00866 **
## Residuals 712 1809178 2541
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1print(model.tables(fit,"means"),digits=3)## Tables of means
## Grand mean
##
## 18.71607
##
## Gender
## female male
## 13.7 24.9
## rep 401.0 321.0
##
## place
## Accra Bonwire Ho Kpetoe Kumasi
## 31 34.9 12.1 12.2 10.6
## rep 208 51.0 207.0 49.0 207.0
##
## Gender:place
## place
## Gender Accra Bonwire Ho Kpetoe Kumasi
## female 21.5 5.0 11.0 7.3 11.4
## rep 122.0 18.0 131.0 32.0 98.0
## male 43.5 54.7 11.5 18.2 11.7
## rep 86.0 33.0 76.0 17.0 109.0Whereby ‘rep’ refers to number of observations in a group. There is a total average of about 18.7 GHS. The male respondents observed are higher than the female. The sample size studied in the major cities are equal same. However, there are few respondents studied in the Kente weaving towns as compare the major cities studied.
- Applying Multi scatterplot in groups
library(car)## Warning: package 'car' was built under R version 4.1.3## Loading required package: carData## Warning: package 'carData' was built under R version 4.1.3##
## Attaching package: 'car'## The following object is masked from 'package:psych':
##
## logitscatterplotMatrix(~Year_birth+WTP_Va+Household_size | Gender, data, smooth=FALSE, regLine=FALSE, ellipse=TRUE, by.groups=TRUE, diagonal=FALSE, legend=list(coords="bottomleft"))
A plots one variable on x and other variable on y (axes are labeled by variables on diagonal of scatterplot matrix),colours specify the groups. Continous variables were listed in code with ~ (~Year_birth+WTP_Va+Household_size), while groupping variable in place of condition, after the | (| Gender).
library(lattice)
xyplot(WTP_Va ~ Household_size | have_kente + place, groups=Gender, data, type="a", ylab="WTP", xlab="Household_size")
Show the plots of amount willingness to pay to conserve Kente on the y-axis by household_size on the axis which captures the study areas and having Kente cloth.
- Significance of differences between the groups
data$Gender<-factor(data$Gender)
data$place<-factor(data$place)
summary(fit<-aov(WTP_Va~Gender+place, data))## Df Sum Sq Mean Sq F value Pr(>F)
## Gender 1 22260 22260 8.643 0.00339 **
## place 4 69373 17343 6.734 2.53e-05 ***
## Residuals 716 1844035 2575
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1That is, both variables influence the independent variable -means in groups defined by these x variables are different.
- What is the difference in means between groups.
- Indicates one factor which is to differentiate groups.
TukeyHSD(fit, "place", ordered=FALSE)## Tukey multiple comparisons of means
## 95% family-wise confidence level
##
## Fit: aov(formula = WTP_Va ~ Gender + place, data = data)
##
## $place
## diff lwr upr p adj
## Bonwire-Accra 3.96048793 -17.72542 25.646392 0.9874112
## Ho-Accra -18.88055246 -32.50603 -5.255079 0.0015336
## Kpetoe-Accra -18.80122014 -40.83968 3.237240 0.1356773
## Kumasi-Accra -20.32857964 -33.95405 -6.703106 0.0004796
## Ho-Bonwire -22.84104039 -44.53726 -1.144825 0.0333507
## Kpetoe-Bonwire -22.76170807 -50.52438 5.000964 0.1654960
## Kumasi-Bonwire -24.28906757 -45.98528 -2.592852 0.0193196
## Kpetoe-Ho 0.07933232 -21.96928 22.127940 1.0000000
## Kumasi-Ho -1.44802718 -15.08991 12.193853 0.9984492
## Kumasi-Kpetoe -1.52735950 -23.57597 20.521248 0.9997123Pairs Ho-Accra, Kumasi-Accra, Kumasi-Bonwire and Ho-Bonwire are significantly different at 0.05 level.There is statistically significant difference between Ho and Accra. Also, there is statistically significant difference between Kumasi and Accra. There is statistically significant difference between Kumasi and Bonwire. Again, there is statistically significant difference between Ho and Bonwire.
- Plot of means by groups given with factors.
plot(TukeyHSD(fit, "place"))
plot(TukeyHSD(fit, "Gender"))
The plots show at 95% family-wise confidence level and differences in mean levels of Gender and study areas(place).
- Assessing the likely factors that affect respondents willingness to pay to establish national centers to preserve kente weaving and interpretation of its symbols and year of birth.
Testing whether or not the independent grouping variable(Gender, have_kente, place, Knowledge_Cloth, Knowledge_Weaving and Knowledge_symbol) simulataneously explains a statistically significant amount of variance in the dependent variable (WTP_Va and year of birth).
Applying MANOVA (multivariate analysis of variance) approach.
Pillai test (trace)test for MANOVA, checks if factors impact output, the higher the value the stronger the impact of factors. H0: factors do not impact output and H1 factors impact output.
fit<-manova(cbind(WTP_Va, Year_birth)~Gender+have_kente+place+Knowledge_Cloth+Knowledge_Weaving+Knowledge_symbol, data)
summary(fit)## Df Pillai approx F num Df den Df Pr(>F)
## Gender 1 0.014825 5.3496 2 711 0.004943 **
## have_kente 1 0.031854 11.6968 2 711 1.004e-05 ***
## place 4 0.076442 7.0737 8 1424 3.358e-09 ***
## Knowledge_Cloth 1 0.006186 2.2126 2 711 0.110164
## Knowledge_Weaving 1 0.009296 3.3359 2 711 0.036142 *
## Knowledge_symbol 1 0.001148 0.4088 2 711 0.664632
## Residuals 712
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1summary(fit, "Pillai") ## Df Pillai approx F num Df den Df Pr(>F)
## Gender 1 0.014825 5.3496 2 711 0.004943 **
## have_kente 1 0.031854 11.6968 2 711 1.004e-05 ***
## place 4 0.076442 7.0737 8 1424 3.358e-09 ***
## Knowledge_Cloth 1 0.006186 2.2126 2 711 0.110164
## Knowledge_Weaving 1 0.009296 3.3359 2 711 0.036142 *
## Knowledge_symbol 1 0.001148 0.4088 2 711 0.664632
## Residuals 712
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1This output show that these factors(Gender, have_kente, place and knowledge_weaving) impact both variables (considered jointly). But perceived knowledge on kente cloth and symbols are statistical insignificant.
- Comparing summary.aov(MANOVA) to double summary(ANOVA)
fit<-aov(WTP_Va~Gender+have_kente+Knowledge_Weaving, data)
summary(fit)## Df Sum Sq Mean Sq F value Pr(>F)
## Gender 1 22260 22260 8.458 0.00375 **
## have_kente 1 12581 12581 4.780 0.02911 *
## Knowledge_Weaving 1 11141 11141 4.233 0.04000 *
## Residuals 718 1889687 2632
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1fit<-aov(Year_birth~Gender+have_kente+Knowledge_Weaving, data)
summary(fit)## Df Sum Sq Mean Sq F value Pr(>F)
## Gender 1 439 439.0 2.581 0.10858
## have_kente 1 2845 2844.7 16.727 4.81e-05 ***
## Knowledge_Weaving 1 2542 2542.3 14.949 0.00012 ***
## Residuals 718 122109 170.1
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1fit3<-car::Manova(lm(cbind(WTP_Va, Household_size)~Gender+have_kente+Knowledge_Weaving, data), type="III")
summary(fit3, multivariate=TRUE)##
## Type III MANOVA Tests:
##
## Sum of squares and products for error:
## WTP_Va Household_size
## WTP_Va 1889687.014 -6502.327
## Household_size -6502.327 5302.137
##
## ------------------------------------------
##
## Term: (Intercept)
##
## Sum of squares and products for the hypothesis:
## WTP_Va Household_size
## WTP_Va 1631.060 1622.527
## Household_size 1622.527 1614.039
##
## Multivariate Tests: (Intercept)
## Df test stat approx F num Df den Df Pr(>F)
## Pillai 1 0.2358740 110.6634 2 717 < 2.22e-16 ***
## Wilks 1 0.7641260 110.6634 2 717 < 2.22e-16 ***
## Hotelling-Lawley 1 0.3086846 110.6634 2 717 < 2.22e-16 ***
## Roy 1 0.3086846 110.6634 2 717 < 2.22e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## ------------------------------------------
##
## Term: Gender
##
## Sum of squares and products for the hypothesis:
## WTP_Va Household_size
## WTP_Va 17186.02631 50.5566698
## Household_size 50.55667 0.1487241
##
## Multivariate Tests: Gender
## Df test stat approx F num Df den Df Pr(>F)
## Pillai 1 0.0091429 3.307969 2 717 0.03715 *
## Wilks 1 0.9908571 3.307969 2 717 0.03715 *
## Hotelling-Lawley 1 0.0092272 3.307969 2 717 0.03715 *
## Roy 1 0.0092272 3.307969 2 717 0.03715 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## ------------------------------------------
##
## Term: have_kente
##
## Sum of squares and products for the hypothesis:
## WTP_Va Household_size
## WTP_Va 8116.70786 90.544590
## Household_size 90.54459 1.010055
##
## Multivariate Tests: have_kente
## Df test stat approx F num Df den Df Pr(>F)
## Pillai 1 0.0046015 1.657272 2 717 0.19139
## Wilks 1 0.9953985 1.657272 2 717 0.19139
## Hotelling-Lawley 1 0.0046228 1.657272 2 717 0.19139
## Roy 1 0.0046228 1.657272 2 717 0.19139
##
## ------------------------------------------
##
## Term: Knowledge_Weaving
##
## Sum of squares and products for the hypothesis:
## WTP_Va Household_size
## WTP_Va 11141.0735 457.14250
## Household_size 457.1425 18.75755
##
## Multivariate Tests: Knowledge_Weaving
## Df test stat approx F num Df den Df Pr(>F)
## Pillai 1 0.0099689 3.609843 2 717 0.027549 *
## Wilks 1 0.9900311 3.609843 2 717 0.027549 *
## Hotelling-Lawley 1 0.0100693 3.609843 2 717 0.027549 *
## Roy 1 0.0100693 3.609843 2 717 0.027549 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1This shows the III type SS, each section is for different explanatory variable. Also, the various tests show that, the factors impact output.
- Estimating the significant impact of some factors (place and gender) on willingness to pay national centers and year of birth of respondents.
Applying Multivariate multiple linear regression approach.
Standard regression function with two dependent variables
mlm<-lm(cbind(WTP_Va, Year_birth)~place+Gender, data)
mlm##
## Call:
## lm(formula = cbind(WTP_Va, Year_birth) ~ place + Gender, data = data)
##
## Coefficients:
## WTP_Va Year_birth
## (Intercept) 26.1977 1980.2152
## placeBonwire 4.0802 -1.6320
## placeHo -18.9043 -3.3136
## placeKpetoe -18.8353 -5.4207
## placeKumasi -20.2706 3.5064
## Gendermale 10.6614 0.8865The results show, there is an impact of each level of factor in each dependent variable, that is coefficients only. In addition, impact is as in individual regressions.
summary(mlm)## Response WTP_Va :
##
## Call:
## lm(formula = WTP_Va ~ place + Gender, data = data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -40.94 -16.86 -7.29 2.71 473.80
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 26.198 3.864 6.780 2.51e-11 ***
## placeBonwire 4.080 7.981 0.511 0.609334
## placeHo -18.904 4.986 -3.792 0.000162 ***
## placeKpetoe -18.835 8.063 -2.336 0.019762 *
## placeKumasi -20.271 5.001 -4.053 5.61e-05 ***
## Gendermale 10.661 3.860 2.762 0.005886 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 50.75 on 716 degrees of freedom
## Multiple R-squared: 0.04734, Adjusted R-squared: 0.04069
## F-statistic: 7.116 on 5 and 716 DF, p-value: 1.653e-06
##
##
## Response Year_birth :
##
## Call:
## lm(formula = Year_birth ~ place + Gender, data = data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -48.902 -6.902 2.392 9.098 24.212
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 1980.2152 0.9910 1998.135 < 2e-16 ***
## placeBonwire -1.6320 2.0470 -0.797 0.42559
## placeHo -3.3136 1.2788 -2.591 0.00976 **
## placeKpetoe -5.4207 2.0681 -2.621 0.00895 **
## placeKumasi 3.5064 1.2828 2.733 0.00643 **
## Gendermale 0.8865 0.9900 0.895 0.37083
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 13.02 on 716 degrees of freedom
## Multiple R-squared: 0.05172, Adjusted R-squared: 0.0451
## F-statistic: 7.811 on 5 and 716 DF, p-value: 3.593e-07Also, the results show that, Ho, Kpetoe and Kumasi as a study areas (place) show a negative significant impact on willingness to pay values to establish national centers to conserve Kente at 0, 0.01, 0 and 0.001 levels respectively. Also male gender indicate a positive siginificant effect on willingness to pay values at 5% level. Again, Bonwire as study shows no significant effect.
Again, the results indicate that, Ho and Kpetoe as a study areas (place) show a negative significant impact on age (year of birth) of respondents at 0.01 levels. Also, Kumasi as a study area indicate a positive significant effect on age (year of birth) of respondents at 0.01 level. Again, male gender indicate no significant effect.
- SS type I
summary(stats::manova(mlm), test="Pillai")## Df Pillai approx F num Df den Df Pr(>F)
## place 4 0.088429 8.2805 8 1432 5.067e-11 ***
## Gender 1 0.011291 4.0827 2 715 0.01726 *
## Residuals 716
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1The factors impact the outputs
- SS type II
car::Manova(mlm, type="II")##
## Type II MANOVA Tests: Pillai test statistic
## Df test stat approx F num Df den Df Pr(>F)
## place 4 0.085149 7.9597 8 1432 1.55e-10 ***
## Gender 1 0.011291 4.0827 2 715 0.01726 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1The factors impact the outputs
- SS type III
car::Manova(mlm, type="III")##
## Type III MANOVA Tests: Pillai test statistic
## Df test stat approx F num Df den Df Pr(>F)
## (Intercept) 1 0.99982 1999239 2 715 < 2.2e-16 ***
## place 4 0.08515 8 8 1432 1.55e-10 ***
## Gender 1 0.01129 4 2 715 0.01726 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1The factors impact the outputs
- The summary can be extended
summary(car::Manova(mlm, type="II"))##
## Type II MANOVA Tests:
##
## Sum of squares and products for error:
## WTP_Va Year_birth
## WTP_Va 1844035.39 26977.06
## Year_birth 26977.06 121317.96
##
## ------------------------------------------
##
## Term: place
##
## Sum of squares and products for the hypothesis:
## WTP_Va Year_birth
## WTP_Va 69373.2199 -920.8781
## Year_birth -920.8781 6178.2892
##
## Multivariate Tests: place
## Df test stat approx F num Df den Df Pr(>F)
## Pillai 4 0.0851486 7.959681 8 1432 1.5501e-10 ***
## Wilks 4 0.9166097 7.954086 8 1430 1.5818e-10 ***
## Hotelling-Lawley 4 0.0890586 7.948476 8 1428 1.6142e-10 ***
## Roy 4 0.0525628 9.408748 4 716 2.0472e-07 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## ------------------------------------------
##
## Term: Gender
##
## Sum of squares and products for the hypothesis:
## WTP_Va Year_birth
## WTP_Va 19651.976 1634.054
## Year_birth 1634.054 135.871
##
## Multivariate Tests: Gender
## Df test stat approx F num Df den Df Pr(>F)
## Pillai 1 0.0112911 4.082674 2 715 0.017257 *
## Wilks 1 0.9887089 4.082674 2 715 0.017257 *
## Hotelling-Lawley 1 0.0114201 4.082674 2 715 0.017257 *
## Roy 1 0.0114201 4.082674 2 715 0.017257 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1The factors strongly impact the outputs
Conclusion
The irrelevance of material aspects of heritage in the definition of cultural heritage has provided convincing arguments for expansion of cultural heritage to include intangible cultural heritage(ICH). The 2003 UNESCO Convention recognizes the importance of preservation of intangible cultural heritage through the safeguarding of ICHs as part of cultural heritage. Therefore, the project seek to evaluate the value of conserving Kente weaving and interpretation of its symbols as an intangible cultural heritage.It can be concluded that, there is significant association between respondents having Kente cloth and study areas. Also, there is an interaction between the independent variables (Gender and having Kente cloth) on willingness to pay values to conserve Kente weaving and interpretation of its symbols. Moreover, these independent variables such as gender, having kente cloth , study areas and perceived knowledge on Kente weaving impact both willingness to pay values and year of birth of respondents. Furthermore, the findings show that, Ho, Kpetoe and Kumasi as a study areas show a negative significant impact on willingness to pay values to establish national centers to conserve Kente. In addition, male gender indicates a positive significant effect on willingness to pay values. Again, the results show that, Ho and Kpetoe as study areas show a negative significant impact on age (year of birth) of respondents. Also, Kumasi as a study area indicate a positive significant effect on age (year of birth) of respondents.
SessionInfo:
sessionInfo()## R version 4.1.0 (2021-05-18)
## Platform: x86_64-w64-mingw32/x64 (64-bit)
## Running under: Windows 10 x64 (build 19043)
##
## Matrix products: default
##
## locale:
## [1] LC_COLLATE=English_United States.1252
## [2] LC_CTYPE=English_United States.1252
## [3] LC_MONETARY=English_United States.1252
## [4] LC_NUMERIC=C
## [5] LC_TIME=English_United States.1252
##
## attached base packages:
## [1] stats graphics grDevices utils datasets methods base
##
## other attached packages:
## [1] lattice_0.20-44 car_3.1-0 carData_3.0-5 gplots_3.1.3
## [5] psych_2.2.5
##
## loaded via a namespace (and not attached):
## [1] bslib_0.3.1 compiler_4.1.0 jquerylib_0.1.4 highr_0.9
## [5] bitops_1.0-7 tools_4.1.0 digest_0.6.29 jsonlite_1.8.0
## [9] evaluate_0.15 nlme_3.1-152 rlang_1.0.1 cli_3.2.0
## [13] rstudioapi_0.13 yaml_2.3.5 parallel_4.1.0 xfun_0.30
## [17] fastmap_1.1.0 stringr_1.4.0 knitr_1.38 sass_0.4.1
## [21] gtools_3.9.2 caTools_1.18.2 grid_4.1.0 R6_2.5.1
## [25] rmarkdown_2.16 magrittr_2.0.1 htmltools_0.5.2 MASS_7.3-57
## [29] abind_1.4-5 mnormt_2.0.2 KernSmooth_2.23-20 stringi_1.7.6
## [33] tmvnsim_1.0-2References
Ahiagble, B.D. The Pride of Ewe Kente; Sub-Saharan Publishers: Accra, Ghana, 2004.
Asamoah-Yaw, E. Kente Cloth: Introduction to History; Ghanam Textiles Inc.: New York, NY, USA, 1999.
Fening, K.O. History of Kente Cloth and Its Value Addition through Design Integration with African Silk for Export Market in Ghana. In Proceedings of the Trainers Course and 4th International Workshop on the Conservation and Utilisation of Commercial Insects Duduville, Nairobi, Kenya, 14 November-8 December 2006.
Lartey, R.L. Integrated Cultural Weaves (Fugu, Kente and Kete) Woven with Organic Dyed Yarns. Master’s Thesis, Department of Integrated Rural Art and Industry, College of Art and Built Environment, Kwame Nkrumah University of Science and Technology, Kumasi, Ghana, 2014.
Vecco, M. A definition of cultural heritage: From the tangible to intangible. J. Cult. Herit. 2010, 11, 321-324.
Vondolia, G.K.; Kusi, A.M.;King, S.R.; Navrud, S. Valuing Intangible Cultural Heritage in Developing Countries. Sustainability 2022, 14, 4484.