SPSS Analysis of the Quantitative Data

In document CHAPTER ONE INTRODUCTION (halaman 104-122)

ICE-GB

4.6 SPSS Analysis of the Quantitative Data

The study employs the use of SPSS analysis so as to describe the significance and insignificance of each category. Statistical analysis makes it possible to describe this. The analysis has been approached manually before the invention of the statistical software tools.

Now that the software is available, it saves time as well as eases the analytical procedures with accuracy and clarity.

This section attempts to quantitatively answer research question two (RQ2) “How do the patterns of use of the preposition of in ICE-Nig. compare to those in ICE-GB?”

105 Figure 4.3 Categories and the percentages shared across the two corpora The uneven distribution of the preposition appears to be clear across the categorization paradigm in both of the corpora with Process Relationship Category emerging as the highest in ICE-Nig. and Attributive Relationship category being the topmost in the ICE-GB. Differences have been described in respect of each category. The level of significance of the differences of the use of the preposition of in each category is stated. As in the tradition of the statistical analysis, this leads to the acceptance or rejection of the null hypotheses, the heading of each category.

p<α

p<α

p<α

p<α

p<α

106

``4.6.1 Null Hypothesis 1: There is equal proportion of frequency in Partitive Relationship Category of the preposition of across ICE-Nig. and ICE-GB.

Table 4.2a Descriptive statistics results for Partitive Relationship frequencies Parameters Percent Observed N Expected N Residual ICE-Nig.

ICE-GB Total

44 56 100

126 162 288

144.0 144.0 288

-18.0 18.0

Partitive

Figure 4.4 Distributions of Partitive Relationship Frequencies across ICE-Nig. and ICE-GB

As Table 4.2a and Figure 4.4 indicate, the percentage of occurrence of the proposition of is lower in ICE-Nig. (44%) as compared with that of ICE-GB (56%).

Table 4.2b Chi-square Results for Partitive Relationship

Parameters Partitive

Chi-Square 4.500a

Df 1

Asymp. Sig. .034

a. 0 cells (0.0%) have expected frequencies less than 5.

The minimum expected cell frequency is 144.0.

107 Based on the results of Goodness of fit Chi Square, X2 (1, N=287) = 4.50, p = .034, (Table 4.2a) this difference was statistically significant. That is, the partitive relationship category has been disproportionately distributed across the two corpora with a higher number of occurences in ICE-GB. This leads to the rejection of the nul hypothesis and it proves that there is a sinificant difference in the use of the preposition in the Partitive category across the two corpora as alpha is <.05

4.6.2 Null Hypothesis 2: There is equal proportion of frequency in Content Relationship Category of the preposition of across ICE-Nig. and ICE-GB.

Table 4.3a Descriptive Statistics Results for Content Relationship Frequencies Parameters Percent Observed N Expected N Residual

ICE-Nig. 63 60 47.5 12.5

ICE-GB 37 35 47.5 -12.5

Total 100 95 95

Content

Figure 4.5 Distributions of Content Relationship Frequencies across ICE-Nig. and ICE-GB

As Table 4.3a and Figure 4.5 indicate, the overall percentage of occurrence of the proposition is higher in ICE-Nig. (63%) as compared with that of ICE-GB (37%).

108 Table 4.3b Chi-square Results for Content Relationship

Based on the results of Goodness of fit Chi Square, X2 (1, N=95) = 6, p = .010 (Table 4.3a) this difference was statistically significant. That is, the content relationship category has been disproportionately distributed across the two corpora with a higher number of occurences in ICE-Nig. This leads to the rejection of the nul hypothesis and it proves that there is a sinificant difference in the use of the preposition in the Content category across the two corpora as alpha is <.05

4.6.3 Null Hypothesis 3: There is equal proportion of frequency in Quantitative Relationship Categories of the preposition of across ICE-Nig. and ICE-GB.

Table 4.4a Descriptive Statistics Results for Quantitative Relationship Frequencies Percent Observed N Expected N Residual

ICE-Nig. 48 61 64.0 -3.0

ICE-GB 52 67 64.0 3.0

Total 100 128 128

Parameters Content

Chi-Square 6.579a

df 1

Asymp. Sig. .010

a. 0 cells (0.0%) have expected frequencies less than 5. The minimum expected cell frequency is 47.5.

109 Figure 4.6 Distributions of Quantitative Relationship Frequencies across ICE-Nig.

and ICE-GB

As Table 4.4a and Figure 4.6 indicate, the overall percentage of occurrence of the proposition is lower in ICE-Nig. (48%) as compared with that of ICE-GB Corpus (52%).

Table 4.4b Chi-square Results for Quantitative Relationship

Based on the results of Goodness of fit Chi Square, X2 (1, N=128) = .281, p = .596, (Table 4.4a) this difference was statistically insignificant. That is, the quantitative relationship category has been closely distributed across the two corpora with slightly higher number of occurences in ICE-GB. This leads to the acceptance of the nul hypothesis and it proves that there is an insinificant difference in the use of the preposition in the Quantitative category across the two corpora as alpha is >.05

4.6.4 Null Hypothesis 4: There is equal proportion of frequency in Extent Relationship Categories of the preposition of across ICE-Nig. and ICE-GB.

Parameters Quantitative

Chi-Square .281a

Df 1

Asymp. Sig. .596

a. 0 cells (0.0%) have expected frequencies less than 5. The minimum expected cell frequency is 64.0.

110 Table 4.5a Descriptive Statistics Results for Extent Relationship Frequencies

Parameters Percent Observed N Expected N Residual

ICE-Nig. 48 57 59.0 -2.0

ICE-GB 52 61 59.0 2.0

Total 100 118 118

Figure 4.7 Distributions of Extent Relationship Frequencies across Nig. and ICE-GB

As Table 4.5a and Figure 4.7 indicate, the overall percentage of occurrence of the proposition is lower in ICE-Nig. (48%) as compared with that of ICE-GB (52%).

Table 4.5b Chi-square results for Extent Relationship

Parameters Extent

Chi-Square df Asymp Sig

.136a 1 .713 a. 0 cells (0.0%) have expected frequencies less than 5.

The minimum expected cell frequency is 59.0.

Based on the results of Goodness of fit Chi Square, X2 (1, N=118) = .136, p = .713, (Table 5a) this difference was statistically insignificant. Extent relationship category has been closely distributed across the two corpora with slightly higher number of occurences in ICE-GB. This leads to the acceptance of the nul hypothesis and it proves that there is an

111 insignificant difference in the use of the preposition in the Extent category across the two corpora as alpha is >.05

4.6.5 Null Hypothesis 5: There is equal proportion of frequency in Source Relationship Categories of the preposition of across ICE-Nig. and ICE-GB.

Table 4.6a Descriptive Statistics Results for Source Relationship Frequencies Parameters Percent Observed N Expected N Residual

ICE-Nig. 53 20 19.0 1.0

ICE-GB 47 18 19.0 -1.0

Total 100 38 38

Figure 4.8 Distribution of Source Relationship Frequency across Nig. and ICE-GB

As Table 4.6a and Figure 4.8 indicate, the overall percentage of occurrence of the proposition is higher in ICE-Nig. (53%) as compared with that of ICE-GB (47%).

Table 4.6b Chi-square Results for Source Relationship

Parameters Source

Chi-Square .105a

Df 1

Asymp. Sig. .746

a. 0 cells (0.0%) have expected frequencies less than 5.

The minimum expected cell frequency is 19.0.

112 Based on the results of Goodness of fit Chi Square, X2 (1, N=38) = 105, p = .746, (Table 4.6a) this difference was statistically insignificant. That is, the source relationship category has been disproportionately distributed across the two corpora with a higher number of occurences in ICE-Nig. This leads to the acceptance of the nul hypothesis and it proves that there is an insinificant difference in the use of the preposition in the Source category across the two corpora as alpha is >.05

4.6.6 Null Hypothesis 6: There is equal proportion of frequency in Attributive Relationship Categories of the preposition of across ICE-Nig. and ICE-GB.

Table 4.7a Descriptive Statistics Results for Attributive Relationship Frequencies Parameters Percent Observed N Expected N Residual

ICE-Nig. 30 82 137.0 -55.0

ICE-GB 70 192 137.0 55.0

Total 100 274 274

Figure 4.9 Distributions of Attributive Relationship Frequencies across ICE-Nig. and ICE-GB

113 As Table 4.7a and Figure 4.9 indicate, the overall percentage of occurrence of the proposition is lower in ICE-Nig. (30%) as compared with that of ICE-GB (70%).

Table 4.7b Chi-Square Results for Attributive Relationship

Parameters Attributive

Chi-Square 44.161a

Df 1

Asymp. Sig. .000

a. 0 cells (0.0%) have expected frequencies less than 5.

The minimum expected cell frequency is 137.0.

Based on the results of Goodness of fit Chi Square, X2 (1, N=274) = 44.161, p = .000, (Table 4.7a) this difference was statistically significant. That is, the attributive relationship category has been disproportionately distributed across the two corpora with a higher number of occurences in ICE-GB. This leads to the rejection of the nul hypothesis and it proves that there is a sinificant difference in the use of the preposition in the Attribution category as alpha is < .05

4.6.7 Null Hypothesis 7: There is equal proportion of frequency in Temporal Relationship Categories of the preposition of across ICE-Nig. and ICE-GB.

Table 4.8a Descriptive Statistics Results for Temporal Relationship Frequencies Parameters Percent Observed N Expected N Residual

ICE-Nig. 53 19 18.0 1.0

ICE-GB 47 17 18.0 -1.0

Total 100 36 36

114 Figure 4.10 Distributions of Temporal Relationship Frequencies across ICE-Nig. and ICE-GB

As Table 4.8 and Figure 4.10 indicate, the overall percentage of occurrence of the proposition is higher in ICE-Nig. (53%) as compared with that of ICE-GB (47%).

Table 4.8b Chi-square Results for Temporal Relationship

Parameters Temporal

Chi-Square .111a

Df 1

Asymp. Sig. .739

a. 0 cells (0.0%) have expected frequencies less than 5.

The minimum expected cell frequency is 18.0.

Based on the results of Goodness of fit Chi Square, X2 (1, N=36) = .111, p =.739, (Table 4.8a) this difference was statistically significant. That is, the Temoral relationship category has been disproportionately distributed across the two corpora with a higher number of occurences in ICE-Nig. This leads to the acceptance of the nul hypothesis and it proves that there is an insinificant difference in the use of the preposition in the Temporal category across the two corpora as alpha is >.05.

4.6.8 Null Hypothesis 8: There is equal proportion of frequency in Cause Relationship Categories of the preposition of across ICE-Nig. and ICE-GB.

115 Table 4.9a Descriptive Statistics Results for Cause Relationship Frequencies

Parameters Percent Observed N Expected N Residual

ICE-Nig. 61 40 33.0 7.0

ICE-GB 39 26 33.0 -7.0

Total 100 66 66

Figure 4.11 Distributions of Cause Relationship Frequencies across ICE-Nig. and ICE-GB

As Table 4.9a and Figure 4.11 indicate, the overall percentage of occurrence of the proposition is higher in ICE-Nig. (61%) as compared with that of ICE-GB (39%).

Table 4.9b Chi-square Results for Cause Relationship

Parameters Cause

Chi-Square 2.970a

df 1

Asymp. Sig. .085

a. 0 cells (0.0%) have expected frequencies less than 5. The minimum expected cell frequency is 33.0.

Based on the results of Goodness of fit Chi Square, X2 (1, N=66) = 2.970, p=.085, (Table 4.9a) this difference was statistically insignificant. That is, the cause relationship category has been disproportionately distributed across the two corpora with a higher number of occurences in ICE-Nig. This leads to the acceptance of the nul hypothesis and it

116 proves that there is an insinificant difference in the use of the preposition in the Cause category across the two corpora as alpha is >.05.

4.6.9 Null Hypothesis 9: There is equal proportion of frequency in Possessive Relationship Cate-gories of the preposition of across ICE-Nig. and ICE-GB.

Table 4.10a Descriptive Statistics Results for Possessive Relationship Frequencies Parameters Percent Observed N Expected N Residual

ICE-Nig. 56 28 25.0 3.0

ICE-GB 44 22 25.0 -3.0

Total 100 50 50

Figure 4.12 Distributions of Possessive Relationship Frequencies across ICE-Nig. and ICE-GB

As Table 4.10a and Figure 4.12 indicate, the overall percentage of occurrence of the pre- position is higher in ICE-Nig. (56%) as compared with that of ICE-GB (44%).

Table 4.10b Chi-square Results for Possessive Relationship

Parameters Possessive

Chi-Square .720a

df 1

Asymp. Sig. .396

a. 0 cells (0.0%) have expected frequencies less than 5. The minimum expected cell frequency is 25.0.

117 Based on the results of Goodness of fit Chi Square, X2 (1, N=50) = .720, p =.396, (Table 4.10a) this difference was statistically significant. That is, the possessive relationship category has been disproportionately distributed across the two corpora with a higher number of occurences in ICE-Nig. This leads to the acceptance of the nul hypothesis and it proves that there is an insinificant difference in the use of the preposition in the Possession category across the two corpora as alpha is >.05.

4.6.10 Null Hypothesis 10: There is equal proportion of frequency in Position/Location Relation -ship Categories of the preposition of across ICE-Nig. and ICE-GB.

Table 4.11a Descriptive Statistics Results for Position/Location Relationship Frequencies

Figure 4.13 Distributions of Position/Location Relationship Frequencies across ICE-Nig. and ICE-GB

Parameters Percent Observed N Expected N Residual

ICE-Nig. 26 20 38.5 -18.5

ICE-GB 74 57 38.5 18.5

Total 100 77 77

118 As Table 4.11a and Figure 4.13 indicate, the overall percentage of occurrence of the proposition is lower in ICE-Nig. (26%) as compared with that of ICE-GB (74%).

Table 4.11b Chi-square Results for Position/Location Relationship

Parameters Position/Location

Chi-Square 17.779a

df 1

Asymp. Sig. .000

a. 0 cells (0.0%) have expected frequencies less than 5.

The minimum expected cell frequency is 38.5.

Based on the results of Goodness of fit Chi Square, X2 (1, N=77) = 17.779, p =.000, (Table 4.11a) this difference was statistically significant. That is, the Position/Location relationship category has been disproportionately distributed across the two corpora with a higher number of occurences in ICE-GB. This leads to the rejection of the nul hypothesis and it proves that there is an insinificant difference in the use of the preposition in the Position/Location category across the two corpora as alpha is >.05.

4.6.11 Null Hypothesis 11: There is equal proportion of frequency in Process Relationship Cate-gories of the preposition of across ICE-Nig. and ICE-GB.

Table 4.12a Descriptive Statistics Results for Process Relationship Frequencies Parameters Percent Observed N Expected N Residual

ICE-Nig. 51 171 167.5 3.5

ICE-GB 49 164 167.5 -3.5

Total 100 335 335

119 Figure 4.14 Distributions of Process Relationship Frequencies across ICE-Nig. and ICE-GB

As Table 4.12a and Figure 4.14 indicate, the overall percentage of occurrence of the proposition is higher in ICE-Nig. (51.0%) as compared with that of ICE-GB (49.0%)

Table 4.12b Chi-square Results for Process Relationship

Parameters Process

Chi-Square .146a

Df 1

Asymp. Sig. .702

a. 0 cells (0.0%) have expected frequencies less than 5.

The minimum expected cell frequency is 167.5.

Based on the results of Goodness of fit Chi Square, X2 (1, N=335) = .146, p =.702, (Table 4.12a) this difference was statistically significant. That is, the process relationship category has been disproportionately distributed across the two corpora with a higher number of occurences in ICE-Nig. This leads to the acceptance of the nul hypothesis and it proves that there is an insinificant difference in the use of the preposition in the Process category across the two corpora as alpha is >.05.

4.6.12 Null Hypothesis 12: There is equal proportion of frequency in Separate-From Relationship Categories of the preposition of across ICE-Nig. and ICE-GB.

120 Table 4.13a Descriptive Statistics Results for Separate-From Relationship Frequencies

Parameters Percent Observed N Expected N Residual

ICE-Nig. 67 8 6.0 2.0

ICE-GB 33 4 6.0 -2.0

Total 100 12 12

Figure 4.15 Distributions of Separate-From Relationship Frequencies across ICE-Nig.

and ICE-GB

As Table 4.13a and Figure 4.15 indicate, the overall percentage of occurrence of the pre-position is higher in ICE-Nig. (67%) as compared with that of ICE-GB (33%).

Table 4.13b Chi-square Results for Separate-From Relationship

Parameters Separate-From

Chi-Square 1.333a

Df 1

Asymp. Sig. .248

a. 0 cells (0.0%) have expected frequencies less than 5.

The minimum expected cell frequency is 6.0.

Based on the results of Goodness of fit Chi Square, X2 (1, N=12) = 1.333, p =.248, (Table 4.13a) this difference was statistically significant. That is, the Separate-From relationship category has been disproportionately distributed across the two corpora with a higher number of occurences in ICE-Nig. This leads to the acceptance of the nul hypothesis

121 and it proves that there is an insinificant difference in the use of the preposition in the Separate-From category across the two corpora as alpha is >.05.

4.6.13 Null Hypothesis 13: There is equal proportion of frequency in Loss Relationship Categories of the preposition of across ICE-Nig. and ICE-GB

Table 4.14a Descriptive Statistics Results for Loss Relationship Frequencies Parameters Percent Observed N Expected N Residual

ICE-Nig. 19 7 18.0 -11.0

ICE-GB 81 29 18.0 11.0

Total 100 36 36

Figure 4.16 Distributions of Loss Relationship Frequencies across Nig. and ICE-GB

As Table 4.14a and Figure 4.16 indicate, the overall percentage of occurrence of the proposition is lower in ICE-Nig. (19%) as compared with that of ICE-GB (81%)

122 Table 4.14b Chi-square Results for Loss Relationship

Parameters Loss

Chi-Square 13.444a

Df 1

Asymp. Sig. .000

a. 0 cells (0.0%) have expected frequencies less than 5.

The minimum expected cell frequency is 18.0.

Based on the results of Goodness of fit Chi Square, X2 (1, N=36) =13 .444, p =.000, (Table 4.14a) this difference was statistically significant. That is, the Loss relationship category has been disproportionately distributed across the two corpora with a higher number of occurences in ICE-GB. This leads to the rejection of the nul hypothesis and it proves that there is a insinificant difference in the use of the preposition in the Loss category across the two corpora as alpha is >.05.

In document CHAPTER ONE INTRODUCTION (halaman 104-122)