Results

1. TOEIC and SRP Regression

Results

In writing up the key variables data, Self-rated proficiency (SRF)/toeic we could say (use this as a model for regression results):

SRP is a significant predictor of toeic scores. b = 46.067, CI 95% = [39.10485, 53.0292], t(596) = 12.99, p < .001. The intercept was 306.953 (so, with an increase of 1 SRP score, we see a toeic increase of 46.067). SRP scores also explained a significant proportion of variance in toeic scores. F(1, 596) = 168.9, p < 0.05. The effect size was small size; adjusted r squared = 0.2195, CI = [0.1625678, 0.2793126].

Plot

***

3. TOEIC and Year ANOVA

Results

Students’ years in university (ie: first year, second year and third year) significantly predicted TOEIC score. Using 1-way ANOVA, we found F(2, 595)=95.94, p<0.001. Between years one and three, we found a cohen’s d of 1.8 (large size) and a Pearson’s r of 0.65 (medium size?). Between years two and three, cohen’s d = 0.93 with an r of 0.37.

Year	n	mean	sd	median	se
1	183	434.86	98.5	450	7.28
2	319	531.17	139.22	515	7.8
3	96	669.17	174.84	700	17.84

## 
## Call:
## lm(formula = TOEIC ~ Year, data = collapseYear)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -439.17  -84.86   -6.17   77.81  403.83 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  434.858      9.975  43.595  < 2e-16 ***
## Year2         96.314     12.513   7.697 5.82e-14 ***
## Year3        234.309     17.005  13.779  < 2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 134.9 on 595 degrees of freedom
##   (227 observations deleted due to missingness)
## Multiple R-squared:  0.2438, Adjusted R-squared:  0.2413 
## F-statistic: 95.94 on 2 and 595 DF,  p-value: < 2.2e-16

## 
##   Simultaneous Tests for General Linear Hypotheses
## 
## Multiple Comparisons of Means: Tukey Contrasts
## 
## 
## Fit: lm(formula = TOEIC ~ Year, data = collapseYear)
## 
## Linear Hypotheses:
##            Estimate Std. Error t value Pr(>|t|)    
## 2 - 1 == 0    96.31      12.51   7.697   <1e-10 ***
## 3 - 1 == 0   234.31      17.00  13.779   <1e-10 ***
## 3 - 2 == 0   137.99      15.71   8.785   <1e-10 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## (Adjusted p values reported -- single-step method)

## 
##   Simultaneous Confidence Intervals
## 
## Multiple Comparisons of Means: Tukey Contrasts
## 
## 
## Fit: lm(formula = TOEIC ~ Year, data = collapseYear)
## 
## Quantile = 2.3408
## 95% family-wise confidence level
##  
## 
## Linear Hypotheses:
##            Estimate lwr      upr     
## 2 - 1 == 0  96.3145  67.0243 125.6047
## 3 - 1 == 0 234.3087 194.5043 274.1132
## 3 - 2 == 0 137.9943 101.2251 174.7635

## 
##  Pairwise comparisons using t tests with pooled SD 
## 
## data:  collapseYear$TOEIC and collapseYear$Year 
## 
##   1       2      
## 2 1.7e-13 -      
## 3 < 2e-16 < 2e-16
## 
## P value adjustment method: bonferroni

Plot

---

4. TOEIC Scores and Contact Details t=test

Results

2 sample T-Test: T-test showed that there was a significant difference in TOEIC scores (M=between students who didn’t (M = 511.67 SE = 6.78) and those who left contact details (M=586.8 SE=15.89). This difference was significant; difference in means = -75.08 [-109.256 -40.904], t(133.28)= -4.3453, p < 0.001. Effect size was a medium size r=0.35.

Contact Details	n	mean	sd	median	se
0	501	511.67	151.77	500	6.78
1	97	586.75	156.52	550	15.89

Plot

***

5. TOEIC and Growth Mindset Regression

Results

A growth mindset predicted TOEIC scores. Using regression analysis we found b= 3.0440 [1.508349, 4.579743], t(596) = 3.893, p = .00011 F(1, 596)=15.15, p<0.001, effect size r sq=0.02316, [0.005303419, 0.05261701]. Intercept was 404.2477

Plot

***

6. TOEIC and Fixed Mindset Regression

Results

Fixed mindsets however was not significant. Intercept of 565.1311; b = -1.378, [-3.213, 0.456], t = -1.476 p = 0.141 F(1, 596)=2.177, p=0.14. Effect size r squared = .002 [0.0, 0.01546194]

Plot

***

9. Contact Details and SRP t-test

Results

Looking at the relationship between students who left an email, and their Self-rated proficiency, we used a 2 sample t-test and found a difference of means between the two groups of -0.794, [-1.116 -0.472], t(140.3)= -4.875, p<0.001, effect size r = 0.38 (medium size).

(I have to explain why the t statistic is a decimal point, no??) - You can explain that it is a Welch’s t-test, which is used when the two samples have unequal variance, and this gives the particular df outcome. Your groups are very unbalanced in size. You should look at the possible consequences of this for the analysis

Contact Details	mean	n	sd	median	se
0	4.542	715	1.507	4.5714	0.056
1	5.336	110	1.602	5.4286	0.153

Plot

***

10. Growth Mindset and SRP Regression

Results

Growth mindset predicted Self-rated proficiency. Intercept: 3.335; b = 0.034, [0.021 0.046] t = 5.273, p < .001; F(1, 823)=27.8, p<0.001. r squared = 0.032 [0.01252844, 0.05954831]

Plot

---

11. Fixed Mindset and SRP Regression

Results

The regression results for fixed mindsets and SRP however were not significant. Intercept = 4.788154, b= -0.005 CI[-0.020 0.010], t = -0.608 p = .543, F(1, 823)=0.3699; p=0.54; adjusted r squared= -0.001 [0, 0.9]

Plot

***

14. Growth Mindset Measure by Estimated Hours of Cram School Regression

Results

For students who attended cram school, an approximate measure of the hours they likely attended was calculated. The following results are only for those who attended cram school. The results look to be almost totally random, wonder if the calculation you used made any sense?

The regression results for estimated hours of cram school attendance and Growth Mindset total were not significant. Intercept = 39.07713, b = -0.0004 CI[-0.005, 0.004], t = -0.189 p = .85, F(1, 599)=0.03568 p = .85; adjusted r squared = -0.00161; unable to obtain a CI for this.

Plot

15. Fixed Mindset Measure by Estimated Hours of Cram School Regression

Results

For students who attended cram school, an approximate measure of the hours they likely attended was calculated. The following results are only for those who attended cram school. The results look to be almost totally random, wonder if the calculation you used made any sense?

The regression results for estimated hours of cram school attendance and Fixed Mindset total were not significant. Intercept = 30.1822, b = -0.0006 CI[-0.004, 0.003], t = -0.336 p = .737, F(1, 599)=0.1126 p = .73; adjusted r squared = -0.001481 - cannot calculate a CI for this

Plot

17. Student Year by Growth Mindset ANOVA

Results

In contrasting what year students are in and their growth mindsets, we used a 1 way ANOVA and found F(2, 822)=0.6458; p=0.52; (year 1&3): d=0.09; r=0.04; (year 2&3): d=0.01; r=0.01. So, although growth mindset increases year by year it did not appear significant.

Year	n	mean	sd	median	se
1	313	38.56	8.2	39	0.46
2	400	39.21	8.25	40	0.41
3	112	39.31	8.64	41	0.82

## 
## Call:
## lm(formula = Total.growth ~ Year, data = collapseYear)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -30.2125  -5.2125   0.6875   5.7875  15.4377 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  38.5623     0.4682  82.369   <2e-16 ***
## Year2         0.6502     0.6251   1.040    0.299    
## Year3         0.7502     0.9120   0.823    0.411    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 8.283 on 822 degrees of freedom
## Multiple R-squared:  0.001569,   Adjusted R-squared:  -0.0008603 
## F-statistic: 0.6458 on 2 and 822 DF,  p-value: 0.5245

## 
##   Simultaneous Tests for General Linear Hypotheses
## 
## Multiple Comparisons of Means: Tukey Contrasts
## 
## 
## Fit: lm(formula = Total.growth ~ Year, data = collapseYear)
## 
## Linear Hypotheses:
##            Estimate Std. Error t value Pr(>|t|)
## 2 - 1 == 0   0.6502     0.6250   1.040    0.546
## 3 - 1 == 0   0.7502     0.9120   0.823    0.685
## 3 - 2 == 0   0.1000     0.8855   0.113    0.993
## (Adjusted p values reported -- single-step method)

## 
##   Simultaneous Confidence Intervals
## 
## Multiple Comparisons of Means: Tukey Contrasts
## 
## 
## Fit: lm(formula = Total.growth ~ Year, data = collapseYear)
## 
## Quantile = 2.3355
## 95% family-wise confidence level
##  
## 
## Linear Hypotheses:
##            Estimate lwr     upr    
## 2 - 1 == 0  0.6502  -0.8096  2.1100
## 3 - 1 == 0  0.7502  -1.3797  2.8801
## 3 - 2 == 0  0.1000  -1.9680  2.1680

## 
##  Pairwise comparisons using t tests with pooled SD 
## 
## data:  collapseYear$Total.growth and collapseYear$Year 
## 
##   1   2  
## 2 0.9 -  
## 3 1.0 1.0
## 
## P value adjustment method: bonferroni

Plot

18. Student Year by Fixed Mindset ANOVA

Results

However, the rate at which fixed mindsets decreased did appear significant. F(2, 822)=4.213; p=0.015; ES(year 1&3): d=-0.3; r=-0.13; ES(year 2&3): d=-0.29; r=-0.12.
***

Year	n	mean	sd	median	se
1	313	30.27	7.03	31	0.4
2	400	30.24	6.98	30	0.35
3	112	28.2	6.88	27	0.65

## 
## Call:
## lm(formula = Total.fixed ~ Year, data = collapseYear)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -21.2425  -5.1964  -0.2425   4.7575  22.7575 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept) 30.27476    0.39477  76.690  < 2e-16 ***
## Year2       -0.03226    0.52705  -0.061  0.95121    
## Year3       -2.07833    0.76900  -2.703  0.00702 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 6.984 on 822 degrees of freedom
## Multiple R-squared:  0.01015,    Adjusted R-squared:  0.007739 
## F-statistic: 4.213 on 2 and 822 DF,  p-value: 0.01512

## 
##   Simultaneous Tests for General Linear Hypotheses
## 
## Multiple Comparisons of Means: Tukey Contrasts
## 
## 
## Fit: lm(formula = Total.fixed ~ Year, data = collapseYear)
## 
## Linear Hypotheses:
##            Estimate Std. Error t value Pr(>|t|)  
## 2 - 1 == 0 -0.03226    0.52705  -0.061   0.9979  
## 3 - 1 == 0 -2.07833    0.76900  -2.703   0.0186 *
## 3 - 2 == 0 -2.04607    0.74663  -2.740   0.0166 *
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## (Adjusted p values reported -- single-step method)

## 
##   Simultaneous Confidence Intervals
## 
## Multiple Comparisons of Means: Tukey Contrasts
## 
## 
## Fit: lm(formula = Total.fixed ~ Year, data = collapseYear)
## 
## Quantile = 2.3367
## 95% family-wise confidence level
##  
## 
## Linear Hypotheses:
##            Estimate lwr      upr     
## 2 - 1 == 0 -0.03226 -1.26381  1.19929
## 3 - 1 == 0 -2.07833 -3.87523 -0.28144
## 3 - 2 == 0 -2.04607 -3.79071 -0.30143

## 
##  Pairwise comparisons using t tests with pooled SD 
## 
## data:  collapseYear$Total.fixed and collapseYear$Year 
## 
##   1     2    
## 2 1.000 -    
## 3 0.021 0.019
## 
## P value adjustment method: bonferroni

Plot

19. Growth Mindset and Contact Details t-test

Results

It was predicted that Students with a higher growth mindset would leave email. Using a 2-sample t-test we found: difference in means: -2.017 CI[-3.7135, -0.3200], t(143.09) = -2.35, p = 0.02, r = 0.193

Contact Details	n	mean	sd	median	se
0	715	38.71	8.23	39	0.31
1	110	40.73	8.4	41.5	0.8

Plot

***

20. Fixed Mindset and Contact Details t-test

Results

However there seemed to be a very small relationship between fixed mindsets and those who left contact addresses. mean difference = 1.35804 [-0.098, 2.814], t(141.96)= 1.844; p = 0.067; ES(r)=0.153 (very small size)

Contact Details	n	mean	sd	median	se
0	715	30.16	6.97	30	0.26
1	110	28.8	7.22	29	0.69

Plot

21. Correlation Between Growth & Fixed Mindsets

Results

Using Spearman rank correlation we found a relationship between growth and fixed mindsets, however it seemed quite small. r = -0.33 (quite low, neg. cor.) (check coloured pie chart for one-to-one detail; polychoric correlation coefficients reported).

22. TOEIC Score by Estimated Hours of Cram School Regression

Results

The regression results for TOEIC Score and estimated hours of Juku were not statistically significant. Intercept = 506.51, b= 0.05, CI[-0.04, 0.14], t = -1.05 p = .294, F(1, 437)=1.103; p = .29; adjusted r squared= -0.0002 cannot calculate CI for effect size

Plot

23. Self-rated Self-rated Proficiency by Estimated Hours of Cram School Regression

Results

The regression results for Self-rated proficiency and estimated hours of Cram School were not statistically significant. Intercept = 4.57, b = 0.0004, CI[-0.0004 0.0012], t = 0.901 p = .368, F(1, 599) = .81; p = .3678; adjusted r squared= -0.0003129 cannot calculate CI for effect size

Plot

Tables for print article

## 
## 
## Table 1 
## 
## Regression results using collapseYear$TOEIC as the criterion
##  
## 
##                  Predictor        b         b_95%_CI beta  beta_95%_CI sr2
##                (Intercept) 404.25** [342.67, 465.83]                      
##  collapseYear$Total.growth   3.04**     [1.51, 4.58] 0.16 [0.08, 0.24] .02
##                                                                           
##                                                                           
##                                                                           
##  sr2_95%_CI     r             Fit
##                                  
##  [.01, .05] .16**                
##                       R2 = .025**
##                   95% CI[.01,.05]
##                                  
## 
## Note. A significant b-weight indicates the beta-weight and semi-partial correlation are also significant.
## b represents unstandardized regression weights. beta indicates the standardized regression weights. 
## sr2 represents the semi-partial correlation squared. r represents the zero-order correlation.
## Square brackets are used to enclose the lower and upper limits of a confidence interval.
## * indicates p < .05. ** indicates p < .01.
## 

## 
## 
## Table 2 
## 
## Regression results using collapseYear$TOEIC as the criterion
##  
## 
##                 Predictor        b         b_95%_CI  beta   beta_95%_CI sr2
##               (Intercept) 565.13** [508.80, 621.46]                        
##  collapseYear$Total.fixed    -1.38    [-3.21, 0.46] -0.06 [-0.14, 0.02] .00
##                                                                            
##                                                                            
##                                                                            
##  sr2_95%_CI    r             Fit
##                                 
##  [.00, .02] -.06                
##                        R2 = .004
##                  95% CI[.00,.02]
##                                 
## 
## Note. A significant b-weight indicates the beta-weight and semi-partial correlation are also significant.
## b represents unstandardized regression weights. beta indicates the standardized regression weights. 
## sr2 represents the semi-partial correlation squared. r represents the zero-order correlation.
## Square brackets are used to enclose the lower and upper limits of a confidence interval.
## * indicates p < .05. ** indicates p < .01.
## 

## 
## 
## Table 3 
## 
## Regression results using collapseYear$ProfScore as the criterion
##  
## 
##                  Predictor      b     b_95%_CI beta  beta_95%_CI sr2 sr2_95%_CI
##                (Intercept) 3.33** [2.84, 3.83]                                 
##  collapseYear$Total.growth 0.03** [0.02, 0.05] 0.18 [0.11, 0.25] .03 [.01, .06]
##                                                                                
##                                                                                
##                                                                                
##      r             Fit
##                       
##  .18**                
##            R2 = .033**
##        95% CI[.01,.06]
##                       
## 
## Note. A significant b-weight indicates the beta-weight and semi-partial correlation are also significant.
## b represents unstandardized regression weights. beta indicates the standardized regression weights. 
## sr2 represents the semi-partial correlation squared. r represents the zero-order correlation.
## Square brackets are used to enclose the lower and upper limits of a confidence interval.
## * indicates p < .05. ** indicates p < .01.
## 

## 
## 
## Table 4 
## 
## Regression results using collapseYear$ProfScore as the criterion
##  
## 
##                 Predictor      b      b_95%_CI  beta   beta_95%_CI sr2
##               (Intercept) 4.79**  [4.32, 5.25]                        
##  collapseYear$Total.fixed  -0.00 [-0.02, 0.01] -0.02 [-0.09, 0.05] .00
##                                                                       
##                                                                       
##                                                                       
##  sr2_95%_CI    r             Fit
##                                 
##  [.00, .01] -.02                
##                        R2 = .000
##                  95% CI[.00,.01]
##                                 
## 
## Note. A significant b-weight indicates the beta-weight and semi-partial correlation are also significant.
## b represents unstandardized regression weights. beta indicates the standardized regression weights. 
## sr2 represents the semi-partial correlation squared. r represents the zero-order correlation.
## Square brackets are used to enclose the lower and upper limits of a confidence interval.
## * indicates p < .05. ** indicates p < .01.
## 

24. NEW Cram School attendance by TOEIC Score

Results

The mean difference of TOEIC scores of those who attended Cram School and those who did not was not statistically significantly different. Mean difference = 27.5984 [-3.338383, 58.535142], t(236.41)= 1.7575; p = 0.0801; ES(r)= 0.114 (very small effect size)

Contact Details	n	mean	sd	median	se
No	157	544.2	176.42	515	14.08
Yes	441	516.61	146.04	500	6.95

Plot

25. NEW Cram School attendance by Self-rated Proficiency

Results

The mean difference of Self-rated Self-rated Proficiencys of those who attended cram school and those who did not was not statistically significantly different. Mean difference = 0.005 [-0.2423701, 0.2530785], t(365.62) = 0.042502; p = .9661; ES(r)= 0.002 (negligable)

Contact Details	n	mean	sd	median	se
No	221	4.652	1.64	4.71	0.11
Yes	604	4.647	1.51	4.71	0.06

Plot

26. NEW Cram School attendance by Contact Details Provided

Results

Results here are as a crosstab table. Pearson’s Chi-squared test used to analyze the results X²(1, N=825) = 1.06709, p = .302. The plot is a mosaic plot - the width of the category represents the number of cases.

Plot

Cell Contents |————————-| | Count | | Row Percent | | Column Percent | | Total Percent | |————————-|

Total Observations in Table: 825

Contact Details	Not Attended	Attended	Row Total
No	196	519	715
	27.413%	72.587%	86.667%
	88.688%	85.927%
	23.758%	62.909%
———————	———–	———–	———–
Yes	25	85	110
	22.727%	77.273%	13.333%
	11.312%	14.073%
	3.030%	10.303%
———————	———–	———–	———–
Column Total	221	604	825
	26.788%	73.212%
———————	———–	———–	———–

27. NEW Year in University by Self-rated Proficiency

Results

Year	n	mean	sd	median	se
3	112	5.22	1.69	5.43	0.16
1	313	4.42	1.47	4.43	0.08
2	400	4.67	1.51	4.71	0.08

## 
## Call:
## lm(formula = ProfScore ~ Year, data = collapseYear, na.action = na.exclude)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -3.6668 -1.0954  0.0475  1.0475  4.1903 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  4.42036    0.08614  51.316  < 2e-16 ***
## Year2        0.24643    0.11501   2.143   0.0324 *  
## Year3        0.79903    0.16780   4.762 2.27e-06 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.524 on 822 degrees of freedom
## Multiple R-squared:  0.02698,    Adjusted R-squared:  0.02461 
## F-statistic: 11.39 on 2 and 822 DF,  p-value: 1.315e-05

## 
##   Simultaneous Tests for General Linear Hypotheses
## 
## Multiple Comparisons of Means: Tukey Contrasts
## 
## 
## Fit: lm(formula = ProfScore ~ Year, data = collapseYear, na.action = na.exclude)
## 
## Linear Hypotheses:
##            Estimate Std. Error t value Pr(>|t|)    
## 2 - 1 == 0   0.2464     0.1150   2.143  0.07971 .  
## 3 - 1 == 0   0.7990     0.1678   4.762  < 0.001 ***
## 3 - 2 == 0   0.5526     0.1629   3.392  0.00207 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## (Adjusted p values reported -- single-step method)

## 
##   Simultaneous Confidence Intervals
## 
## Multiple Comparisons of Means: Tukey Contrasts
## 
## 
## Fit: lm(formula = ProfScore ~ Year, data = collapseYear, na.action = na.exclude)
## 
## Quantile = 2.3358
## 95% family-wise confidence level
##  
## 
## Linear Hypotheses:
##            Estimate lwr     upr    
## 2 - 1 == 0  0.2464  -0.0222  0.5151
## 3 - 1 == 0  0.7990   0.4071  1.1910
## 3 - 2 == 0  0.5526   0.1721  0.9331

## 
##  Pairwise comparisons using t tests with pooled SD 
## 
## data:  collapseYear$ProfScore and collapseYear$Year 
## 
##   1       2     
## 2 0.0973  -     
## 3 6.8e-06 0.0022
## 
## P value adjustment method: bonferroni

Here are effect sizes for the Year by SRP data

Effect size for difference between 1st & 3rd years

## Mean Differences ES: 
##  
##  d [ 95 %CI] = -0.52 [ -0.74 , -0.3 ] 
##   var(d) = 0.01 
##   p-value(d) = 0 
##   U3(d) = 30.06 % 
##   CLES(d) = 35.59 % 
##   Cliff's Delta = -0.29 
##  
##  g [ 95 %CI] = -0.52 [ -0.74 , -0.3 ] 
##   var(g) = 0.01 
##   p-value(g) = 0 
##   U3(g) = 30.09 % 
##   CLES(g) = 35.61 % 
##  
##  Correlation ES: 
##  
##  r [ 95 %CI] = -0.22 [ -0.31 , -0.13 ] 
##   var(r) = 0 
##   p-value(r) = 0 
##  
##  z [ 95 %CI] = -0.23 [ -0.32 , -0.13 ] 
##   var(z) = 0 
##   p-value(z) = 0 
##  
##  Odds Ratio ES: 
##  
##  OR [ 95 %CI] = 0.39 [ 0.26 , 0.58 ] 
##   p-value(OR) = 0 
##  
##  Log OR [ 95 %CI] = -0.95 [ -1.34 , -0.55 ] 
##   var(lOR) = 0.04 
##   p-value(Log OR) = 0 
##  
##  Other: 
##  
##  NNT = -8.79 
##  Total N = 425

Effect size for difference between 2nd & 3rd years

## Mean Differences ES: 
##  
##  d [ 95 %CI] = -0.35 [ -0.57 , -0.14 ] 
##   var(d) = 0.01 
##   p-value(d) = 0 
##   U3(d) = 36.14 % 
##   CLES(d) = 40.1 % 
##   Cliff's Delta = -0.2 
##  
##  g [ 95 %CI] = -0.35 [ -0.56 , -0.14 ] 
##   var(g) = 0.01 
##   p-value(g) = 0 
##   U3(g) = 36.16 % 
##   CLES(g) = 40.11 % 
##  
##  Correlation ES: 
##  
##  r [ 95 %CI] = -0.15 [ -0.23 , -0.06 ] 
##   var(r) = 0 
##   p-value(r) = 0 
##  
##  z [ 95 %CI] = -0.15 [ -0.23 , -0.06 ] 
##   var(z) = 0 
##   p-value(z) = 0 
##  
##  Odds Ratio ES: 
##  
##  OR [ 95 %CI] = 0.53 [ 0.36 , 0.77 ] 
##   p-value(OR) = 0 
##  
##  Log OR [ 95 %CI] = -0.64 [ -1.03 , -0.26 ] 
##   var(lOR) = 0.04 
##   p-value(Log OR) = 0 
##  
##  Other: 
##  
##  NNT = -11.88 
##  Total N = 512

Effect size for difference between 1st & 2nd years

## Mean Differences ES: 
##  
##  d [ 95 %CI] = -0.17 [ -0.32 , -0.02 ] 
##   var(d) = 0.01 
##   p-value(d) = 0.03 
##   U3(d) = 43.35 % 
##   CLES(d) = 45.29 % 
##   Cliff's Delta = -0.09 
##  
##  g [ 95 %CI] = -0.17 [ -0.32 , -0.02 ] 
##   var(g) = 0.01 
##   p-value(g) = 0.03 
##   U3(g) = 43.36 % 
##   CLES(g) = 45.29 % 
##  
##  Correlation ES: 
##  
##  r [ 95 %CI] = -0.08 [ -0.16 , -0.01 ] 
##   var(r) = 0 
##   p-value(r) = 0.03 
##  
##  z [ 95 %CI] = -0.08 [ -0.16 , -0.01 ] 
##   var(z) = 0 
##   p-value(z) = 0.03 
##  
##  Odds Ratio ES: 
##  
##  OR [ 95 %CI] = 0.74 [ 0.56 , 0.97 ] 
##   p-value(OR) = 0.03 
##  
##  Log OR [ 95 %CI] = -0.3 [ -0.57 , -0.04 ] 
##   var(lOR) = 0.02 
##   p-value(Log OR) = 0.03 
##  
##  Other: 
##  
##  NNT = -22.97 
##  Total N = 713

Plot

28. NEW Year in University and Contact Details

Results

Results here are as a crosstab table. Pearson’s Chi-squared test used to analyze the results X²(2, N = 825) = 8.366843, p = .015. The plot is a mosaic plot - the width of the category represents the number of cases.

Cell Contents
Count
Row Percent
Column Percent
Total Percent
————————-

University Year	No Contact Details	Contact Details	Row Total
3	89	23	112
	79.464%	20.536%	13.576%
	12.448%	20.909%
	10.788%	2.788%
—————–	———–	———–	———–
1	282	31	313
	90.096%	9.904%	37.939%
	39.441%	28.182%
	34.182%	3.758%
—————–	———–	———–	———–
2	344	56	400
	86.000%	14.000%	48.485%
	48.112%	50.909%
	41.697%	6.788%
—————–	———–	———–	———–
Column Total	715	110	825
	86.667%	13.333%
—————–	———–	———–	———–

Plot

28. NEW Cram School Hours and Contact Details Provided

Results

Contacted	n	mean	sd	median	se
No	715	162.01	162.98	120	6.1
—	—	——	——	—	—–
Yes	110	162.55	163.38	160	15.58

The mean difference of cram school hours for those who did and did not leave contact details was not statistically significantly different. Mean difference = -0.5 [-33.59497, 32.53203], t(144.41) = -0.031771; p = .9747; ES(r) = 0.003 (negligable)

Plot

Here are effect sizes for the Year by TOEIC data

Effect size for difference between 1st & 3rd years

## Mean Differences ES: 
##  
##  d [ 95 %CI] = -1.8 [ -2.09 , -1.52 ] 
##   var(d) = 0.02 
##   p-value(d) = 0 
##   U3(d) = 3.56 % 
##   CLES(d) = 10.1 % 
##   Cliff's Delta = -0.8 
##  
##  g [ 95 %CI] = -1.8 [ -2.09 , -1.51 ] 
##   var(g) = 0.02 
##   p-value(g) = 0 
##   U3(g) = 3.6 % 
##   CLES(g) = 10.16 % 
##  
##  Correlation ES: 
##  
##  r [ 95 %CI] = -0.65 [ -0.71 , -0.58 ] 
##   var(r) = 0 
##   p-value(r) = 0 
##  
##  z [ 95 %CI] = -0.78 [ -0.9 , -0.66 ] 
##   var(z) = 0 
##   p-value(z) = 0 
##  
##  Odds Ratio ES: 
##  
##  OR [ 95 %CI] = 0.04 [ 0.02 , 0.06 ] 
##   p-value(OR) = 0 
##  
##  Log OR [ 95 %CI] = -3.27 [ -3.8 , -2.75 ] 
##   var(lOR) = 0.07 
##   p-value(Log OR) = 0 
##  
##  Other: 
##  
##  NNT = -5.1 
##  Total N = 279

Effect size for difference between 2nd & 3rd years

## Mean Differences ES: 
##  
##  d [ 95 %CI] = -0.93 [ -1.17 , -0.69 ] 
##   var(d) = 0.01 
##   p-value(d) = 0 
##   U3(d) = 17.58 % 
##   CLES(d) = 25.51 % 
##   Cliff's Delta = -0.49 
##  
##  g [ 95 %CI] = -0.93 [ -1.17 , -0.69 ] 
##   var(g) = 0.01 
##   p-value(g) = 0 
##   U3(g) = 17.63 % 
##   CLES(g) = 25.55 % 
##  
##  Correlation ES: 
##  
##  r [ 95 %CI] = -0.37 [ -0.45 , -0.28 ] 
##   var(r) = 0 
##   p-value(r) = 0 
##  
##  z [ 95 %CI] = -0.38 [ -0.48 , -0.29 ] 
##   var(z) = 0 
##   p-value(z) = 0 
##  
##  Odds Ratio ES: 
##  
##  OR [ 95 %CI] = 0.18 [ 0.12 , 0.28 ] 
##   p-value(OR) = 0 
##  
##  Log OR [ 95 %CI] = -1.69 [ -2.12 , -1.26 ] 
##   var(lOR) = 0.05 
##   p-value(Log OR) = 0 
##  
##  Other: 
##  
##  NNT = -6.18 
##  Total N = 415

Effect size for difference between 1st & 2nd years

## Mean Differences ES: 
##  
##  d [ 95 %CI] = -0.76 [ -0.95 , -0.58 ] 
##   var(d) = 0.01 
##   p-value(d) = 0 
##   U3(d) = 22.22 % 
##   CLES(d) = 29.43 % 
##   Cliff's Delta = -0.41 
##  
##  g [ 95 %CI] = -0.76 [ -0.95 , -0.58 ] 
##   var(g) = 0.01 
##   p-value(g) = 0 
##   U3(g) = 22.25 % 
##   CLES(g) = 29.46 % 
##  
##  Correlation ES: 
##  
##  r [ 95 %CI] = -0.35 [ -0.42 , -0.27 ] 
##   var(r) = 0 
##   p-value(r) = 0 
##  
##  z [ 95 %CI] = -0.36 [ -0.45 , -0.27 ] 
##   var(z) = 0 
##   p-value(z) = 0 
##  
##  Odds Ratio ES: 
##  
##  OR [ 95 %CI] = 0.25 [ 0.18 , 0.35 ] 
##   p-value(OR) = 0 
##  
##  Log OR [ 95 %CI] = -1.39 [ -1.73 , -1.05 ] 
##   var(lOR) = 0.03 
##   p-value(Log OR) = 0 
##  
##  Other: 
##  
##  NNT = -6.85 
##  Total N = 502

Here are effect sizes for the Year by Growth data

Effect size for difference between 1st & 3rd years

## Mean Differences ES: 
##  
##  d [ 95 %CI] = -0.09 [ -0.31 , 0.13 ] 
##   var(d) = 0.01 
##   p-value(d) = 0.41 
##   U3(d) = 46.41 % 
##   CLES(d) = 47.46 % 
##   Cliff's Delta = -0.05 
##  
##  g [ 95 %CI] = -0.09 [ -0.31 , 0.13 ] 
##   var(g) = 0.01 
##   p-value(g) = 0.41 
##   U3(g) = 46.41 % 
##   CLES(g) = 47.46 % 
##  
##  Correlation ES: 
##  
##  r [ 95 %CI] = -0.04 [ -0.13 , 0.06 ] 
##   var(r) = 0 
##   p-value(r) = 0.41 
##  
##  z [ 95 %CI] = -0.04 [ -0.14 , 0.06 ] 
##   var(z) = 0 
##   p-value(z) = 0.41 
##  
##  Odds Ratio ES: 
##  
##  OR [ 95 %CI] = 0.85 [ 0.57 , 1.26 ] 
##   p-value(OR) = 0.41 
##  
##  Log OR [ 95 %CI] = -0.16 [ -0.56 , 0.23 ] 
##   var(lOR) = 0.04 
##   p-value(Log OR) = 0.41 
##  
##  Other: 
##  
##  NNT = -41.19 
##  Total N = 425

Effect size for difference between 2nd & 3rd years

## Mean Differences ES: 
##  
##  d [ 95 %CI] = -0.01 [ -0.22 , 0.2 ] 
##   var(d) = 0.01 
##   p-value(d) = 0.91 
##   U3(d) = 49.52 % 
##   CLES(d) = 49.66 % 
##   Cliff's Delta = -0.01 
##  
##  g [ 95 %CI] = -0.01 [ -0.22 , 0.2 ] 
##   var(g) = 0.01 
##   p-value(g) = 0.91 
##   U3(g) = 49.52 % 
##   CLES(g) = 49.66 % 
##  
##  Correlation ES: 
##  
##  r [ 95 %CI] = 0 [ -0.09 , 0.08 ] 
##   var(r) = 0 
##   p-value(r) = 0.91 
##  
##  z [ 95 %CI] = 0 [ -0.09 , 0.08 ] 
##   var(z) = 0 
##   p-value(z) = 0.91 
##  
##  Odds Ratio ES: 
##  
##  OR [ 95 %CI] = 0.98 [ 0.67 , 1.43 ] 
##   p-value(OR) = 0.91 
##  
##  Log OR [ 95 %CI] = -0.02 [ -0.4 , 0.36 ] 
##   var(lOR) = 0.04 
##   p-value(Log OR) = 0.91 
##  
##  Other: 
##  
##  NNT = -299.28 
##  Total N = 512

Effect size for difference between 1st & 2nd years

## Mean Differences ES: 
##  
##  d [ 95 %CI] = -0.08 [ -0.23 , 0.07 ] 
##   var(d) = 0.01 
##   p-value(d) = 0.3 
##   U3(d) = 46.85 % 
##   CLES(d) = 47.77 % 
##   Cliff's Delta = -0.04 
##  
##  g [ 95 %CI] = -0.08 [ -0.23 , 0.07 ] 
##   var(g) = 0.01 
##   p-value(g) = 0.3 
##   U3(g) = 46.86 % 
##   CLES(g) = 47.78 % 
##  
##  Correlation ES: 
##  
##  r [ 95 %CI] = -0.04 [ -0.11 , 0.03 ] 
##   var(r) = 0 
##   p-value(r) = 0.3 
##  
##  z [ 95 %CI] = -0.04 [ -0.11 , 0.03 ] 
##   var(z) = 0 
##   p-value(z) = 0.3 
##  
##  Odds Ratio ES: 
##  
##  OR [ 95 %CI] = 0.87 [ 0.66 , 1.13 ] 
##   p-value(OR) = 0.3 
##  
##  Log OR [ 95 %CI] = -0.14 [ -0.41 , 0.13 ] 
##   var(lOR) = 0.02 
##   p-value(Log OR) = 0.3 
##  
##  Other: 
##  
##  NNT = -46.78 
##  Total N = 713

Here are effect sizes for the Year by Fixed data

Effect size for difference between 1st & 3rd years

## Mean Differences ES: 
##  
##  d [ 95 %CI] = 0.3 [ 0.08 , 0.51 ] 
##   var(d) = 0.01 
##   p-value(d) = 0.01 
##   U3(d) = 61.64 % 
##   CLES(d) = 58.29 % 
##   Cliff's Delta = 0.17 
##  
##  g [ 95 %CI] = 0.3 [ 0.08 , 0.51 ] 
##   var(g) = 0.01 
##   p-value(g) = 0.01 
##   U3(g) = 61.62 % 
##   CLES(g) = 58.28 % 
##  
##  Correlation ES: 
##  
##  r [ 95 %CI] = 0.13 [ 0.03 , 0.22 ] 
##   var(r) = 0 
##   p-value(r) = 0.01 
##  
##  z [ 95 %CI] = 0.13 [ 0.03 , 0.23 ] 
##   var(z) = 0 
##   p-value(z) = 0.01 
##  
##  Odds Ratio ES: 
##  
##  OR [ 95 %CI] = 1.71 [ 1.15 , 2.53 ] 
##   p-value(OR) = 0.01 
##  
##  Log OR [ 95 %CI] = 0.54 [ 0.14 , 0.93 ] 
##   var(lOR) = 0.04 
##   p-value(Log OR) = 0.01 
##  
##  Other: 
##  
##  NNT = 10.79 
##  Total N = 425

Effect size for difference between 2nd & 3rd years

## Mean Differences ES: 
##  
##  d [ 95 %CI] = 0.29 [ 0.08 , 0.5 ] 
##   var(d) = 0.01 
##   p-value(d) = 0.01 
##   U3(d) = 61.53 % 
##   CLES(d) = 58.21 % 
##   Cliff's Delta = 0.16 
##  
##  g [ 95 %CI] = 0.29 [ 0.08 , 0.5 ] 
##   var(g) = 0.01 
##   p-value(g) = 0.01 
##   U3(g) = 61.51 % 
##   CLES(g) = 58.2 % 
##  
##  Correlation ES: 
##  
##  r [ 95 %CI] = 0.12 [ 0.03 , 0.21 ] 
##   var(r) = 0 
##   p-value(r) = 0.01 
##  
##  z [ 95 %CI] = 0.12 [ 0.03 , 0.21 ] 
##   var(z) = 0 
##   p-value(z) = 0.01 
##  
##  Odds Ratio ES: 
##  
##  OR [ 95 %CI] = 1.7 [ 1.16 , 2.49 ] 
##   p-value(OR) = 0.01 
##  
##  Log OR [ 95 %CI] = 0.53 [ 0.15 , 0.91 ] 
##   var(lOR) = 0.04 
##   p-value(Log OR) = 0.01 
##  
##  Other: 
##  
##  NNT = 10.91 
##  Total N = 512

Effect size for difference between 1st & 2nd years

## Mean Differences ES: 
##  
##  d [ 95 %CI] = 0.3 [ 0.15 , 0.45 ] 
##   var(d) = 0.01 
##   p-value(d) = 0 
##   U3(d) = 61.71 % 
##   CLES(d) = 58.34 % 
##   Cliff's Delta = 0.17 
##  
##  g [ 95 %CI] = 0.3 [ 0.15 , 0.45 ] 
##   var(g) = 0.01 
##   p-value(g) = 0 
##   U3(g) = 61.7 % 
##   CLES(g) = 58.34 % 
##  
##  Correlation ES: 
##  
##  r [ 95 %CI] = 0.15 [ 0.07 , 0.22 ] 
##   var(r) = 0 
##   p-value(r) = 0 
##  
##  z [ 95 %CI] = 0.15 [ 0.07 , 0.22 ] 
##   var(z) = 0 
##   p-value(z) = 0 
##  
##  Odds Ratio ES: 
##  
##  OR [ 95 %CI] = 1.72 [ 1.31 , 2.25 ] 
##   p-value(OR) = 0 
##  
##  Log OR [ 95 %CI] = 0.54 [ 0.27 , 0.81 ] 
##   var(lOR) = 0.02 
##   p-value(Log OR) = 0 
##  
##  Other: 
##  
##  NNT = 10.71 
##  Total N = 713