INTRODUCTION

As Scotland has a demographic deficit, the Scottish government has been exploring the idea of introducing a social integration strategy into Scotland to create a welcoming environment for all migrants. The Scottish government claims to have an overarching aim to create a country with equality, inclusiveness, normalise racial and ethnic differences, and develop integrated communities (Trevena, 2019). Qualitative research on behalf of the Scottish government, has also identified that Scotland is a more welcoming environment than England for EU migrants. This research inferred that the older age group; and younger people in the most deprived socioeconomic groups are less welcoming than younger and middle-aged adults (Strategic Analysis Team, 2018; Trevena, 2019, p.18). To explore these phenomena, data from the British Social Survey 2019 (BSA 2019) will be used to focus on sample data of Scottish and English participants between the age of 18 and 90 to ascertain whether these qualitative reports replied upon by the Scottish government are robust, have external validity, and can be generalised across the whole population.

Therefore, to explore these phenomena, the null hypothesis - People living in England compared to Scotland are less welcoming of migration as they believe that it changes cultural life.

The alternative hypothesis: The Scottish population are more welcoming of EU migration than English people and do not perceive that migration negatively alters their culture life.

These hypothesises were chosen as, among other things, focus group participants claimed to have a fear of losing ‘traditional Scottish/local identity and culture and voiced a fear of EU migration ‘taking over’ (Trevena, 2019, p.16).

DEPENDENT VARIABLE - CULTURAL PERCEPTION

Table 1. AVERAGE CULTURAL PERCEPTION

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   1.000   5.000   6.500   6.307   8.000  10.000

MiCultur - This variable is a good fit as a dependent variable to test this hypothesis as it measures the cultural perceptions of migration in relation to cultural life (Cultural Perception). This variable has a suitable Likert Scale to be used in linear regression to measure this phenomenon as it has an overall mean of 6.362 which is greater than the average count of 10 levels and therefore should not impact the regression analysis. 0 determines that cultural life has been undermined because of migration into the UK, rising to 10 representing participants who perceive that British cultural life is enriched due to migration. However, this dependant variable has limitations as it measures whether migration undermines cultural life which is not specific to EU migration. Nevertheless, it will measure the participants’ view of migration and the impact they believe that it has on their cultural life.

MISSINGNESS

Although the BSA 2019 indicates that all its data is meaningful, observations were removed from the analysis for participants who refused to answer or did not know the answer as their responses will create data to test the null hypothesis. Therefore, participants who refused to answer ‘99’, or did not know the answer ‘98’, or were not applicable ‘-1’, ‘-2’, and NA have been removed throughout the whole analysis.

Table 2. PERCENTATE OF POPULATION BY CULTURAL PERCEPTION

##    Cultural_Perception   n percent
## 1           Undermined  17    3.16
## 2                    2  23    4.28
## 3                    3  31    5.76
## 4                    4  36    6.69
## 5              Neither 116   21.56
## 6                    6  46    8.55
## 7                    7  78   14.50
## 8                    8  90   16.73
## 9                    9  39    7.25
## 10            Enriched  62   11.52

Table 3.

## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

The table of proportions and histogram demonstrates that out of the entire sample set of 3,224, 538 participants will be analysed. Levels 0 to 4 of this observation represent 19.88% of the participants who perceive that migration undermines cultural life. However, 21.56% neither believed nor disbelieve that cultural life is undermined or enriched. In contrast, 58.55% of the sample set perceive that that migration enriches British cultural life. These levels of perception vary between levels 7,8, 10, who are twice as likely to perceive cultural enrichment due to migration compared to the levels of enrichment in levels 6 and 9.

In order to understand the demographics of the total population in the sample set, this paper will take a specific to general approach to analyse independent variables highlighted in the Scottish governments reports.

COUNTRY

Firstly, we will analyse the IV for country. Wales will not be identified in the paper to focus on the difference in the populations’ perceptions of the impact on cultural life due to migration between Scotland and England.

Table 4. PERCENTAGE OF POPULATION BY COUNTRY

##   Country1   n percent
## 1  England 483   89.78
## 2 Scotland  55   10.22

89.78% of the sample are resident in England and 10.22% in Scotland. The BSA, 2019 stipulates that its data is a representative survey of adults over 18+ living in the UK. The BSA, 2019 research only surveys private households. Additionally it worth noting that participants living North of the Caledonian Canal in Scotland are also excluded from this sample set.

Table 5. JITTER PLOT The Likert scale for Scotland indicates that fewer people believe that migration undermines cultural life. Most of the dots appear in the category that neither has a belief nor disbelieves that migration impacts cultural life, which is consistent with the analysis of the DV in Tables 2 and 3. The observations for Scotland in levels 1, 3, and 4 are so few; although valid observations, they may appear as outliers during later regressions analysis. Overall, for Scotland, most of the observations indicate a perception that migration enriches cultural life.

England’s observations are consistent with the DV and Scotland’s observations, indicating a neutral perception in category 5. Levels 7, 8, and 10 are again more prevalent compared to 6 and 9. The observations for England are also consistent with the DV pattern identified in Tables 2 and 3. Although the Jitter plot is not a percentage analysis, the overall observations indicate that England believes that migration enriches cultural life. However, the jitter pattern in level one indicates that more people in England believe that cultural life is undermined by migration when compared to Scotland. However, to test if these observations are statistically significant, we will continue our analysis with a t-test to test the null hypothesis.

H0:μEngland=μScotland

People living in England compared to Scotland are less welcoming of EU migration as they believe that it changes cultural life.

HA:μEngland<μScotland

As there is an expectation in the direction of the mean in this analysis - that the Scottish population are more welcoming of EU migration than English people and do not perceive that migration negatively alters their culture life; a one-tail test is used.

Table 6. BIVARIATE LINEAR REGRESSION

## 
## Call:
## lm(formula = MiCultur ~ Country1, data = BSA2019subset1)
## 
## Residuals:
##    Min     1Q Median     3Q    Max 
## -5.333 -1.333  0.297  1.667  3.927 
## 
## Coefficients:
##                  Estimate Std. Error t value Pr(>|t|)    
## (Intercept)        6.3333     0.1087  58.263   <2e-16 ***
## Country1Scotland  -0.2606     0.3400  -0.767    0.444    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 2.389 on 536 degrees of freedom
## Multiple R-squared:  0.001095,   Adjusted R-squared:  -0.0007686 
## F-statistic: 0.5876 on 1 and 536 DF,  p-value: 0.4437

This bivariate linear regression tests that the actual value of the estimate is 0. We reject the null hypothesis as the p-value is less than 0.05; there is a difference.

Data indicates that the DV is interpreting the coefficient country relative to IV level for England.

The range between England and Scotland in relation to their cultural perception that migration changes cultural life varies based on a normal distribution of 68% that the observations will fall between (6.3333-0.1087) = 6.2246 to (6.3333+0.1087) = 6.442 and the standard deviation is 0.1087 disbursements around the mean (Foggarty, p.122).

The mean-variance between England and Scotland is relatively small -0.2606. The mean perception of the impact of migration on cultural life in England is 6.3333 compared to Scotland’s 6.0727. The t-test indicates that people living in Scotland are 6.3333/2.389= 2.6510255 less likely to believe that migration will change cultural life. The DV cultural perception in Table 1. mean was 6.362, which indicates that when accounting for country both the IV and DV are close to this mean. Although, surprisingly, the mean falls within level 6 when accounting for the variance for these countries rather than level 5, which was more prominent on the DV histogram and jitter plot.

Even though the sample size is small, the degree of freedom tested n 536 -1 interpreted whether this is a representative sample. Hence, the F test ascertains if the variation among these two groups is significant. The P test value for the F test indicates 0.4437 is not statistically significant. If we draw these samples again, the outcome might vary, which is not generalisable to the population. Additionally, the R-squared indicates that 0% of the variation is explained between the DV and IV.

REGION

Table 7.

##             UK_Regions  n   percent
## 1           North East 31  6.418219
## 2           North West 65 13.457557
## 3 Yorkshire and Humber 58 12.008282
## 4        East Midlands 43  8.902692
## 5        West Midlands 41  8.488613
## 6      East of England 50 10.351967
## 7               London 45  9.316770
## 8           South East 84 17.391304
## 9           South West 66 13.664596

Analysis of these regions indicates that there is not an even distribution of participants throughout the United Kingdom. Therefore when interpreting the data it is important to bear in mind that the North East of England is highly underrepresented. For example, Jarrow in Newcastle Upon Tyne in particular was one of the areas targeted by UKIP which had a majority vote to leave the EU.

MULTIPLE LINEAR REGRESSION

Having accounted for England and Scotland’s small variations we will now test the null hypothesis accounting for these regionalised demographics.

Table 8.

## 
## Call:
## lm(formula = MiCultur ~ Country + UK_Regions, data = BSA2019subset1)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -6.1786 -1.5200  0.0233  1.8245  4.7419 
## 
## Coefficients: (1 not defined because of singularities)
##                                Estimate Std. Error t value Pr(>|t|)    
## (Intercept)                      5.2581     0.4239  12.403  < 2e-16 ***
## Country                              NA         NA      NA       NA    
## UK_RegionsNorth West             0.5573     0.5152   1.082 0.279902    
## UK_RegionsYorkshire and Humber   0.9143     0.5251   1.741 0.082301 .  
## UK_RegionsEast Midlands          0.7187     0.5561   1.292 0.196880    
## UK_RegionsWest Midlands          0.8883     0.5618   1.581 0.114500    
## UK_RegionsEast of England        1.2619     0.5396   2.339 0.019761 *  
## UK_RegionsLondon                 1.4753     0.5509   2.678 0.007667 ** 
## UK_RegionsSouth East             1.9205     0.4960   3.872 0.000123 ***
## UK_RegionsSouth West             1.0904     0.5139   2.122 0.034378 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 2.36 on 474 degrees of freedom
##   (55 observations deleted due to missingness)
## Multiple R-squared:  0.04714,    Adjusted R-squared:  0.03106 
## F-statistic: 2.931 on 8 and 474 DF,  p-value: 0.00332

The analysis highlights assumption 3, multicollinearity between the variables, undermining the significance of these analyses as there is a direct correlation between the variables. One way to circumnavigate this problem could be to load regional data from another BSA data set or similar, but due to time constraints, Region will be removed from this analysis to avoid violating this assumption.

AGE

Analysis of the Scottish government also indicates that age is a significant factor in relation to the impact that migration has on British cultural life. In this sample set the minimum age is 18 and the maximum age range is 99 with a mean of 53.06 years of age.

Table 9.

##     Age   n percent
## 1 18-24  26    4.83
## 2 25-34  75   13.94
## 3 35-44  83   15.43
## 4 45-54 102   18.96
## 5 55-64  80   14.87
## 6 65-97 172   31.97
##   mean_age minimum_age maximum_age
## 1 52.87918          18          90
##   Age_Variance Age_Standard_Deviation
## 1     300.8774               17.34582

31.97% of the population are in the higher age group. This is interesting as Trevena’s (2019), analysis infers that the older aged group were in a majority that opposed migration than middle aged and younger people. This variable will be added to expand this analysis and test the null hypothesis.

The mean age in the sample set is 52.87918, and the standard deviation is 17.34582 disbursements around the mean. Therefore, ‘based on a normal distribution 68% of the observations should lie between 35.53 (52.87-17.34) and (52.87+17.34) = 70.21, exemplified clearly within the density plot above visualising that the observations relative to age have a normal distribution.

Table 10. The box plot in table 10. indicates that the participants who were least likely to perceive that migration undermined cultural life across the United Kingdom in Levels 0 to 4, the first percentile is 35 years and the third percentile 70, with a mean of 65. In comparison, the class with a neutral perception, the first percentile was 43 years of age with a third percentile of 65 and a mean of 55. In the first percentile, participants in groups 6 and 9 are an approximate age of 39 years with equal third percentiles of 65 years of age with a mean of 57 and 47, respectively. Finally, those who perceived the most enrichment of cultural life in groups 7, 8, and 10, the first percentile range was from 35 years to 65 with a mean age of 45 to 53. Thus, for the whole UK, the group in the dominant level, 5, has a mean age of 55. Although those in the younger class with a mean age of 48 years perceived an enrichment of cultural life due to migration, whereas the older age group with a mean age of 65 perceived that migration undermined cultural life, which as you can see from the box plot below the average age in Scotland is 50.

The outliers visualised in groups 4,5 and 6 between the age of 18 to 22 and 70+ will not be removed as all the participants are integral to the Scottish government’s reports.

Table 11. Table 12.

## 
## Call:
## lm(formula = MiCultur ~ Country + RageE, data = BSA2019subset1)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -6.0198 -1.5446  0.1162  1.7855  4.5644 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  8.02455    0.49584  16.184  < 2e-16 ***
## Country     -0.29181    0.33404  -0.874    0.383    
## RageE       -0.02640    0.00584  -4.522 7.57e-06 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 2.347 on 535 degrees of freedom
## Multiple R-squared:  0.03786,    Adjusted R-squared:  0.03426 
## F-statistic: 10.53 on 2 and 535 DF,  p-value: 3.281e-05

When accounting for age in addition to the country, we reject the null hypothesis as the p-value is again less than 0.05. This multi-linear regression also on the face of it supports the qualitative evidence of Trevena (2019), that age is a highly significant factor when accounting for a person’s cultural perception in England and Scotland concerning migration.

This variance altered the mean for England of 6.3333when accounting for country. Calling for age has increased this mean to 8.02455. When calling country, the mean for Scotland was 6.0727, which increased to 7.73274 when also accounting for the age variable; there iss a slight increase in the variation between the IV and DV from -0.2606 to -0.29181.

Even though the sample size is small, the F test ascertained if the variation among the two groups is statistically significant. If we draw these samples again, the outcome is likely to be the same and therefore it is generalisable to the population providing the data meets the statistical assumptions. Additionally, the R-squared indicates that 3% of the variation is explained between the DV and IV. As this is extremely low further independent variables will be added to this model.

Table 13.

##      Sex   n  percent
## 1   Male 248 46.09665
## 2 Female 290 53.90335

Table 14.

## 
## Call:
## lm(formula = MiCultur ~ Country + Age + Sex, data = BSA2019subset1)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -6.2186 -1.6457  0.1783  1.7024  4.7098 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  6.69795    0.62004  10.802   <2e-16 ***
## Country     -0.20583    0.33590  -0.613   0.5403    
## Age25-34     0.02021    0.53393   0.038   0.9698    
## Age35-44     0.40077    0.52746   0.760   0.4477    
## Age45-54    -0.17215    0.51576  -0.334   0.7387    
## Age55-64    -0.35975    0.53109  -0.677   0.4985    
## Age65-97    -0.99612    0.49321  -2.020   0.0439 *  
## SexFemale    0.32569    0.20366   1.599   0.1104    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 2.341 on 530 degrees of freedom
## Multiple R-squared:  0.05132,    Adjusted R-squared:  0.03879 
## F-statistic: 4.096 on 7 and 530 DF,  p-value: 0.0002151

There are a total of 538 people in this sample, 53% female and 46% male. Sex is not a statistically significant variable, although we still reject the the null hypothesis. The intercept for England has reduced from 8.02455 when accounting for country and age to 6.69795 when also account for the IV sex. 5% of the variation is now accounted for although the F statistic continues to remain statistically significant if we were to draw the same variables again. However, the government reports do not not indicate that Gender is a determinant but it does infers that people in socially deprived areas are more prone to opposing migration.

Table 15.

## 
## Call:
## lm(formula = MiCultur ~ Country + Age + Sex + HH_Monthly_Income, 
##     data = BSA2019subset1)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -5.8735 -1.5344  0.1167  1.8528  4.5922 
## 
## Coefficients:
##                                Estimate Std. Error t value Pr(>|t|)    
## (Intercept)                     5.98846    0.68261   8.773  < 2e-16 ***
## Country                        -0.07191    0.33120  -0.217  0.82820    
## Age25-34                       -0.02377    0.52915  -0.045  0.96418    
## Age35-44                        0.08688    0.53014   0.164  0.86988    
## Age45-54                       -0.37981    0.51257  -0.741  0.45903    
## Age55-64                       -0.33438    0.52733  -0.634  0.52629    
## Age65-97                       -0.77349    0.49213  -1.572  0.11662    
## SexFemale                       0.33661    0.19943   1.688  0.09204 .  
## HH_Monthly_Income£921-1,240    -0.42535    0.45320  -0.939  0.34840    
## HH_Monthly_Income£1,241-1,590  -0.33207    0.45426  -0.731  0.46509    
## HH_Monthly_Income£1,591-2,030   0.40359    0.43981   0.918  0.35922    
## HH_Monthly_Income£2,031-2,560   0.42023    0.40790   1.030  0.30338    
## HH_Monthly_Income£2,561-3,160   0.87003    0.43434   2.003  0.04568 *  
## HH_Monthly_Income£3,161-3,890   0.75849    0.42928   1.767  0.07783 .  
## HH_Monthly_Income£3,891-4,4910  1.26696    0.44169   2.868  0.00429 ** 
## HH_Monthly_Income£4,911-6,610   0.99072    0.49396   2.006  0.04541 *  
## HH_Monthly_Income£6,611-plus    1.53816    0.42830   3.591  0.00036 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 2.286 on 521 degrees of freedom
## Multiple R-squared:  0.1113, Adjusted R-squared:  0.08398 
## F-statistic: 4.077 on 16 and 521 DF,  p-value: 2.157e-07

We continue to reject the null hypothesis when we account for a person’s income. A person’s monthly income increases the R-squared and explains 11 percent of the variation between England and Scotland. The mean for England has decreased again from 6.69795 to 5.98846, which indicates a neutral perception of migration on cultural life. Scotland is slightly lower at -0.07191 below this intercept at 5.91655. The intercept is in relation to the lowest earners with income below £921. This analysis does not appear to support the government’s report, which infers that teenagers in socially deprived areas are more prone to oppose migration. Participants earning £921-1,240 also have a neutral perception, although people in the highest group increase to 7.604119, indicating that the higher earners believe that migration enriches cultural life. Unfortunately, this IV that reflects the monthly income cannot be analysed alongside the IV which identifies the source of this income due to multicollinearity.

Table 16.

## 
## Call:
## lm(formula = MiCultur ~ Country + Age + Sex + Employment_Status, 
##     data = BSA2019subset1)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -6.2191 -1.4946  0.0722  1.7809  5.2460 
## 
## Coefficients:
##                                       Estimate Std. Error t value Pr(>|t|)    
## (Intercept)                            6.76320    0.64192  10.536  < 2e-16 ***
## Country                               -0.10979    0.33419  -0.329 0.742652    
## Age25-34                               0.01839    0.54855   0.034 0.973274    
## Age35-44                               0.27437    0.55369   0.496 0.620430    
## Age45-54                              -0.06511    0.54075  -0.120 0.904206    
## Age55-64                              -0.47125    0.56966  -0.827 0.408478    
## Age65-97                              -1.19393    0.62738  -1.903 0.057591 .  
## SexFemale                              0.29131    0.20065   1.452 0.147158    
## Employment_StatusOccupational Pension  0.74382    0.41994   1.771 0.077108 .  
## Employment_StatusPrivate Pension      -0.51070    0.69107  -0.739 0.460247    
## Employment_StatusState Pension        -0.30010    0.45799  -0.655 0.512585    
## Employment_StatusIncome Support       -0.75804    1.62460  -0.467 0.640982    
## Employment_StatusPension Credit       -5.18215    2.29901  -2.254 0.024609 *  
## Employment_StatusDisability Benefit   -0.80984    0.68094  -1.189 0.234872    
## Employment_StatusPIP                  -3.73956    1.33243  -2.807 0.005196 ** 
## Employment_StatusUniversal Credit     -1.83432    0.51932  -3.532 0.000449 ***
## Employment_StatusState or Tax Credit  -0.42136    1.62363  -0.260 0.795342    
## Employment_StatusSavings & Investment -1.24797    0.95517  -1.307 0.191947    
## Employment_StatusStudent               1.50619    1.07137   1.406 0.160366    
## Employment_StatusParents              -0.29906    1.68634  -0.177 0.859307    
## Employment_StatusOther                 0.81399    0.74567   1.092 0.275508    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 2.278 on 517 degrees of freedom
## Multiple R-squared:  0.1236, Adjusted R-squared:  0.0897 
## F-statistic: 3.646 on 20 and 517 DF,  p-value: 2.298e-07

Focusing on the source of income IV without the monthly income IV indicates that people with an occupational pension are not quite significant. However, the population reliant on pension credits has a -5.18215 statistical significance and a much lower perception of migration than employed people. Hence, English people on pension credits have a rating of 1.58105. Equally, people on PIP and Universal Credit believe that migration undermines cultural life. Even though Scotland’s levels only differ by -0.10979, this outcome still supports the Scottish government’s analysis that the socially deprived people in both countries have a lower tolerance of migration in relation to cultural life Trevena (2019, p.15).

Table 17. This density plot highlights whether the participants think that migrants should be allowed to work freely within the UK if British firms can sell to the EU. Introducing this variable as opposed to IV focusing on the personal prejudice of the participants avoids incorporating an exogenous IV and violating assumption 4. For both countries, the impact of migration on cultural life between level 1 and 2 in England extending to level 4 in Scotland, this class of people were against EU migration even if there was a trade benefit. This class was also identified in tables 15 and 16 above as belonging to the most deprived groups in society. In England, there is a relationship in levels 7 and 10 of the impact of migration on cultural life and those favouring migration when there is a trade relationship. Previous analysis has indicated that this group represents the highest earners although less prevalent in Scotland the levels 4 to 10 with a peak at 8 who are in favour of migration when there is a trade benefit.

Table 18.

## 
## Call:
## lm(formula = MiCultur ~ Country + Age + Sex + HH_Monthly_Income + 
##     Personal_View_of_EU_Migration, data = BSA2019subset1)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -6.0108 -1.4403  0.1703  1.4856  6.0575 
## 
## Coefficients:
##                                                      Estimate Std. Error
## (Intercept)                                           4.60933    0.68326
## Country                                              -0.22151    0.30223
## Age25-34                                              0.16635    0.48261
## Age35-44                                              0.22282    0.48346
## Age45-54                                              0.04079    0.46876
## Age55-64                                              0.15445    0.48270
## Age65-97                                             -0.42839    0.44974
## SexFemale                                             0.31258    0.18302
## HH_Monthly_Income£921-1,240                          -0.35865    0.41335
## HH_Monthly_Income£1,241-1,590                        -0.32862    0.41432
## HH_Monthly_Income£1,591-2,030                         0.08288    0.40194
## HH_Monthly_Income£2,031-2,560                         0.20458    0.37223
## HH_Monthly_Income£2,561-3,160                         0.57269    0.39746
## HH_Monthly_Income£3,161-3,890                         0.30544    0.39444
## HH_Monthly_Income£3,891-4,4910                        0.85179    0.40473
## HH_Monthly_Income£4,911-6,610                         0.42494    0.45377
## HH_Monthly_Income£6,611-plus                          1.02652    0.39355
## Personal_View_of_EU_Migration2. Probably Against      0.18433    0.37902
## Personal_View_of_EU_Migration3. Probably In Favour    1.30361    0.33664
## Personal_View_of_EU_Migration4. Definately In Favour  2.59513    0.33525
##                                                      t value Pr(>|t|)    
## (Intercept)                                            6.746 4.07e-11 ***
## Country                                               -0.733 0.463941    
## Age25-34                                               0.345 0.730471    
## Age35-44                                               0.461 0.645076    
## Age45-54                                               0.087 0.930694    
## Age55-64                                               0.320 0.749116    
## Age65-97                                              -0.953 0.341272    
## SexFemale                                              1.708 0.088258 .  
## HH_Monthly_Income£921-1,240                           -0.868 0.385982    
## HH_Monthly_Income£1,241-1,590                         -0.793 0.428054    
## HH_Monthly_Income£1,591-2,030                          0.206 0.836716    
## HH_Monthly_Income£2,031-2,560                          0.550 0.582819    
## HH_Monthly_Income£2,561-3,160                          1.441 0.150223    
## HH_Monthly_Income£3,161-3,890                          0.774 0.439070    
## HH_Monthly_Income£3,891-4,4910                         2.105 0.035809 *  
## HH_Monthly_Income£4,911-6,610                          0.936 0.349478    
## HH_Monthly_Income£6,611-plus                           2.608 0.009360 ** 
## Personal_View_of_EU_Migration2. Probably Against       0.486 0.626940    
## Personal_View_of_EU_Migration3. Probably In Favour     3.872 0.000122 ***
## Personal_View_of_EU_Migration4. Definately In Favour   7.741 5.22e-14 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 2.082 on 518 degrees of freedom
## Multiple R-squared:  0.2667, Adjusted R-squared:  0.2398 
## F-statistic: 9.916 on 19 and 518 DF,  p-value: < 2.2e-16

When we account for migration and trade in addition to migration and cultural life, we reject the null hypothesis. However, the mean has dropped to 4.60933 and an even lower tolerance of migration in Scotland of 4.38782. The high earners have a higher Likert score than anticipated, but a rating of 5.63585 does not correlate with the density plot in table 17. In addition, age is no longer a statistically significant variable. Furthermore, in relation to the intercept representing the class of people who were against migration even if there is a trade benefit, the class favouring migration increased between 1.30361 and 2.59513 above the mean. The R-squared explains 26% of the variation, and the F test confirms that this is a representative sample.

Table 19.

## 
## Call:
## lm(formula = MiCultur ~ Country + Age + Sex + HH_Monthly_Income + 
##     Personal_View_of_EU_Migration + Employment_Perception, data = BSA2019subset1)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -6.7027 -1.2379  0.1161  1.2133  5.1464 
## 
## Coefficients:
##                                                      Estimate Std. Error
## (Intercept)                                           2.41327    0.89353
## Country                                              -0.18923    0.27435
## Age25-34                                              0.01629    0.44259
## Age35-44                                             -0.08024    0.44347
## Age45-54                                              0.15434    0.42710
## Age55-64                                             -0.05326    0.43988
## Age65-97                                             -0.52439    0.40867
## SexFemale                                             0.17586    0.16687
## HH_Monthly_Income£921-1,240                          -0.29134    0.37870
## HH_Monthly_Income£1,241-1,590                        -0.39315    0.37946
## HH_Monthly_Income£1,591-2,030                         0.13764    0.36683
## HH_Monthly_Income£2,031-2,560                         0.14781    0.34042
## HH_Monthly_Income£2,561-3,160                         0.64955    0.36423
## HH_Monthly_Income£3,161-3,890                         0.34308    0.35843
## HH_Monthly_Income£3,891-4,4910                        0.74108    0.36969
## HH_Monthly_Income£4,911-6,610                         0.12629    0.41632
## HH_Monthly_Income£6,611-plus                          0.80634    0.36030
## Personal_View_of_EU_Migration2. Probably Against      0.01273    0.34850
## Personal_View_of_EU_Migration3. Probably In Favour    0.70818    0.31641
## Personal_View_of_EU_Migration4. Definately In Favour  1.61739    0.32023
## Employment_Perception2                                1.42386    0.81268
## Employment_Perception3                                1.39745    0.76316
## Employment_Perception4                                2.29723    0.82012
## Employment_PerceptionNutural                          2.61161    0.70632
## Employment_Perception6                                3.19536    0.74309
## Employment_Perception7                                3.51649    0.72897
## Employment_Perception8                                4.03804    0.74252
## Employment_Perception9                                4.86970    0.80741
## Employment_PerceptionCreates Jobs                     4.99649    0.79622
##                                                      t value Pr(>|t|)    
## (Intercept)                                            2.701 0.007147 ** 
## Country                                               -0.690 0.490662    
## Age25-34                                               0.037 0.970657    
## Age35-44                                              -0.181 0.856494    
## Age45-54                                               0.361 0.717969    
## Age55-64                                              -0.121 0.903681    
## Age65-97                                              -1.283 0.200017    
## SexFemale                                              1.054 0.292448    
## HH_Monthly_Income£921-1,240                           -0.769 0.442055    
## HH_Monthly_Income£1,241-1,590                         -1.036 0.300668    
## HH_Monthly_Income£1,591-2,030                          0.375 0.707661    
## HH_Monthly_Income£2,031-2,560                          0.434 0.664329    
## HH_Monthly_Income£2,561-3,160                          1.783 0.075126 .  
## HH_Monthly_Income£3,161-3,890                          0.957 0.338926    
## HH_Monthly_Income£3,891-4,4910                         2.005 0.045533 *  
## HH_Monthly_Income£4,911-6,610                          0.303 0.761754    
## HH_Monthly_Income£6,611-plus                           2.238 0.025657 *  
## Personal_View_of_EU_Migration2. Probably Against       0.037 0.970873    
## Personal_View_of_EU_Migration3. Probably In Favour     2.238 0.025641 *  
## Personal_View_of_EU_Migration4. Definately In Favour   5.051 6.14e-07 ***
## Employment_Perception2                                 1.752 0.080366 .  
## Employment_Perception3                                 1.831 0.067663 .  
## Employment_Perception4                                 2.801 0.005287 ** 
## Employment_PerceptionNutural                           3.697 0.000241 ***
## Employment_Perception6                                 4.300 2.05e-05 ***
## Employment_Perception7                                 4.824 1.86e-06 ***
## Employment_Perception8                                 5.438 8.37e-08 ***
## Employment_Perception9                                 6.031 3.13e-09 ***
## Employment_PerceptionCreates Jobs                      6.275 7.47e-10 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.886 on 509 degrees of freedom
## Multiple R-squared:  0.4085, Adjusted R-squared:  0.376 
## F-statistic: 12.56 on 28 and 509 DF,  p-value: < 2.2e-16

To understand the motivation for the demographics, we will add an IV which identifies that most of the population fear that migration is threatening jobs (Trevena, 2019, p.15). This IV has altered the mean from 4.60933 to 2.41327 , and the variation with Scotland’s view is -0.18923 . The relationship between the DV and those that earn a higher monthly income and those favouring migration when there are trade benefits is still statistically significant. However, most pertinent variation between the DV and IV is the association for those who believe that migration will threaten their jobs which is statistically significant throughout the whole population. Including this variable explains nearly 40% of the variation between England and Scotland, and once again, the F statistic indicates this is a representative sample.

Table 20.

##                                   GVIF Df GVIF^(1/(2*Df))
## Country                       1.044361  1        1.021940
## Age                           1.505529  5        1.041763
## Sex                           1.046069  1        1.022775
## HH_Monthly_Income             1.644905  9        1.028035
## Personal_View_of_EU_Migration 1.388075  3        1.056174
## Employment_Perception         1.655458  9        1.028400

The low numbers confirm that multicollinearity does not exist among the variables. Furthermore, a test for cluster robust standard errors will not be carried out as the BSA 2019 stipulated that only one person in each household was interviewed.

CORRECTLY SPECIFIED and OMITTED VARIABLE BIAS - Assumption 1

The data in this model is correctly specified. We are interested in the difference between England and Scotland and their perception of the impact of migration on cultural life. The Scottish government’s qualitative reports relate to a perception that cultural life is being undermined and that Scotland is more welcoming than England. Additionally, the reports indicate that age and social deprivation also impact a negative view towards migration. Furthermore, the government reports refer to other factors such as scarcity of resources such as the NHS and housing and fear of crimes relating to migrants. Unfortunately, appropriate variables do not exist in the BSA2019; therefore, there is no omitted variable bias in relation to these narrow parameters of this study.

Outliers were previously identified in table 10 concerning participants below the age of 20. However, as IV age was integral to the government’s reports, it was not removed to avoid omitted variable bias. Although, a small number of 0 Likert scores were removed from the MiJobs to run a linear regression plot.

Table 21.

lm(MiCultur ~ Country + Age1 + Sex + HH_Inc_pm + Personal_View_of_EU_Migration + Employment_Perception2, data=BSA2019subset1) %>% summary
## 
## Call:
## lm(formula = MiCultur ~ Country + Age1 + Sex + HH_Inc_pm + Personal_View_of_EU_Migration + 
##     Employment_Perception2, data = BSA2019subset1)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -5.6415 -1.4954  0.1586  1.5025  5.9218 
## 
## Coefficients:
##                                                      Estimate Std. Error
## (Intercept)                                           3.34647    0.57853
## Country                                              -0.30754    0.29268
## Age1 55 to 90                                        -0.47851    0.18701
## SexFemale                                             0.27768    0.17929
## HH_Inc_pm £2,561 - 4,490                              0.63806    0.20714
## HH_Inc_pm4,911 - plus                                 0.78666    0.25231
## Personal_View_of_EU_Migration2. Probably Against      0.04081    0.37231
## Personal_View_of_EU_Migration3. Probably In Favour    1.06134    0.33534
## Personal_View_of_EU_Migration4. Definately In Favour  2.29804    0.33534
## Employment_Perception2Creates_jobs                    1.82536    0.41272
##                                                      t value Pr(>|t|)    
## (Intercept)                                            5.784 1.25e-08 ***
## Country                                               -1.051  0.29385    
## Age1 55 to 90                                         -2.559  0.01078 *  
## SexFemale                                              1.549  0.12203    
## HH_Inc_pm £2,561 - 4,490                               3.080  0.00218 ** 
## HH_Inc_pm4,911 - plus                                  3.118  0.00192 ** 
## Personal_View_of_EU_Migration2. Probably Against       0.110  0.91275    
## Personal_View_of_EU_Migration3. Probably In Favour     3.165  0.00164 ** 
## Personal_View_of_EU_Migration4. Definately In Favour   6.853 2.02e-11 ***
## Employment_Perception2Creates_jobs                     4.423 1.18e-05 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 2.047 on 528 degrees of freedom
## Multiple R-squared:  0.2776, Adjusted R-squared:  0.2653 
## F-statistic: 22.54 on 9 and 528 DF,  p-value: < 2.2e-16

Consolidation has once again demonstrated the significance of age, higher earning favouring migration and a fear that migration could also threaten jobs.

RESIDUALS V LEVERAGE - Assumption 2 Heteroscedastity/ Homoscedasticity

Table 22. Log Log

Table 22.a Table 23. Semi-log This Semi-log model is preferable to the log log model. The observations in this Semi-log plot are mostly clustered around 0.01 to 0.02 and are normally distributed as they are evenly spread at the high and low points around the 0 line. As anticipated, there are a few outliers and they do not compromise the even distribution of the residuals around the mean throughout, and therefore the variance is constant.

Table 23.a QQ PLOT OF RESIDUALS These variables indicate no clear linear pattern, although in the semi-log models, Table 23 and 23.a, the red line is almost flat: it is not in the residuals, which is preferable. As anticipated, the residual variables do not pull away from the line very much as the original histogram in table 3 was not skewed.

However, when we account for the overall model. The outliers are clearly visible in both tables but the residuals follow an almost normal distribution.

BSA2019subset1 %>%
  lm(MiCultur ~ Country + Age1 + Sex + HH_Inc_pm + Personal_View_of_EU_Migration + Employment_Perception2,
     data = .) %>%
  summary()
## 
## Call:
## lm(formula = MiCultur ~ Country + Age1 + Sex + HH_Inc_pm + Personal_View_of_EU_Migration + 
##     Employment_Perception2, data = .)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -5.6415 -1.4954  0.1586  1.5025  5.9218 
## 
## Coefficients:
##                                                      Estimate Std. Error
## (Intercept)                                           3.34647    0.57853
## Country                                              -0.30754    0.29268
## Age1 55 to 90                                        -0.47851    0.18701
## SexFemale                                             0.27768    0.17929
## HH_Inc_pm £2,561 - 4,490                              0.63806    0.20714
## HH_Inc_pm4,911 - plus                                 0.78666    0.25231
## Personal_View_of_EU_Migration2. Probably Against      0.04081    0.37231
## Personal_View_of_EU_Migration3. Probably In Favour    1.06134    0.33534
## Personal_View_of_EU_Migration4. Definately In Favour  2.29804    0.33534
## Employment_Perception2Creates_jobs                    1.82536    0.41272
##                                                      t value Pr(>|t|)    
## (Intercept)                                            5.784 1.25e-08 ***
## Country                                               -1.051  0.29385    
## Age1 55 to 90                                         -2.559  0.01078 *  
## SexFemale                                              1.549  0.12203    
## HH_Inc_pm £2,561 - 4,490                               3.080  0.00218 ** 
## HH_Inc_pm4,911 - plus                                  3.118  0.00192 ** 
## Personal_View_of_EU_Migration2. Probably Against       0.110  0.91275    
## Personal_View_of_EU_Migration3. Probably In Favour     3.165  0.00164 ** 
## Personal_View_of_EU_Migration4. Definately In Favour   6.853 2.02e-11 ***
## Employment_Perception2Creates_jobs                     4.423 1.18e-05 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 2.047 on 528 degrees of freedom
## Multiple R-squared:  0.2776, Adjusted R-squared:  0.2653 
## F-statistic: 22.54 on 9 and 528 DF,  p-value: < 2.2e-16

The semi-log show the coefficient intercept 3.34647 and indicates changes in the IV to the percentage change of the DV. Therefore, changes to the IV will impact a percentage change in our DV. For example, the coefficient for Country -0.30754 *100. The percentage change in the DV is approximately 30.74%.

Table 24.

## 
##  studentized Breusch-Pagan test
## 
## data:  .
## BP = 13.235, df = 9, p-value = 0.1523

Table 25.

## 
##  RESET test
## 
## data:  .
## RESET = 0.89893, df1 = 2, df2 = 526, p-value = 0.4076

Both the Breusch-Pagen and RESET test are greater than 0.05 and therefore we fail to reject the null hypothesis, and therefore homoscedasticity exists, which satisfies assumption 2.

Table 26.

## 
## Call:
## lm(formula = MiCultur ~ Country + Age1 + Sex + HH_Inc_pm + Personal_View_of_EU_Migration + 
##     Employment_Perception2, data = BSA2019subset1)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -5.6415 -1.4954  0.1586  1.5025  5.9218 
## 
## Coefficients:
##                                                      Estimate Std. Error
## (Intercept)                                           3.34647    0.57853
## Country                                              -0.30754    0.29268
## Age1 55 to 90                                        -0.47851    0.18701
## SexFemale                                             0.27768    0.17929
## HH_Inc_pm £2,561 - 4,490                              0.63806    0.20714
## HH_Inc_pm4,911 - plus                                 0.78666    0.25231
## Personal_View_of_EU_Migration2. Probably Against      0.04081    0.37231
## Personal_View_of_EU_Migration3. Probably In Favour    1.06134    0.33534
## Personal_View_of_EU_Migration4. Definately In Favour  2.29804    0.33534
## Employment_Perception2Creates_jobs                    1.82536    0.41272
##                                                      t value Pr(>|t|)    
## (Intercept)                                            5.784 1.25e-08 ***
## Country                                               -1.051  0.29385    
## Age1 55 to 90                                         -2.559  0.01078 *  
## SexFemale                                              1.549  0.12203    
## HH_Inc_pm £2,561 - 4,490                               3.080  0.00218 ** 
## HH_Inc_pm4,911 - plus                                  3.118  0.00192 ** 
## Personal_View_of_EU_Migration2. Probably Against       0.110  0.91275    
## Personal_View_of_EU_Migration3. Probably In Favour     3.165  0.00164 ** 
## Personal_View_of_EU_Migration4. Definately In Favour   6.853 2.02e-11 ***
## Employment_Perception2Creates_jobs                     4.423 1.18e-05 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 2.047 on 528 degrees of freedom
## Multiple R-squared:  0.2776, Adjusted R-squared:  0.2653 
## F-statistic: 22.54 on 9 and 528 DF,  p-value: < 2.2e-16

The model has met the assumptions and is therefore robust. The final model indicates that we reject the null hypothesis. The R-squared explains 27% of the variation between England and Scotland. The t-test indicates that 3.3464/2.047= 1.6347826 living in Scotland are likely to believe that migration will undermine cultural life compared to England. The mean-variance for England is 3.4823 compared to 3.1398 for Scotland. Thus, both countries believe that migration undermines cultural life rather than enriches it.

This model supports the qualitative evidence posed by the Scottish government concerning the older age group who are slightly more inclined than their English counterparts to opposed migration due to a fear to cultural life. Gender has continuously just failed to reach a level statistically significant to interpret meaningfully. The population on a higher wage fall in the neutral category in relation to their perception of migration compared to the people on a low income, especially when there is an EU trade benefit. Finally, it is highly statistically significant that the whole population fears that jobs are threatened due to migration. Even though the sample size is small, the degree of freedom tested n 501 -1 interpreted whether this is a representative sample. Hence, the P-value of the F test confirms that the variation among the two groups is significant. Thus, if the samples were drawn again, the outcome would not vary, which is generalizable to the population. Inter alia, this quantitative highlights the Scottish government’s qualitative reports are not completely generalisable as it has violated some assumptions in relation to omitted variable bias in relation to high earners and multicollinarity between regional and country variations.