Te Taunaki Public Service Census: Inclusion deep dive

  1. Section 01
    Introduction
  2. Section 02
    Inclusion Deep Dive Overview and Approach
  3. Section 03
    Findings on inclusion
  4. Section 04
    Feelings of inclusion for different groups and communities
  5. Section 05
    How work and environmental factors influence inclusion
  6. Section 06
    Workplace experiences and inclusion
  7. Section 07
    Conclusion
  8. Section 08
    APPENDIX A - Methodology

Scale development

Questions in Te Taunaki were drawn from a variety of sources or created based on stakeholder needs. This meant that we were not using a previously validated scale to measure the concept of inclusion.

In general, single item scales are less reliable for analysis[10] and therefore we sought to develop a multi-item scale with a mean score that could capture the concept of inclusion.

An exploratory factor analysis was conducted on the questions in the inclusion section of Te Taunaki. We identified four questions that loaded on a single factor and had good reliability (alpha=0.84). Adding other questions from the inclusion section reduced the reliability of the scale and therefore were eliminated from the final scale. These are reported on their own.

Items used in the inclusion scale:

  • I feel accepted as a valued member of the team
  • I feel comfortable being myself at work/with my colleagues
  • The agency I work for supports and actively promotes an inclusive workplace,
  • The people in my workgroup behave in an accepting manner towards people from diverse backgrounds

Items excluded from the scale, reported individually:

  • I am valued for the range of cultural expertise I bring to the job
  • I have access to employee led networks relevant to me
  • I feel comfortable working with people from backgrounds other than my own

Analysis

An initial Ordinary Least Squares regression was performed using R on the Te Taunaki data (obs=39977) using most (23) of the demographic and employment variables collected in the Census.  Note that there are likely to be multiple factors that explain people’s feeling of inclusion that were not captured in the Census.

The target variable was the average of the Likert scores (1-5 scale) over the four questions described above for each respondent. This initial analysis resulted in the following variables being removed as they were not adding explanatory power to the model: 

  • Migrant Status 
  • Religion 
  • Caring Responsibilities 
  • Tenure  
  • Occupation Group 
  • Job Contract (PSA/individual) 
  • Employment Type (permanent/temporary). 

The final variable set consisted of: 

  • Disability 
  • Mental Health 
  • Age Group 
  • Gender 
  • Sexual Identity 
  • Intersex/Transgender 
  • Ethnicity 
  • Highest Qualification
  • Flexible Working 
  • Customer facing 
  • Management Level 
  • Salary (scale) 
  • Region (Wellington/other) 

The results were as follows:  

Variable 

Co-efficient 

Disability (base – no disability)

Disability                               

-0.167*** 

Mental Health (base – no mental health condition) 

Mental Health condition  

-0.174*** 

Age (as an increasing scale)

Age group (10-year bands)                               

-0.015*** 

Gender (base – female)

Male                                      

0.029*** 

Other                                     

-0.101 

Unknown                                   

-0.135** 

Sexual Identity (base – straight/heterosexual)

Bisexual                          

-0.068*** 

Gay                               

-0.049 

Lesbian                           

-0.031 

Other                             

-0.159*** 

unknown                           

-0.148*** 

Intersex (base – no intersex variation)

Intersex                                       

-0.076 

Transgender (base – not transgender)

Transgender                                    

-0.187** 

Ethnicity (base – ‘no’)

European=Yes                            

0.032** 

Māori=Yes                               

-0.079*** 

Pacific=Yes                             

-0.061*** 

Asian=Yes                               

-0.001 

MELAA=Yes                               

-0.064* 

Other=Yes                               

-0.087** 

Highest qualification (as an increasing scale)

Highest qualification                                

-0.004 

Flexible working (base - Don't have/ Do not want)

Don’t have / Want             

-0.216*** 

Have / Don’t want more         

0.152*** 

Have / Want more       

-0.079*** 

Not Specified       

-0.012

Customer Facing (base – ‘no’)

In a customer facing role

-0.037*** 

Management level (base - no managerial responsibilities)

Manager                                     

 

0.082*** 

Prefer not to answer                        

0.210*** 

Dont know                                   

0.001 

Salary (as an increasing scale)

Salary Band                                       

0.005*** 

Region (base – rest of country)

Wellington                         

0.043*** 

 

Constant                                         

4.203*** 

  

  

N                                                

34504 

R2                                               

0.058 

Adjusted R2                                      

0.058 

chi2                                             

2,073.434*** (df = 30)   

***Significant at the 1% level. 

  

 **Significant at the 5% level. 

  

 *Significant at the 10% level. 

  

 An additional ordered Logistic Regression Model, which looked across the five-point Likert scale, was run on each of the four questions individually, with categorical variables used instead of scales for Age, Highest Qualification and Salary. This was done to look at the individual variables separately and as a robustness check on our main OLS model.  

The results were similar to the main model, providing reassurance the main OLS model using the scale was providing robust results. There were some interesting differences between the four individual Likert questions and these have been discussed in the main text.

 The results were as follows: 

Variable 

Co-efficient by inclusion question

accepted as a valued member of the team 

comfortable being myself at work

people in my workgroup behave in an accepting manner 

agency I work for supports an inclusive workplace 

Disability (base – no disability)

Disability                               

0.676*** 

0.721*** 

0.767*** 

0.786*** 

Mental Health (base – no mental health condition) 

Mental Health condition  

0.690*** 

0.600*** 

0.780*** 

0.783*** 

Age group (base – 25-34 years)

Under 25 years       

 

1.219*** 

 

1.167*** 

 

1.313*** 

 

1.282*** 

35 to 44 years       

0.829*** 

0.884*** 

0.947* 

0.917*** 

45 to 54 years       

0.753*** 

0.862*** 

0.898*** 

0.916*** 

55 to 64 years       

0.760*** 

0.941* 

0.975 

0.869*** 

65 years or over     

0.856*** 

1.172*** 

1.087 

0.942 

Unknown Age          

0.589*** 

0.684** 

0.752* 

0.779* 

Gender (base – female)

Male                                      

 

1.102*** 

 

1.000 

 

1.193*** 

 

1.204*** 

Other                                     

0.951 

0.766 

0.780 

0.844 

Unknown                                   

0.827 

0.719** 

0.896 

0.824 

Sexual Identity (base – straight/heterosexual)

Bisexual                          

 

1.011 

 

0.790*** 

 

0.759*** 

 

0.874** 

Gay                               

0.919 

0.890 

0.939 

0.894 

Lesbian                           

1.177* 

1.104 

0.887 

0.899 

Other                             

0.974 

0.723*** 

0.592*** 

0.660*** 

Unknown                           

0.761*** 

0.701*** 

0.731*** 

0.753*** 

Intersex (base – no intersex variation)

Intersex

0.784 

1.117 

0.664 

1.100 

Transgender (base – not transgender)

Transgender

0.905 

0.573*** 

0.753* 

0.658** 

Ethnicity (base – ‘no’)

European=Yes                            

 

0.986 

 

1.015 

 

1.103*** 

 

1.139*** 

Māori=Yes                               

0.935** 

0.912*** 

0.771*** 

0.788*** 

Pacific=Yes                             

0.945 

0.950 

0.811*** 

0.841*** 

Asian=Yes                               

0.962 

0.991 

0.897** 

1.018 

MELAA=Yes                               

0.979 

0.805** 

0.925 

0.998 

Other=Yes                               

0.760*** 

0.744*** 

0.919 

0.871 

Highest qualification (base - Bachelor Honours Degree or Postgraduate Certificate/Diploma)

No Qualifications               

 

 

0.789*** 

 

 

1.006 

 

 

0.959 

 

 

0.825** 

High School/Secondary School  

0.977 

1.130*** 

1.005 

1.084** 

Level 1 to 4 Certificate        

0.946 

1.182*** 

1.008 

1.036 

Level 5 or 6 Diploma            

0.954 

1.075* 

1.000 

1.011 

Bachelor’s Degree or Level 7  

1.025 

1.060* 

1.021 

1.035 

Master’s Degree                 

0.975 

0.943* 

0.963 

0.897*** 

PhD/Doctoral Degree             

1.146* 

0.980 

1.139* 

0.799*** 

Other  

0.913 

0.901 

0.813 

0.972 

Don't know                      

1.125 

0.870 

1.066 

1.020 

Prefer not to answer            

0.725*** 

0.840* 

0.876 

0.738*** 

Flexible working (base – no, does not have)

Yes, has flexible work                      

 

1.444*** 

 

1.182*** 

 

1.349*** 

 

1.606*** 

Not Elsewhere Included   

0.889 

0.759** 

0.831 

0.788** 

Customer Facing (base – ‘no’)

In a customer facing role

 

0.906*** 

 

1.050** 

 

0.915*** 

 

0.852*** 

Management level (base - no managerial responsibilities)

Manager                                     

 

 

1.127*** 

 

 

1.230*** 

 

 

1.004 

 

 

1.292*** 

Prefer not to answer                        

1.149 

1.073 

0.980 

1.764*** 

Dont know                                   

1.008 

1.057 

0.931 

1.095 

Salary band (base – $100,000 - $109,999)

Less than $40,000

 

0.679*** 

 

0.738*** 

 

0.851 

 

0.86 

$40,000 - $44,999

1.017 

1.039 

1.047 

1.035 

$45,000 - $49,999

1.178** 

1.143* 

1.544*** 

1.702*** 

$50,000 - $54,999

0.996 

0.96 

1.161*** 

1.161*** 

$55,000 - $59,999

0.941 

0.929 

1.073 

1.064 

$60,000 - $64,999

0.851*** 

0.872*** 

0.919* 

0.889** 

$65,000 - $69,999

0.827*** 

0.935 

0.833*** 

0.824*** 

$70,000 - $74,999

0.859*** 

0.897** 

0.923 

0.93 

$75,000 - $79,999

0.876** 

1.032 

0.962 

1.006 

$80,000 - $84,999

0.925 

0.982 

0.984 

1.084 

$85,000 - $89,999

0.963 

0.983 

0.939 

0.987 

$90,000 - $94,999

1.012 

0.982 

1.062 

1.105* 

$95,000 - $99,999

0.975 

0.981 

0.973 

1.096 

$110,000 - $119,999

1.092 

1.146** 

1.130** 

1.214*** 

$120,000 - $129,999

1.279*** 

1.226*** 

1.271*** 

1.267*** 

$130,000 - $139,999

1.175** 

1.210*** 

1.242*** 

1.372*** 

$140,000 - $149,999

1.310*** 

1.339*** 

1.288*** 

1.376*** 

$150,000 - $199,999

1.308*** 

1.354*** 

1.196*** 

1.387*** 

$200,000 or more

2.040*** 

1.840*** 

1.691*** 

1.825*** 

Don't know

0.867 

0.885 

1.068 

0.921 

Prefer not to answer

0.838*** 

0.872** 

0.949 

0.963 

Region (base – rest of country)

Wellington                         

 

1.023 

 

0.903*** 

 

1.159*** 

 

1.109*** 

N 

36370 

36386 

36159 

36176 

R2                                               

0.045 

0.037 

0.043 

0.062 

C (Area under the ROC curve)

0.593

0.582

0.593

0.610

***Significant at the 1% level. 

  

 

 

 

 **Significant at the 5% level. 

  

 

 

 

 *Significant at the 10% level. 

  

 

 

 

 

[10] Guidelines for choosing between multi-item and single-item scales for construct measurement: a predictive validity perspective | SpringerLink

  1. Section 07
    Conclusion