You are here: Home / Statistics / Growing Up in Ireland Cohort ‘98 at Age 25: Childhood Influences on Education and Finances / Appendix

Growing Up in Ireland Cohort ‘98 at Age 25: Childhood Influences on Education and Finances

Appendix

CSO statistical release, 23 February 2026, 11am

A CSO Frontier Series Output

This release is categorised as a CSO Frontier Series Output. Particular care must be taken when interpreting the statistics in this release as it may use new methods which are under development and/or data sources which may be incomplete, for example, new administrative data sources.

Multiple logistic regression models were examined taking a stepwise approach. The final models used were selected using Akaike Information Criterion which compares the relative quality of different statistical models for a given set of data to find the model that best explains the data without being overly complex, balancing goodness of fit with model simplicity. The variables for inclusion as confounders in the model were selected to reduce the risk of overestimation of the effect of the predictor variables. See Background Notes for more information on the development of these models.

Table A1, A2 and A3 show the final models used in the logistic regression analyses for this release and the associated marginal effect for each variable in the model.

Table A1 Relative probability of respondents' obtaining a level 7 degree or higher by age 25 – final model¹

Table A1 Relative probability of respondents' obtaining a level 7 degree or higher by age 25 – final model¹
Independent Variable at Wave 1	Reference Group	Average Marginal Effect ²	Standard Error	Pr(>\|t\|)³
Predictor Variables
TV in bedroom at age 9
No TV in bedroom	TV in bedroom	0.10	0.02	<0.001***

Parental expectation of child’s education
Will obtain a degree or higher	Will not obtain a degree	0.19	0.02	<0.001***

Parent uses public library for child
Yes	No	0.07	0.02	0.003**

Control Variables
Housing tenure at Wave 1
Owner occupied	Other tenure status	0.12	0.04	0.001**

Primary caregiver highest level of education at Wave 1
Degree or higher	Less than degree	0.05	0.03	0.053

Social class (SC) of household at Wave 1
SC3-4	SC1-2	-0.04	0.03	0.103
SC5-6	SC1-2	-0.07	0.04	0.063

Households equivalised income (quintile) at Wave 1
Quintile 2	Quintile 1 (Lowest)	0.04	0.04	0.322
Quintile 3	Quintile 1 (Lowest)	0.06	0.04	0.145
Quintile 4	Quintile 1 (Lowest)	0.08	0.04	0.073
Quintile 5 (Highest)	Quintile 1 (Lowest)	0.13	0.04	0.002**

Ability of parent’s household to make ends meet when parent was 16
At least some difficulty making ends meet	Did not have difficulty making ends meet	-0.04	0.02	0.099

Age of primary caregiver at Wave 1
Per 1 year increase in age	NA	0.01	0.002	<0.001***

¹ The average marginal effects from the logistic regressions on the respondent obtaining at least an NFQ level 7 degree by the age of 25 (dependent variable). The number of observations was 3,028. Nagelkerke pseudo R² = 0.24.
²The average marginal effects on the probability of obtaining at least an NFQ level 7 degree in a logistic regression can be interpreted as the average increase in the probability of obtaining a degree (all else being equal) when the relevant independent variable is changed from the reference group e.g. respondents who did not have a TV in their room at age 9 increases the probability of the respondent having a degree at age 25 by 10% (0.10), on average in the sample, relative to respondents who had a TV in their room at age 9.
³ Probability value associated with the t-statistic. P-values of < 0.05 signify the presence of a statistically significant relationship. * p < 0.05; p < 0.01; * p < 0.001

Open in Excel: Table A1 Relative probability of respondents' obtaining a level 7 degree or higher by age 25 – final model (XLS 20KB)

Table A2 Relative probability of respondents’ having at least some difficulty making ends meet at age 25 – final model¹

Table A2 Relative probability of respondents’ having at least some difficulty making ends meet at age 25 – final model¹
Independent Variable at Wave 1	Reference Group	Average Marginal Effect ²	Standard Error	Pr(>\|t\|)³
Predictor Variables
Households equivalised income (quintile) at Wave 1
Quintile 2	Quintile 1 (Lowest)	-0.06	0.04	0.209
Quintile 3	Quintile 1 (Lowest)	-0.04	0.05	0.412
Quintile 4	Quintile 1 (Lowest)	-0.10	0.04	0.021*
Quintile 5 (Highest)	Quintile 1 (Lowest)	-0.10	0.05	0.026*

Control Variables
Sex of respondent
Female	Male	0.15	0.02	<0.001***

Housing tenure at Wave 1
Owner occupied	Other tenure status	-0.04	0.04	0.259

Social Class (SC) of Household at Wave 1
SC3-4	SC1-2	0.05	0.03	0.046*
SC5-6	SC1-2	0.11	0.04	0.007**

Ability of respondent’s household to make ends meet at Wave 1
At least some difficulty making ends meet	Did not have difficulty making ends meet	0.06	0.03	0.042*

Ability of parent’s household to make ends meet when parent was 16
At least some difficulty making ends meet	Did not have difficulty making ends meet	0.02	0.02	0.33

Age of primary caregiver at Wave 1
Per 1 year increase in age	NA	-0.01	0.002	0.024*

¹ The average marginal effects from the logistic regressions on the respondent having at least some difficulty making ends meet at the age of 25 (dependent variable). The number of observations was 3,038. Nagelkerke pseudo R² = 0.10.
²The average marginal effects on the probability of having at least some difficulty making ends meet in a logistic regression can be interpreted as the average change in the probability of having at least some difficulty making ends meet (all else being equal) when the relevant independent variable is changed from the reference group e.g. the respondent who, at the age of 9, lived in a household in the 5th equivalised income quintile (highest) decreases the probability of the respondent having at least some difficulty making ends meet at age 25 by 10% (-0.10), on average in the sample, relative to respondents who lived in a household in the 1st equivalised income quintile (lowest).
³ Probability value associated with the t-statistic. P-values of < 0.05 signify the presence of a statistically significant relationship. * p < 0.05; p < 0.01; * p < 0.001

Open in Excel: Table A2 Relative probability of respondents’ having at least some difficulty making ends meet at age 25 – final model (XLS 19KB)

Table A3 Relative probability of respondents' having ‘a lot’ or ‘a little’ difficulty making loan repayments at age 25 – final model¹

Table A3 Relative probability of respondents' having ‘a lot’ or ‘a little’ difficulty making loan repayments at age 25 – final model¹
Independent Variable at Wave 1	Reference Group	Average Marginal Effect ²	Standard Error	Pr(>\|t\|)³
Predictor Variables
Households equivalised income (quintile) at Wave 1
Quintile 2	Quintile 1 (Lowest)	-0.04	0.03	0.188
Quintile 3	Quintile 1 (Lowest)	-0.02	0.04	0.612
Quintile 4	Quintile 1 (Lowest)	-0.06	0.03	0.079
Quintile 5 (Highest)	Quintile 1 (Lowest)	-0.08	0.03	0.024*

Control Variables
Sex of respondent
Female	Male	0.04	0.02	0.018*

Ability of parent’s household to make ends meet when parent was 16
At least some difficulty making ends meet	Did not have difficulty making ends meet	0.02	0.02	0.238

Housing tenure at Wave 1
Owner occupied	Other tenure status	-0.02	0.03	0.422

Social Class (SC) of Household at Wave 1
SC3-4	SC1-2	0.01	0.02	0.479
SC5-6	SC1-2	0.01	0.03	0.805

Ability of respondent’s household to make ends meet at Wave 1
At least some difficulty making ends meet	Did not have difficulty making ends meet	0.01	0.02	0.538

Age of primary caregiver at Wave 1
Per 1 year increase in age	NA	-0.003	0.002	0.035*

Household experiencing enforced deprivation at Wave 1
Yes	No	0.04	0.05	0.354

¹ The average marginal effects from the logistic regressions on the respondent having ‘a lot’ or ‘a little’ difficulty making loan repayments at the age of 25 (dependent variable). The number of observations was 3,038. Nagelkerke pseudo R² = 0.05.
²The average marginal effects on the probability of having ‘a lot’ or ‘a little’ difficulty making loan repayments in a logistic regression can be interpreted as the average increase in the probability of having a lot or a little difficulty making loan repayments (all else being equal) when the relevant independent variable is changed from the reference group e.g. the respondent who, at the age of 9, lived in a household in the 5th equivalised income quintile (highest) decreases the probability of the respondent having a lot or a little difficulty making loan repayments at age 25 by 8% (-0.08), on average in the sample, relative to respondents who lived in a household in the 1^st equivalised income quintile (lowest).
³ Probability value associated with the t-statistic. P-values of < 0.05 signify the presence of a statistically significant relationship. * p < 0.05; p < 0.01; * p < 0.001

Open in Excel: Table A3 Relative probability of respondents' having ‘a lot’ or ‘a little’ difficulty making loan repayments at age 25 – final model (XLS 20KB)

Next chapter: Data

Why you can Trust the CSO

Learn about our data and confidentiality safeguards, and the steps we take to produce statistics that can be trusted by all.

go to trust & transparency

Growing Up in Ireland Cohort ‘98 at Age 25: Childhood Influences on Education and Finances

Appendix

Appendix

A CSO Frontier Series Output

Contents

Related

Why you can Trust the CSO

Contact:

Links:

Careers

Follow: