Results from TALIS Starting Strong 2024

Annex C. Technical notes on analyses in this report

Use of staff and setting weights

The statistics presented in this report were derived from data obtained through samples of settings, setting leaders and staff (see Annex B). For these statistics to describe a country or subnational entity they need to reflect the whole population from which they were drawn and not merely the sample used to collect them. Survey weights must therefore be used to obtain design-unbiased estimates of population or model parameters.

Final weights allow the production of country- or subnational entity-level estimates from the observed sample data. The estimation weight indicates how many population units are represented by a sampled unit. The final weight is the combination of many factors reflecting the probabilities of selection at the various stages of sampling and the response obtained at each stage. Other factors may also come into play as dictated by special conditions to maintain the unbiasedness of the estimates (e.g. adjustment for staff working in more than one centre). A detailed description of the sampling and weighting procedures can be found in the TALIS Starting Strong 2024 Technical Report (OECD, forthcoming[1]).

Data from New Zealand were weighted differently. Due to sparse response that led to unanticipated exclusions for some segments of the target population these samples (both ISCED 02 and settings for children under age 3) cannot be adequately adjusted through weighting to represent early childhood education and care (ECEC) settings and staff in New Zealand overall. Nonetheless, staff and setting weights were applied to adjust for different probabilities of selection into the sample, but not to make the samples representative of the overall target population for New Zealand. Thus, estimates presented in this volume for New Zealand only represent respondents included in the sample and not the population targeted by the survey.

Statistics in this report that represent settings and statistics based solely on leader responses were estimated using leader weights (LEADERWGT). Leader weights are used to make the sample representative of the population of ECEC settings in each country and subnational entity (with the exception of New Zealand); results are not representative of the population of leaders. Results that represent staff, which are based only on responses of staff or on responses of staff and leaders (i.e. responses from setting leaders were merged with staff responses), were estimated using staff weights (STAFFWGT). Staff weights are used to make the sample representative of the population of ECEC staff in each country and subnational entity (with the exception of New Zealand).

Coverage of the sample

Not all respondents answer every question, either because they omit a response or give an invalid response. Coverage is defined as the weighted percentage of respondents who provide a valid answer to a given question in the survey, regardless of their answer. Coverage daggers are shown in every table to indicate the level of coverage: no symbol means at least 75% of the expected population was covered; one dagger (†) means at least 50% but less than 75%; and a double-dagger (‡) means less than 50% was covered.

Coverage is calculated based on the section of the sample that could answer the relevant question in the survey. Some respondents did not have the possibility of answering some questions due to filters in the questionnaire (i.e. the question being not logically applicable) or to use of different versions of the questionnaire targeted to different types of setting leaders and staff (see the TALIS Starting Strong 2024 Technical Report for more detail (OECD, forthcoming[1]). These respondents were excluded from coverage and estimate calculations. In addition, respondents in home-based settings were excluded from coverage and estimate calculations wherever the analysis was not appropriate for this type of setting. These exclusions are indicated in notes below the relevant tables in Annex D.

Standard errors and significance tests

Standard errors

The statistics in this report represent estimates based on samples of staff and settings, rather than values that could be calculated if every staff member and setting leader in every country or subnational entity had answered every question. Consequently, it is important to measure the degree of uncertainty of the estimates. In the Starting Strong Teaching and Learning International Survey (TALIS Starting Strong) 2024, each estimate has an associated degree of uncertainty that is expressed through a standard error. The reported standard errors were computed with a Balanced Repeated Replication (BRR) method, which provides a simple way of producing a consistent estimator that takes into account the sampling strategies used in the survey (a detailed description can be found in the TALIS Starting Strong 2024 Technical Report (OECD, forthcoming[1]). The use of confidence intervals provides a way to make inferences about the population statistics in a manner that reflects the uncertainty associated with the sample estimates. From an observed sample statistic and assuming a normal distribution, its corresponding confidence interval at the 95% level can be constructed as the estimated parameter, plus or minus 1.96 times the associated standard error. In this way, it can be inferred that the corresponding parameter for the population would lie within the confidence interval in 95 out of 100 replications of the measurement on different samples drawn from the same population.

Differences between sub-groups

Differences between sub-groups either for staff (e.g. teachers and assistants) or setting characteristics (e.g. settings with a high concentration of children from socio-economically disadvantaged homes and settings with a low concentration of children from socio-economically disadvantaged homes) were tested for statistical significance. All differences marked in bold in the data tables of this report are statistically significantly different from 0 at the 95% level. This means that differences were labelled as statistically significant when a difference of that size or greater, in either direction, would be observed less than 5% of the time in the sample estimates, if there were actually no difference in the corresponding population values. In the case of differences between sub-groups, standard errors were calculated by taking into account that the two sub-samples are not independent, with the help of the BRR method.

Statistics based on regressions

Country- and subnational entity-specific regression analyses were performed to examine the associations between different variables. TALIS Starting Strong 2024 used a multivariable regression model for linear and logistic regression analyses. The primary goal was to accurately estimate the independent contribution of each predictor while adjusting for potential confounders. This model allows expressing the outcome as a function of all relevant variables simultaneously – where each coefficient (or odds ratio, OR) represents the expected change in the outcome (or in the likelihood of the outcome rather than the reference category, for ORs) when that predictor increases by one unit (or takes the indicated value rather than the reference category, for categorical variables), holding all other variables constant. Regression analyses presented in this report handle missing data by listwise deletion. The percentage of the sample on which the regression is estimated is indicated by the coverage signs presented in the tables (see above for details).

Two types of models were used in this volume. Multiple linear regression was used in those cases where the dependent (or outcome) variable was considered continuous, such as scales. Binary logistic regression was employed when the dependent (or outcome) variable was a binary categorical variable, such as a high versus a low share of children from socio-economically disadvantaged homes.

Linear regression models

Linear regression models were used where the dependent (or outcome) variable was continuous. The models were specified as follows:

$Y=\mathrm{ }{\beta }_0+{\beta }_1{X}_1+{\beta }_2{X}_2+\dots +{\beta }_p{X}_p+ϵ$

Where Y is the predicted variable, $X_1,X_2,\dots ,\mathrm{ }{X}_{\mathrm{p}}$ is the set of regressors, ${\beta }_1,{\beta }_2,\dots ,\mathrm{ }{\beta }_{\mathrm{p}}$ is the set of the coefficients that measure the strength of association, and $ϵ$ is an error term. Coefficients presented in the tables are unstandardised, were estimated using the Ordinary Least Squares method and represent the expected change in the Y variable with a change in the predictor, holding all other predictors constant. The magnitude of the unstandardised coefficient linking a predictor and an outcome in a linear regression is referred to as the “strength of association.” If the predictor is a categorical or binary variable, the coefficient represents the change in Y associated with being in the specified category rather than the reference category. If the predictor is a continuous variable, the coefficient represents the change in Y associated with a one-unit change in the predictor. Associated standard errors were calculated using the BRR method. Coefficients were tested for difference from 0 at the 95% confidence level using a Student’s t-test. Coefficients are indicated to be statistically significantly different from 0 when an estimate of that size or greater, in either direction, would be observed less than 5% of the time in the sample estimates, if the corresponding population value were actually 0.

Logistic regression models (odds ratios)

Logistic regression (logit) models were used where the dependent (or outcome) variable was binary. The models were specified as follows:

$Logit\left(\pi \left(y\right)\right)=\ln \left(\frac{\pi \left(y\right)}{1-\pi \left(y\right)}\right)=\mathrm{ }{\beta }_0+{\beta }_1{X}_1+{\beta }_2{X}_2+\dots +{\beta }_p{X}_p+ϵ$

Where Y is the predicted variable, with probability $\pi \left(y\right)$ for outcome y instead of the reference category, $X_1,X_2,\dots ,\mathrm{ }{X}_{\mathrm{p}}$ is the set of regressors, ${\beta }_1,{\beta }_2,\dots ,\mathrm{ }{\beta }_{\mathrm{p}}$ is the set of the coefficients that measure the strength of association, and $ϵ$ is an error term. Associated standard errors were calculated using the BRR method.

In the tables, logit coefficients are not reported directly. Instead, ORs are presented. The OR is a measure of the relative likelihood of a particular outcome across two groups. In this case, ORs are equivalent to the ratio between the probability of observing outcome y instead of the reference category when a predictor takes on a particular value instead of the reference category, and the probability of observing the same outcome in Y when the predictor does not take that value. Given that $p_{XY}$ is the probability of observing a given outcome for Y given a particular value for X, the OR can be defined as follows:

$OR= \frac{p_{11}/p_{12}}{p_{21}/p_{22}}$

Where $p_{11}/p_{12}$ represents the “odds” of observing the indicated outcome (1) rather than the reference category (2) in Y, when X takes on its indicated outcome (1); and $p_{21}/p_{22}$ represents the “odds” of observing the outcome (1) rather than the reference category (2) in Y, when X takes on the value of its reference category (2).

If the predictor is a categorical or binary, the OR represents the change in likelihood of Y taking on value y rather than the reference category associated when the predictor takes the specified value rather than its reference category. If the predictor is continuous, the coefficient represents the change in the likelihood of Y taking on value y associated with a one-unit change in the predictor.

In this report, the OR is obtained by applying an exponential transformation to the estimated logit coefficient. Throughout the report, ORs were calculated as the exponential of the logit coefficient, so that for any regressor j and its associated coefficient ${\beta }_{\mathrm{j}}$, its OR is:

${OR}_j=e^{{\beta }_j}$

Since this transformation is non-linear, statistical inference is performed on the original coefficient rather than on the OR itself. Hypothesis tests (Student’s t-tests) use the coefficient estimates and their standard errors. P-values are computed directly from the original coefficients and standard errors, using the inverse of the normal cumulative distribution function. This ensures accurate statistical inference and avoids errors that may arise from testing transformed values like ORs.

The same procedure can be applied to all predictors individually, with no modifications needed for each variable. The resulting p-value reflects the probability of observing an estimate of that size or greater, in either direction, if the true population logit coefficient were 0 (or the corresponding population OR were 1). In this report, ORs are indicated as statistically significant when an estimate of their associated logit coefficient of that size or greater, in either direction, would be observed less than 5% of the time in the sample estimates, if the corresponding population value for the logit coefficient were actually 0. This is equivalent to an associated p-value of less than 0.05.

Predictors used in all regressions

For each regression model, a consistent set of predictors (detailed below) and a more restricted set of predictors specific to the analytical goals of each model were included. The selection of predictors was grounded in substantive theory and prior research. This helped avoid overfitting and ensured that the included predictors have a meaningful interpretation. Multicollinearity was assessed by estimating bivariable variance inflation factors on the set of predictors used in all regressions. This analysis confirmed that the model’s independent effects are estimable with acceptable precision.

Predictors in all staff models

• Level of formal education: Based on the highest level of formal education completed by staff (variable: SS2G11). It was grouped into three categories: secondary education or below (ISCED level 3 or below), post-secondary non-tertiary or short-cycle tertiary education (ISCED levels 4 or 5), and bachelor’s degree or equivalent or higher (ISCED level 6 or higher; used as the reference category).

• Experience: Based on total experience as ECEC staff (variable: SS2G08B). It was grouped into three categories: 3 years or less, between 4 and 9 years, and 10 years or more (used as the reference category).

• Employment status: Based on the contract agreement staff have to work at their ECEC setting (variable: SS2G05). It was grouped into two categories: permanent employment and others (which encompasses fixed-term, self-employed or no employment contract; used as the reference category).

• Working hours: Based on working hours of ECEC staff (variable: SS2G06). It was grouped into two categories: full-time and part-time (whether staff indicated working 50-90% of full-time hours or less than 50% of full-time hours; used as the reference category).

• Staff role: Based on the category that staff indicated best represents their role in general when working at their ECEC setting (variable: SS2G03). It was grouped into two categories: teacher or ECEC setting leader and all other staff roles (which encompasses assistants, staff for individual children, staff for special tasks, interns and others; used as the reference category).

• Setting location: Based on the ECEC setting’s location (variable: SL2G13). It was grouped into two categories: a municipality with more than 15 000 people and a municipality with 15 000 people or less (used as the reference category).

• Setting type of management: Based on the organisation responsible for the day-to-day management of the ECEC setting (variable: SL2G19). It comprises two categories: publicly managed and privately managed (used as the reference category).

• Setting size (or target group size) tertiles: The position of the ECEC setting in the national distribution of ECEC setting size, based on the number of children enrolled in the ECEC setting (variable: SL2G25) at the setting level, or on the number of children in the target group (variable: SS2G35) at the target group level (relevant tables specify whether the setting size or target group size is used). Depending on the outcome, this variable refers either to the number of children enrolled in the ECEC setting or the number of children in the target group. The first [bottom] tertile was used as the reference category. Coefficients are therefore a comparison to the smallest ECEC settings (or target groups) in each country or subnational entity.

• Number of staff per child in the setting (or target group) tertiles: The position of the ECEC setting in the national distribution of number of staff per children, based on the sum of staff in the setting (based on variable: SL2G22) and the number of children enrolled in the setting (variable: SL2G25) at the setting level, or on the sum of staff in the target group (based on variable: SS2G34) and the number of children in the target group (variable: SS2G35) at the target group level (relevant tables specify whether the number of staff per child in the setting or target group is used). Depending on the outcome, this variable refers either to the number of staff per children in the ECEC setting or the number of staff per children in the target group. The first [bottom] tertile was used as the reference category. Coefficients are therefore a comparison to the ECEC settings (or target groups) where there are the fewest number of staff per child in each country or subnational entity.

Predictors in all setting/leader models

• Level of formal education: Based on the highest level of formal education completed by leaders (variable: SL2G03). It was grouped into three categories: short-cycle tertiary education or below (ISCED level 5 or below), bachelor’s degree or equivalent (ISCED level 6), and master’s degree or equivalent or higher (ISCED level 7 or higher; used as the reference category).

• Experience: Based on total experience as an ECEC setting leader (variable: SL2G05B). It was grouped into three categories: 3 years or less, between 4 and 9 years, and 0 years or more (used as the reference category).

• Setting location: Based on the ECEC setting’s location (variable: SL2G13). It was grouped into two categories: a municipality with more than 15 000 people and a municipality with 15 000 people or less (used as the reference category).

• Setting type of management: Based on the organisation responsible for the day-to-day management of the ECEC setting (variable: SL2G19). It comprises two categories: publicly managed and privately managed (used as the reference category).

• Setting size tertiles: The position of the ECEC setting in the national distribution of ECEC setting size, based on the number of children enrolled in the ECEC setting (variable: SL2G25) at the setting level. For ECEC setting/leader outcomes, this variable refers to the number of children enrolled in the ECEC setting. The first [bottom] tertile was used as the reference category. Coefficients are therefore a comparison to the smallest ECEC settings in each country or subnational entity.

• Number of staff per child in the setting tertiles: The position of the ECEC setting in the national distribution of number of staff per children, based on the sum of staff in the setting (based on variable: SL2G22) and the number of children enrolled in the setting (variable: SL2G25) at the setting level. For ECEC setting/leader outcomes, this variable refers to the number of staff per child in the ECEC setting. The first [bottom] tertile was used as the reference category. Coefficients are therefore a comparison to the ECEC settings where there are the fewest number of staff per child in each country or subnational entity.

Use of complex variables

Setting (or target group) size

Tertiles refer to 3 groups comprised approximately between the 1st and 33rd percentiles (first [bottom] tertile), the 34th and the 66th percentiles, and the 67th and 100th percentiles, in the weighted distribution of setting (or target group) size. Setting size is the number of children enrolled in the ECEC setting (variable: SL2G25). Target group size is the number of children in the target group (variable: SS2G35). Tertiles were calculated within each country and subnational entity, meaning that the values that go into each percentile vary with each country and subnational entity. These tertiles are of unequal size but of equal value, meaning that they group all observations with the same value in the same bin (typically the lower one). Therefore, the number of (weighted) observations in each bin is different, but all observations with the same value of setting (or target group) size are grouped in the same bin. For the setting size tertiles in Quebec (Canada), a further recoding was done, grouping all settings with less than 10 children enrolled, and dividing the rest of the settings (ranging from 29 to 191 enrolled children) into 2 similar-size groups. This procedure creates “tertiles” that are of very unequal (weighted) size, but have a clear cut in values, fit for analytical purposes.

Number of staff per child in the setting (or target group)

Tertiles refer to 3 groups comprised approximately between the 1st and 33rd percentiles, the 34th and the 66th percentiles, and the 67th and 100th percentiles, in the weighted distribution of staff per child in the setting (or in the target group). Staff per child in the setting is the number of staff working in the setting (sum of items A-D of variable SL2G22) divided by the number of children enrolled in the ECEC setting (variable: SL2G25). Staff per child in the target group is the number of staff in the target group (the sum of items A-D of variable SS2G34) divided by the number of children in the target group (variable: SS2G35). Tertiles were calculated within each country and subnational entity, meaning that the values that go into each percentile vary with each country and subnational entity. These tertiles are of unequal size but of equal value, meaning that they group all observations with the same value in the same bin (typically the lower one). Therefore, the number of (weighted) observations in each bin is different, but all observations with the same value of setting (or target group) size are grouped in the same bin.

Proportions of children with various characteristics in the setting (or target group)

Proportion variables were created to represent the share of children with specific characteristics (eight different characteristics captured in SL2G27a-h) relative to the total number of children in the setting (SL2G25) or the share of children with specific characteristics (five different characteristics captured in SS2G36a-e) relative to the total number of children in the target group (SS2G35). Each proportion was calculated as (SL2G27x/SL2G25) × 100 for the setting and (SS2G36x/SS2G35) × 100 for the target group. Values resulting from division by zero or exceeding 125% were set to missing. The threshold of 125% was chosen instead of 100% because, during the cleaning process, the interrelation between the pairs of variables was intentionally relaxed to retain as many valid values as possible. Categorical dummy variables were then generated from the proportions, coded as 0 for less than 10% and 1 for 10% or more. Specific characteristics could overlap, meaning a child could be counted in more than one category.

Staff working hours and reasons for part-time work

Staff working hours and reasons for part-time work refers to the current working hours of staff as full time or part time (variable: SS2G06) and, in case of part-time work, the reason for part-time work (variable: SS2G07). It has three categories: full time, voluntary part time and involuntary part time. An ECEC staff is considered to be working full time when they indicated full-time working hours in SS2G06. An ECEC staff is considered to be voluntarily working part time when they indicated either part time in SS2G06 or their answer was omitted or invalid, and answered they were only looking for part-time work in SS2G07. An ECEC staff is considered to be involuntarily working part time when they indicated either part time in SS2G06 or their answer was omitted or invalid, and answered they were only offered part-time work in SS2G07.

Type of management and profit status

The categorical variable “type of management and profit status” was derived to represent the combined classification of management type and profit status for each setting (SL2G19 and SL2G21). Three categories were generated: publicly managed settings, privately managed for-profit settings and privately managed non-profit settings. Settings were classified as publicly managed cases regardless of their profit status if the leader indicated public management in SL2G19.

Use of scales

Several scales are used in this report, in particular in regression analyses. TALIS Starting Strong 2024 scales were constructed using the same international model parameters to be statistically equivalent across countries and subnational entities, facilitating scale score comparability across ECEC systems. In general, scale scores created for TALIS Starting Strong 2024 can be used in different types of analyses, including comparisons of country and subnational entity means for each individual scale.

TALIS Starting Strong 2024 scale scores are standardised to have a standard deviation of 2 across all education systems participating in TALIS Starting Strong 2024. A value of 10 corresponds to the item mid‑point value of the response scale. A detailed description of the construction and validation of TALIS Starting Strong 2024 scales can be found in Chapter 11 of the TALIS 2024 Technical Report (OECD, forthcoming[1]).

Adequacy of physical space

The scale of “adequacy of physical space” (S2LSQSET; referred to as “Structural quality: Adequacy of setting” in the TALIS 2024 Technical Report (OECD, forthcoming[1])) was constructed using leader responses (“strongly disagree”, “disagree”, “agree”, “strongly agree”) to seven statements which apply to their ECEC setting (SL2G28A, SL2G28B, SL2G28C, SL2G28D, SL2G28E, SL2G28F and SL2G28G): “There is adequate ventilation for healthy airflow”; “There is an adequate amount of furniture for each child (e.g. chairs, tables)”; “Indoor space is arranged so that children can easily move around”; “The background noise level in the classroom/playgroup/group does not interfere with interactions”; “The roof, ceiling, walls and floor are in good shape (e.g. have no leaks, mould)”; “Children can access a bathroom easily”; and “Outdoor space is arranged so that children can easily move around”. Higher scores on this scale indicate more positive perceptions of the physical space in the ECEC setting.

Barriers to structural quality: Adequacy of physical resources

The scale of “barriers to structural quality: adequacy of physical resources” (S2LSQPRS) was constructed using leaders’ responses (“not at all”, “to some extent”, “quite a bit”, “a lot”) to four statements on issues with the availability of physical resources that hinder the setting’s capacity to provide a quality environment for children’s development, well-being and learning (SL2G42A, SL2G42B, SL2G42C and SL2G42D): “Shortage or inadequacy of play or learning materials (e.g. picture books, building blocks, paint)”; “Shortage or inadequacy of indoor space”; “Shortage or inadequacy of outdoor play space”; and “Shortage or inadequacy of digital resources and tools”. Higher scores on this scale indicate greater perceived barriers to structural quality related to the adequacy of physical resources.

Barriers to structural quality: Adequacy of staffing

The scale of “barriers to structural quality: adequacy of staffing” (S2LSQSTA) was constructed using leaders’ responses (“not at all”, “to some extent”, “quite a bit”, “a lot”) on staffing issues that hinder the setting’s capacity to provide a quality environment for children’s development, well-being and learning (SL2G42I, SL2G42J, SL2G42K, SL2G42L, SL2G42M, SL2G42N and SL2G42O): “<ECEC staff> absences”; “Lack of opportunities or support for <ECEC staff>’s professional development”; “Shortage of <ECEC staff> with competence in working with children from <socio-economically disadvantaged homes>“; “Shortage of <ECEC staff> with competence in working with children with special education needs”; “Shortage of <ECEC staff> with competence in working with children speaking a language other than the language(s) used in the <ECEC setting>“; “Shortage of qualified staff”; and “Shortage of staff relative to the number of enrolled children”. This scale is available only for pre-primary settings. Higher scores on this scale reflect greater perceived barriers to structural quality in relation to staffing adequacy.

Behavioural support in the classroom/playroom

The scale of “behavioural support in the classroom/playroom” (S2SOCIALR) was constructed using staff responses (“never”, “rarely”, “occasionally”, “often”, “always”) to five statements which apply to their work in the target group at their ECEC setting (SS2G41A, SS2G41B, SS2G41C, SS2G41D and SS2G41F): “I explained the rules children should follow at the start of a new activity”; “I addressed children’s disruptive behaviour by explaining why they should stop”; “I showed appreciation to children that behaved well during an activity”; “I helped children understand the consequences if they did not follow the rules”; and “I redirected a child to an appropriate activity if they were disturbing other children”. Higher scores on this scale reflect a greater level of behavioural support provided in the classroom/playroom.

Child-centred planning

The scale of “child-centred planning” (S2CCAFL) was constructed using staff responses (“not at all”, “once a week”, “a few times a week”, “once a day”, “more than once a day”) to four statements about how frequently they used the following practices in their work during the most recent complete calendar week (SS2G32D, SS2G32E, SS2G32F and SS2G32G): “Intentionally organised the <classroom/playgroup/group> to facilitate children’s exploration and play”; “Planned activities designed to allow children to take the initiative”; “Involved children in making plans for the day”; and “Adapted daily activities in response to children’s natural rhythms (e.g. hunger, fatigue)”. Higher scores on this scale indicate a greater frequency of use of child-centred approaches to planning.

Daily adaptive pedagogical practices

The scale of “daily adaptive pedagogical practices” (S2PRACTR) was constructed using staff responses (“never”, “rarely”, “occasionally”, “often”, “always”) to five statements which apply to their work in the target group at their ECEC setting (SS2G40A, SS2G40B, SS2G40C, SS2G40E and SS2G40F): “I explained how a new activity relates to children’s lives”; “I gave different activities to suit different children’s interests”; “I gave different activities to suit different children’s levels of development”; “I recognised children’s individual interests regardless of gender”; and “I adapted my activities to meet the needs of dual/second language learners”. Higher values on this scale reflect greater use of daily adaptive pedagogical practices.

Distributed leadership

The scale of “distributed leadership” (S2LDLEAD) was constructed using leaders’ responses (“not at all”, “to some extent”, “quite a bit”, “a lot”) to six statements on leadership practices (SL2G41A, SL2G41B, SL2G41C, SL2G41D, SL2G41F and SL2G41G): “<ECEC staff> have opportunities to actively participate in <ECEC setting> decisions”; “Parents or guardians have opportunities to actively participate in <ECEC setting> decisions”; “Children have opportunities to actively participate in decisions”; “There is a culture of shared responsibility for <ECEC setting> issues”; “There is a collaborative <ECEC setting> culture which is characterised by mutual support”; and “This <ECEC setting> encourages staff to lead new initiatives”. Higher scores in this scale indicate greater distributed leadership within the ECEC setting.

Engagement in collaborative practices

The scale of “engagement in collaborative practices” (S2COLLR) was constructed using staff responses (“never”, “less than monthly”, “monthly”, “weekly”, “daily”) to five statements on collaborative practices they engage in (SS2G42A, SS2G42B, SS2G42C, SS2G42D and SS2G42F): “Provide feedback to other <ECEC staff> about their practice”; “Engage in joint activities across different groups, including age groups”; “Exchange learning or pedagogical materials with colleagues”; “Co-design activities or content for children’s development, well-being and learning with colleagues”; and “Discuss the evaluation of children’s development, well-being and learning with other <ECEC staff>”. Higher values on this scale indicate a greater level of engagement in collaborative practices.

Facilitating emotional development

The scale of “facilitating emotional development” (S2PREMR) was constructed using staff responses (“not at all”, “once a week”, “a few times a week”, “once a day”, “more than once a day”) to three statements about how frequently they used the following practices in their work during the most recent complete calendar week (SS2G33A, SS2G33B and SS2G33D): “Talked with children about feelings”; “Helped children understand their feelings (e.g. when they are excited)”; and “Helped children express their feelings”. Higher values on this scale reflect more frequent activities to facilitate the emotional development of children.

Facilitating language development

The scale of “facilitating language development” (S2PRLANR) was constructed using staff responses (“never”, “rarely”, “occasionally”, “often”, “always”) to five statements which apply to their work in the target group at their ECEC setting (SS2G39A, SS2G39B, SS2G39C, SS2G39D and SS2G39F): “I encouraged children to lead conversations”; “I asked supportive questions to encourage children to give longer explanations”; “I supported children to express their ideas by asking them questions”; “I rephrased or repeated statements to make sure I understood the children”; and “I used words or phrasing that was new to children”. Higher scores on this scale reflect more activities aimed at facilitating language development among children.

Facilitating literacy development

The scale of “facilitating literacy development” (S2PROLIR) was constructed using staff responses (“not at all”, “once a week”, “a few times a week”, “once a day”, “more than once a day”) to seven statements about how frequently they used the following practices in their work during the most recent complete calendar week (SS2G29A, SS2G29B, SS2G29D, SS2G29E, SS2G29F, SS2G29G and SS2G29H): “Played word games with the children”; “Read books to children”; “Encouraged pre-verbal children to point to pictures when looking at a book”; “Sang songs with or to the children”; “Engaged children in activities that helped them pick out sounds in words (e.g. phonics games)”; “Used synonyms for words to help explain their meaning when talking with children”; and “Used activities for children to make marks (e.g. finger painting, pressing into dough/playdough)”. Higher scores on this scale reflect more frequent activities aimed at facilitating literacy development among children.

Facilitating numeracy development

The scale of “facilitating numeracy development” (S2PRONUR) was constructed using staff responses (“not at all”, “once a week”, “a few times a week”, “once a day”, “more than once a day”) to five statements about how frequently they used the following practices in their work during the most recent complete calendar week (SS2G30A, SS2G30B, SS2G30C, SS2G30D and SS2G30F): “Used sorting activities by shape or colour”; “Played number games”; “Sang songs with numbers”; “Helped children to use numbers or to count”; and “Encouraged children to measure items (e.g. while cooking or building with blocks)”. Higher values on this scale indicate more frequent engagement in activities that support children’s numeracy development.

Facilitating parent/guardian engagement

The scale of “facilitating parent/guardian engagement” (S2PARENR) was constructed using staff responses (“strongly disagree”, “disagree”, “agree”, “strongly agree”) to five statements on parent/guardian engagement practices as applied to their ECEC setting (SS2G43A, SS2G43C, SS2G43D, SS2G43E and SS2G43F): “Parents or guardians are informed about daily activities on a regular basis”; “Parents or guardians are encouraged by <ECEC staff> to participate in regular classroom/playgroup/group activities”; “Parents or guardians are encouraged by <ECEC staff> to contribute to classroom/playgroup/group level decisions”; “Parents or guardians are encouraged by <ECEC staff> to take on ECEC setting-level tasks”; and “Parents or guardians are given opportunities to actively participate in ECEC setting decisions”. Higher scores on this scale indicate that <ECEC staff> agree more strongly that parent/guardian engagement is supported in their ECEC setting.

Facilitating play

The scale of “facilitating play” (S2PRINTR) was constructed using staff responses (“not at all”, “once a week”, “a few times a week”, “once a day”, “more than once a day”) to three statements about how frequently they used the following practices in their work during the most recent complete calendar week (SS2G32A, SS2G32B and SS2G32C): “Allowed children to take the lead when I played with them”; “Responded positively to non-verbal invitations to play (e.g. a child passing me a toy)”; and “Let a child play alone if they were deeply engaged in play”. Higher values on this scale reflect more frequent activities to facilitate play among children.

Facilitating prosocial behaviour

The scale of “facilitating prosocial behaviour” (S2PRSOC) was constructed using staff responses (“not at all”, “once a week”, “a few times a week”, “once a day”, “more than once a day”) to four statements about how frequently they used the following practices in their work during the most recent complete calendar week (SS2G31A, SS2G31B, SS2G31C and SS2G31D): “Encouraged sharing amongst children”; “Encouraged children if they helped each other”; “Encouraged children playing in small groups to include other children”; and “Encouraged children if they comforted each other”. Higher values on this scale indicate a higher frequency of activities that facilitate prosocial behaviour among children.

Leader innovations in the ECEC setting

The scale of “leader innovations in the ECEC setting” (S2LEADSPL2) was constructed using leader responses (“never”, “less than monthly”, “monthly”, “weekly”, “daily”) to three statements on leadership practices they engage in (SL2G32F, SL2G32G and SL2G32H): “I worked on developing, revising or communicating a <vision> for this ECEC setting”; “I made changes in <ECEC setting> policies or practices based on monitoring results and external evaluation”; and “I facilitated improvements in the indoor or outdoor space used for children’s development, well-being and learning”. Higher scores on this scale reflect a greater frequency of innovative leadership practices within the ECEC setting.

Leader support for staff in pedagogical learning

The scale of “leader support for staff in pedagogical learning” (S2LEADSPL1) was constructed using leaders’ responses (“never”, “less than monthly”, “monthly”, “weekly”, “daily”) to four statements on how often they engaged in various leadership activities over the past year (SL2G32A, SL2G32B, SL2G32C and SL2G32D): “I collaborated with <ECEC staff> to improve how children play together”; “I observed <ECEC staff> practices and staff-child interactions”; “I provided feedback to <ECEC staff> based on my observations”; and “I took actions to support co-operation among <ECEC staff> to develop new approaches to ECEC practices”. Higher scores on this scale indicate more frequent leadership practices aimed at supporting the pedagogical learning of ECEC staff.

Level of stress and its impact on staff well-being

The scale of “level of stress and its impact on staff well-being” (S2STRESS) was constructed using staff responses (“not at all”, “to some extent”, “quite a bit”, “a lot”) to four statements about their experience as an <ECEC staff> at their ECEC setting (SS2G46A, SS2G46B, SS2G46C and SS2G46D): “I experience stress in my work”; “My job leaves me time for my personal life”; “My job negatively impacts my mental health”; and “My job negatively impacts my physical health”. Higher scores reflect greater perceived stress (and less perceived well-being).

Neighbourhood environment

The scale of “neighbourhood environment” (S2LNIGHBR) was constructed using leader responses (“strongly disagree”, “disagree”, “agree”, “strongly agree”) to five statements which apply to their ECEC setting (SL2G14A, SL2G14B, SL2G14C, SL2G14D and SL2G14F): “There is litter lying around”; “There is vandalism”; “People from diverse backgrounds experience insults or attacks (e.g. different ethnicity or culture, migrant background, sexual or gender minorities)”; “There are public places where children can play safely”; and “There are drug-related problems”. Higher values on this scale indicate more critical, or negative, perceptions of the neighbourhood environment.

Promoting awareness and acceptance of diversity

The scale of “promoting awareness and acceptance of diversity” (S2PAAD) was constructed using staff responses (“not at all”, “to some extent”, “quite a bit”, “a lot”) on the applicability of five statements to the ECEC setting (SS2G23A, SS2G23C, SS2G23D, SS2G23E and SS2G23G): “Books and pictures feature people from a variety of ethnic and cultural groups”; “Books and pictures feature people with different physical needs”; “Books and pictures feature diverse family structures (e.g. single parent families, families with parents from sexual or gender minorities)”; “Children play with toys and artefacts from cultures other than the ethnic majority”; and “Activities aim to raise awareness about the different languages of the <dual/second language learners> in the <classroom/playgroup/group>“. Higher scores on this scale indicate a stronger emphasis on promotion awareness and acceptance of diversity.

Satisfaction with setting leadership

The scale of “satisfaction with setting leadership” (S2SSLEAD) was constructed using staff responses (“strongly disagree”, “disagree”, “agree”, “strongly agree”) to seven statements on the ECEC setting leadership (SS2G44A, SS2G44B, SS2G44C, SS2G44D, SS2G44E, SS2G44F and SS2G44G): “The <ECEC setting leader> has a clear vision for this ECEC setting”; “The <ECEC setting leader> encourages co-operation among <ECEC staff> to develop new ideas in their practices”; “The <ECEC setting leader> encourages <ECEC staff> to take responsibility for improving their practices”; “The <ECEC setting leader> encourages all staff to have a say on important decisions”; “The <ECEC setting leader> promotes a culture of shared professional learning”; “The <ECEC setting leader> communicates effectively with <ECEC staff>“; and “I need more support from my <ECEC setting leader>“. Higher scores on this scale reflect greater satisfaction with the leadership at the ECEC setting.

Satisfaction with work at this ECEC setting

The scale of “satisfaction with work at this ECEC setting” (S2JOBSAR) was constructed using staff responses (“strongly disagree”, “disagree”, “agree”, “strongly agree”) to four statements about their experience as an ECEC staff at their ECEC setting (SS2G45C, SS2G45E, SS2G45F and SS2G45G): “I enjoy working at this <ECEC setting>”; “I am satisfied with the support that I receive from parents or guardians in this <ECEC setting>“; “I would recommend this <ECEC setting> as a good place to work”; and “I would like to change to another <ECEC setting> if that were possible”. Higher scores indicate greater satisfaction with work at the ECEC setting.

Self-efficacy, overall

The scale “self-efficacy, overall” (S2SELF) was constructed using staff responses (“not at all”, “to some extent”, “quite a bit”, “a lot”) to ten statements about the different tasks they could do in their work with children (SS2G24A, SS2G24B, SS2G24C, SS2G24D, SS2G24E, SS2G24F, SS2G24G, SS2G24H, SS2G24I and SS2G24J): “Support the development of children from disadvantaged backgrounds”; “Adapt your work to individual child needs”; “Help children develop their capacity to learn independently”; “Help children prepare for transitioning to other <ECEC settings> or to <ISCED 2011 level 1> schools”; “Stimulate children’s interest in cultural differences and commonalities”; “Help children to interact with each other and show good social behaviour (e.g. sharing, helping others)”; “Help children develop creativity and problem solving”; “Use <digital resources and tools> to facilitate working with children”; “Provide all children with a feeling of security”; and “Use a variety of monitoring and assessment strategies to follow children’s development, well-being and learning”. Higher values on this scale indicate greater perceived self-efficacy.

Sources of work stress: Aspects related to capacity, resources and time

The scale of “sources of work stress: aspects related to capacity, resources and time” (S2STRESO) as constructed using staff responses (“not at all”, “to some extent”, “quite a bit”, “a lot”) to five statements about the sources of stress at their ECEC setting (SS2G47C, SS2G47D, SS2G47G, SS2G47J and SS2G47L): “Having too much administrative work to do (e.g. filling out forms)”; “Having extra duties due to absent <ECEC staff>“; “Addressing parent or guardian concerns”; “Having too many tasks to do at the same time”; and “Lacking resources (e.g. financial support, material resources)”. Higher scores in this scale indicate greater perceived stress in terms of capacity, resources and time.

Sources of work stress: Aspects specific to children’s development, well-being and learning

The “scale of sources of work stress: aspects specific to children’s development”, well-being and learning (S2STRESC) as constructed using staff responses (“not at all”, “to some extent”, “quite a bit”, “a lot”) to five statements about the sources of stress at their ECEC setting (SS2G47E, SS2G47F, SS2G47H, SS2G47I and SS2G47K): “Being held responsible for children’s development, well-being and learning”; “Managing <classroom/playgroup/group> behaviour”; “Accommodating children with special education needs”; “Having too many children in my <classroom/playgroup/group>“; and “Having to adapt my practices with children due to unexpected disruptions (e.g. natural disasters, public health emergencies/pandemics, humanitarian crises”. Higher scores indicate greater perceived stress in relation to children’s development, well-being and learning.

Reference

[1] OECD (forthcoming), Teaching and Learning International Survey (TALIS) 2024 Technical Report, OECD Publishing, Paris, forthcoming.