CHANGING PATTERNS OF TRANSITION TO ADULTHOOD IN MOLDOVA BEFORE AND AFTER THE 1990S TRANSFORMATION

DOI: https://doi.org/10.36004/nier.es.2024.2-07

JEL Classification: C38, J11, J13, P36

UDC: 316.346.32-053.6(478)

Maxim SLAV

National Institute for Economic Research, Academy for Economic Studies of Moldova

https://orcid.org/0000-0002-3061-1692

slav.maxim@ase.md

Received 15 November 2024

Accepted for publication 2 December 2024

SUMMARY

The transition to adulthood (TA) patterns are often classified into two broad categories: traditional and modern. This classification is usually confirmed by the research of trends of the age of transition events in developed countries. Fewer studies concentrate on or cover post-socialist societies. Both case and comparative studies suggested a more complex picture and challenged a uniform move from traditional to modern TA. This study explores the Moldovan society by classifying the patterns of the cohorts born in the pre-Gorbachev Soviet Era. The Latent Profile Analysis is applied to the Gender and Generation Survey 2020. To account for the gender- and period-specific trends, the sample was divided into four subsamples and analyzed independently. This allowed us to detect the clusters of TA in the continuum between the modern and traditional patterns. The changes in the patterns’ timing and frequencies were analyzed. The study detects a substantial gender difference in trends. The male TA patterns witnessed slow and robust modernization. At the same time, the female TA patterns saw traditionalization and its reversal for the last Soviet-born birth cohorts. The analysis also suggests that the trends started and developed before the 1990s transformation. The study adds to our understanding of TA patterns in developing Europe and gives several methodological suggestions for further research.

Keywords: Transition to adulthood, life course, gender, generations, Latent Profile Analysis, Moldova

INTRODUCTION

There are various approaches to transition to adulthood (TA). This paper contributes to the line that focuses on the timing of events. Classically, the list has six positions: partnership, marriage, and childbirth, which are parts of the demographic transition (Mitrofanova, 2019), and finishing education, first employment, and leaving the parents, which are parts of the socio-economic transition. The list is not universal. The individuals also value them unequally (Nelson, 2009). For the reasons discussed in the methodology section, this study focuses on education, marriage, and childbirth. The research questions are as follows. What were the types of TA in Moldova among those born between the Interbellum and Perestroika? What were the trends in timing for the types? What are the trends in frequencies?

The means are the most usual way to analyze trends. For example, Billari and Liefbroer (2010) suggested that the TA in Europe can be analyzed with two poles: the traditional and the modern one. The traditional TA is defined as “early, contracted, and simple” (p. 60), and the modern one is “late, protracted, and complex” (p. 60). These findings were also supported by theoretical considerations, specifically, the Second Demographic Transition theory (Lesthaeghe & Van de Kaa, 1986). The theory links the overarching changes in demographic behavior, primarily fertility, with the large-scale values shift towards individualism, post-materialism, and the worldviews that allow for greater uncertainty (Lesthaeghe, 2014, Lesthaeghe, 2020; Sobotka, 2008). While the theory is criticized for its poor applicability to developing countries outside of Europe and “new Europes” a presupposition that the changes happen between long-term stable stages of development, and blinds spots linked with the sustainability of the traditional values or re-traditionalization, it is agreed that the theory is at least partially correct (Zaidi & Morgan, 2017). Other definitions of traditional and modern TAs exist, but they typically overlap. For example, in the post-Soviet context, Mitrofanova (2019, pp. 189–190) defined female traditional TA as a condensed demographic transition that is done to the detriment of the economic transition (i.e., the former is short and intense, while the latter is typically postponed, especially if there was a child). The male traditional TA consists of early economic transition and postponed demographic one, with both intense. Thus, Mitrofanova defined TA modernization as the convergence of the patterns rather than postponement for both genders.

However, the trends in means are undesirable for several reasons. First, they may erase the less pronounced but existing trends. Even if a new pattern overtakes an old one, both persist for some time (Raz-Yurovich, 2017 argues for it theoretically; Sobotka, 2008 shows it regarding the Second Demographic Transition). The trend in means may not show the earliest signs, as they would only be outliers. The less pronounced trends are erased this way. The cluster analysis allows for evading this problem. While TA is a priori unpredictable (Dennison, 2016), classifying it has three major advantages. First, events frequently happen in sets. Shotgun marriages and failure to continue education due to early pregnancy are one of the examples. In the context of the theory of TA modernization, it would correspond to determining the timing and the variation of the event. Second, aggregating similar TA patterns will show changes in TA composition over time. Some trends can be overarching, e.g., attributed to the structural environment (e.g., raising costs of having children, Werding, 2014) or related to the changes in individuals’ values (Sobotka, 2008). However, they may have differing effects on the individuals. For example, economic hardships are known to put contradictory pressure on the values of traditional men and women (Enneli & Enneli, 2017). Third, TA has several strong predictors, including the parental socio-economic position (SEP). Billari et al. (2019) exhaustingly showed that SEP influences not only the opportunities but also the desires and motivation of the individuals. Overall, it can be said that these mechanisms were found in every country under consideration (France, Austria, and Bulgaria).

The rest of the paper is organized as follows. In the literature review, the results of the previous cluster analyses of TA in other societies are listed. Next, in the Methodology section, the data, its subsampling, and the choice of the clustering methods are discussed in detail. Next, the results are presented, and the trends that they reveal. Finally, the profiles and trends are compared to those in other countries. The limitations of the method and the alternatives are also covered in the last section.

LITERATURE REVIEW

In this section, the profiles from the previous research are reviewed. The findings are scarce for Moldova, especially due to the data availability, but the research on the other countries to the East of the Hajnal line (Hajnal, 1965) may add context.

Crismaru (2024) is the only researcher who classified TAs in Moldova. She based her study on the GGS survey and focused on the socio-economic positions of those aged 35 and younger. She primarily focused on the (un-) privileged positions of the youth and the (binary) facts if, as of the moment of the interview, they were married, had a child, had a job, etc. She found four types of youth. Those having complex transitions are predominantly employed and tend to be men, live in an urban area, and have university degrees. On the other pole, there is precarious transition. On average, this cluster’s members have lower education and are unemployed. The third pattern holds an intermediate position regarding education but also has a majority of women on maternity leave. Finally, the fourth class predominantly consists of female homemakers. The gender compositions of the contemporary TA clusters show that gender differences also exist in Moldova. For example, an average male with uncertain transition graduates by the age of 18, marries 6 years later, and has a firstborn in less than a year. An average female with uncertain transition graduates when 19 years old, marries in 2.5 years, and gives birth in less than a year. Crismaru achieves this by stressing the individuals’ socio-economic position. Note that in this case, the female graduations happen almost 1.5 years after the male ones, which is the reverse of the classical gender gap in events’ timing. In the context of patterns’ classification by timing alone, these two patterns may be viewed as separate and distinct. It is thus more plausible to divide the sample by gender.

On the Russian data, Mitrofanova (2019) showed that the TA differs by the area of residence (rural vs. urban) and by the level of education. She suggested that the convergence with the West European (modern) TA can be found among those born in the 1980s, but those from urban and more educated strata exhibited it earlier. She also suggested three types of transition. The 1930s-1960s birth cohorts exhibited traditional TAs, the 1980s having the modern ones, and the 1970s having the transitional type of TA with the risk of marriage significantly rising due to norms and laws shifting overnight. The TAs were different for different genders. Women tended to have a fast demographic transition to the detriment of the socio-economic one. For men, the socio-economic transition preceded the demographic one. At the same time, for both genders, prolonging and postponement existed, which means that the Russian TA modernization was about inter-gender convergence. While its beginning, overall, shifted for the later generations, the transition became more and more intensive, with the 1970s generations having it the shortest (many of the 1980s generations had not finished their TA by the moment of the data collection). Earlier, economic hardships were found to intensify TAs (Enneli & Enneli, 2017), but even though the 1970s generations experienced the 1990s economic crisis, the intensification these individuals experienced looks like a part of a trend.

Lesnard et al. (2016) have similar findings for some post-socialist countries. There, especially in Bulgaria, Slovakia, and to lesser extents Poland, Hungary, and Estonia, shift towards the patterns that skip the period of living alone, having a job, or the whole socio-economic transition. At the same time, most of the Central and Northern European countries converge towards the patterns that value independence, i.e., leaving the parents, finding a job, and finishing education. A later comparative study of TA confirmed that the post-socialist countries (Russia-Estonia) are still relatively closer to each other than to the Western countries in spite of cultural differences (Mitrofanova, 2023; Mitrofanova & Makarov, 2023). The researchers allow for the possibility of East-West non-convergence because of the cultural and social differences (Billari et al., 2019) but generally accept that Europe to the East of the Hajnal line participates in the Second Demographic Transition (Sobotka, 2008; Zakharov, 2008). If the traditional vs. modern duality is accepted, the comparative European studies show that in the 1930-1950s birth cohort, the average TA was becoming more traditional. At the same time, for some genders, countries, and events, the cohort with the earliest and the most uniformly invariant timing could be the 1940s or 1960s birth cohort.

To sum up, the literature review shows that the patterns of TA substantially changed for the 1940-1980s birth cohort. Even accounting for the variance, two poles can be found. The traditional TA is early and short, and the modern TA is late and protracted. Overall, the countries moved in the direction of the second type of TA. However, the comparative studies also showed that the trends are different in different cultures. The TAs in the post-socialist countries are earlier and more congested compared to those in the West. The studies also show that the genders must have different paths to the modern transition to adulthood, with the female TA having to change more drastically. The Moldova-specific study showed that the women may have drastically diverging paths, with some having radically traditional and others having radically modern TA.

Overall, it is hypothesized that the Moldovan 1940-1980s generations will show no less than two TA patterns: traditional and modern. The modern patterns happen later and are longer than the traditional ones. As the more chaotic TA pattern, the variance of the events’ timing should also be higher for the modern profile. There may be two objectives. First, the modern TA of past generations can overlap with the traditional TA of the contemporary generations. This is the case if the modern TA also existed in the earlier cohorts and both modern and traditional poles were under the influence of the trends. The probability of this happening increases if one accounts for gender (i.e., modern female TA of the past can overlap with traditional TA of the present). This suggests dividing the sample by gender and generation. Second, I also expect the set of patterns to not be limited to two because of the generations that had their TA during the 1990s transformation.

METHODS

To check the hypothesis, I used the Gender and Generation 2020 Survey (GGS, 2020) in Moldova. The oldest individuals were aged 79, which corresponded to the 1941 birth year. The cut-off point for the youngest generation was set at 35 years old, corresponding to the 1985 birth year cohort. In this sense, the study is complementary to the one by Crismaru (2024), who considered individuals aged 35 or younger. The cut-off point was introduced due to the issue of self-censorship, which is described to a greater extent below. As the literature review showed, for each TA profile, there is a gap between men and women vis-à-vis the timing of their events. Thus, the sample is divided into four by gender and the birth cohort. This allows dividing between those who experienced the 1990s economic and political crisis during their TA and those who did not. To do this, I used the share of people whose last event happened after December 31, 1991. January 1, 1961, was chosen as the division date as for the 1960 birth cohort, the share was 9%, and for the 1961 cohort, it was 13%, i.e., exceeded 10%. For shortness, the 1941 – 1960 cohort will be referred to as Early, and the 1961 – 1985 cohort will be referred to as Late.

GGS in Moldova gives information on four events of the respondents: the first partnership (with the exception of several special cases, it is coded as lhi04_1 or dem30b), the first marriage (dem_28b or lhi_05b), the first childbirth (lhi29 with a condition on childbirth lhi26), and the education ending (dem_08). The data on the first partnership were not used as a result for three reasons. First, it is highly correlated with the data on marriage, with Pearson’s correlation varying between 0.81 and 0.92 for some subsamples (see Fig. 1 for one of the correlation plots). This leads to the lack of convergence[1] of the models that take correlation into account and additional weight to marriage for the other models. Second, it is incomplete, and due to the way the questionnaire was constructed, we do not know when some of the respondents had their partnership started (specifically, in the cases when the respondent lives separately from their spouse). Third, partnerships were stigmatized in the Soviet Union (Mitrofanova, 2019), which implies that the event was not a milestone for the older generations (and was united with marriage), and the data on it is subject to (deliberate or unconscious) self-censorship. Even though two other very useful milestones (first employment and separation) are not covered here, one can still spot and classify patterns of transition with education, marriage, and childbirth.

Figure 1. Scatter Plot of the Age of Partnership and the Age of Marriage. Pearson correlation = 0.92.

Source: Author’s Own Calculations Based on Gender and Generation Survey, 2020, for 1941 – 1960 Males.

The Latent Class family model was chosen because it consistently performs better than other clustering techniques, including K-means and HCPC (Nadeekantha et al., 2023; Preud’homme et al., 2021). It is a more classical approach to use longitudinal versions of Latent Class Analysis (Helske & Helske, 2019). The greatest disadvantage of the Hidden Mixture Markov Chains is the assumption of the independence of events. In many instances, this is not the case. Covariances are also likely to be different for different transition patterns. For example, marriages are likely to have a stronger link with childbirth in traditional patterns than in modern ones. Latent Profile Analysis Model 6 (Rosenberg & Van Lissa, 2018) allows for this flexibility. As I show in the Analysis section, this addition sometimes leads to ambiguity but is generally well-founded.

There are two problems arising from the Latent Profile Analysis. First, the number of profiles is likely to be overinflated due to the skewness of the data. LPA assumes that the distribution of variables (inside one profile) is normal (non-skewed), which is sometimes impossible. This is exemplified below, in the Fig. 2 and 3, which show the data for the early male cohort. Specifically, the pink and the red profiles (closer to the lower values) represent a traditional TA: early graduation (the age of 15 for the red one, the age of 17 for the pink one), marriage at 22, and childbirth two years later (for both profiles). In the scope of this study, this is a substantial reason to treat them as one. An additional argument in favor of it is the fact that the red profile accounts for 4% of the subsample, which is acceptable but quite undesirable. Finally, when facing such profiles, I also account for the events’ variance and timing distribution. In the context of the study, any event’s variance higher than 40 means that the event’s timing is unpredictable.

Figure 2. An Example of Profile Overinflation. Bars: Graduation Dates. Distribution Functions: Hypothetical Seven Profiles Model.

Source: Author’s Own Calculations Based on Gender and Generation Survey, 2020, for 1941 – 1960 Males.

Figure 3. An Example of Profile Overinflation. Bars: Marriage Dates. Distribution Functions: Hypothetical Seven Profiles Model.

Source: Author’s Own Calculations Based on Gender and Generation Survey, 2020, for 1941 – 1960 Males.

Second, it does not allow for missing data, and the data for the three events had to be imputed. The data are unlikely to be missing at random (i.e., it is more likely that people who never marry are from the modern profiles), which leads to a bias (Heymans & Twisk, 2022). This was also the reason for cutting off the individuals aged 34 or younger. The data on their TA should be imputated to a far greater extent, and it will be based on the previous cohorts’ data alone. It will thus lead to a bias and add less substantial information. Recognizing that it can skew the results, the data on the missing variables is presented in Table 1.

With the algorithm exemplified above, the number of profiles was chosen for each subsample (see Table 2 for the fit indices). In every case, AIC suggests more profiles. However, for each subsample, the models with seven profiles and more had at least one unacceptably small profile (i.e., the size is 3% or less, which could only be acceptable if the entropy was 80% or more (Spurk et al., 2020)).

Table 1. Missing Values Distribution Across Subsamples.

Source: Author’s Own Calculations Based on Gender and Generation Survey, 2020.

Table 2. Fit indices for the Latent Profile Analysis Models with Covariates for Four Subsamples.

Source: Author’s Own Calculations Based on Gender and Generation Survey, 2020.

Analysis

Tables 3 and 4 depict the means and the variances of the events for men and women, by profile and generation. Figures 5 and 6 display them, aggregated by five-year cohorts. In most cases, the covariances were only significant between the marriages and childbirth, and the data on them is not provided here.

Table 3. Profiles of TA for men. Mean values of the events’ timing are without parentheses and the variance of the events’ timings are in parentheses. Note that these are the events’ variances, and not the means’ variances.

Source: Author’s Own Calculations Based on Gender and Generation Survey, 2020.

Table 4. Profiles of TA for women. Mean values of the events’ timing are without parentheses and the variance of the events’ timings are in parentheses. Note that these are the events’ variances, and not the means’ variances.

Source: Author’s Own Calculations Based on Gender and Generation Survey, 2020.

The most standing out profiles are the Unpredictable ones, whose demographic transition was very late and uncertain. The variance of their education timing was also relatively high for both genders. While it is suggested to primarily consider the profile as technical (i.e., there may be many patterns inside of it, but the number of individuals does not allow for detecting them), the profile shows that in the earlier cohorts, there was a great number of paths that one could have turned to. One of such diversities is suggested by the fact that the mean timing of childbirth for the Unpredictable women born in 1941-1960 is 1.5 years lower than the timing of their marriage.

For both genders, one can find corresponding profiles in both cohorts by the timing and the timing’s variance. In some cases, the indicators virtually did not change. For example, in both cohorts, the intermediate male profile is finishing education at 19, marrying in early 23, and having a child by the age of 24. For some profiles, the variance was the only criterion to classify an individual. For example, the Great Delay profile consists of those having a high variance of demographic events and a low variance of (typically, secondary) education. The Resuming Education profile unites those with the average timing of demographic transition and highly unpredictable timing of education. Notice that the profile is quite different for men and women. For the former, the mean age of re-education fell from 26 to 23.1, while for the latter, the mean age rose from 23 to 28. The Unpredictable profile unites those with high variance for all the events and without a well-defined pattern of TA.

The Unpredictable profile is the only profile that does not have a corresponding profile in the later generation. For both genders’ late cohort, there is a profile without a clear pattern of demographic transition, but their education timing was rather stable and early. This suggests that the Unpredictable ones are actually delaying the demographic transition, but this group contains those who finished school early and those who resumed their education. For men, the latter subgroup vanished. Combined with the fact that the female Resuming education profile declined significantly after the 1960/1961 boundary, this highlights the decline in the number of those coming back to education among women.

When evaluating the changes in the corresponding profiles, one can highlight several trends. For men, the timing of the demographic transition rose for the Early School and Resuming education profiles by 1.5-2 years. At the same time, there were no changes for the Intermediate ones, which makes them the most stable male profile. Note that if, for the early cohort, the Intermediate profile was in between what can be called the Traditional profile and the more uncertain ones, this position disappeared in the late cohort. On the contrary, for the Late cohort, there was an inter-profile negative correlation between the timing of graduation and the timing of the demographic events, as the profiles that have early graduations have relatively late marriages and childbirths and vice versa. For women, on the contrary, the timing of the demographic events, on average, fell for the Traditional and Intermediate profiles by around 1 year. To some extent, this change was compensated by the changes in the composition, as the share of the Intermediate TA rose and the share of the Traditional TA fell after the 1960/1961 boundary (see Figure 6). In other words, there is a substantial share of TAs that can be labeled as Traditional before the 1960/1961 boundary that are labeled as Intermediate after it. However, the changes are not a side effect of a methodological nature (see Figure 4, which shows that the distributions for both Intermediate and Traditional profiles shift towards younger ages). Such changes show that it was reasonable to divide between men and women and to expect different trends.

Figure 4. The Distribution of the Marriage Timing for the Five-Year Cohorts to Both Sides of the 1960/1961 Boundary.

Source: Author’s Own Calculations Based on Gender and Generation Survey, 2020.

Figure 5. Shares of the Male Profiles, 1941-1960 and 1961-1985 cohorts.

Source: Author’s Own Calculations Based on Gender and Generation Survey, 2020.

Figure 6. Shares of the Female Profiles, 1941-1960 and 1961-1985 cohorts.

Source: Author’s Own Calculations Based on Gender and Generation Survey, 2020.

The female Resuming education and Intermediate patterns are the only patterns with a growing age of graduation. On average, the index fell for men (20.3 years in the first cohort vs. 18.5 years in the second cohort) and stayed the same for women (19 years). While this can be explained by the shifts in the Soviet re-education policies (see above), this can also be the result of sampling (i.e., the cutting-off of the youngest generations) and migration.

When analyzing the changes in profile composition, one can notice that the male profiles show a greater level of continuity than the female ones. While the people born in 1960 and those born in 1961 can be expected to have very similar profiles of TA, they are analyzed together with different peer groups: the former are analyzed together with those born in the 1940s and the latter are analyzed with those born in the 1980s. For men, this did not influence the profiles’ composition, and the 1960/1961 boundary was passed without visible discontinuity (see Figs. 4 and 5). For women, the composition changed more visibly. The Early School pattern was split. Most of this profile’s members had their demographic transition earlier, shifting both marriage and childbirth one year towards lower ages. Note that in both cohorts the events’ timing is quite low-variance; together with mean timing, this allows classifying them as the traditional profiles. Similarly, the Resuming Education profile split and the mean timing of their education shifted towards higher ages. Overall, while there is a justification for such a shift, this suggests that one needs a greater number of female sub-cohorts for further analysis.

One can find the trends in the popularity of certain transition patterns. For example, the number of those re-educating sharply fell for men from 41.5% to just 15.6% between 1940 and 1961 and continued to fall further to just 8.1% for those born in 1966-1970. The share then slowly grew to 20% for the 1981-1985 cohort. The last two cohorts were 25 between 1991 and 1995 and between 2006 and 2010, correspondingly, and thus had differing re-education opportunities. However, the sharp decline in the beginning must correspond to the changes in the Soviet education policy. This early decline is virtually non-existent for the female “Resuming education” profile, and, as discussed above, the profile virtually evaporates.

CONCLUSION AND DISCUSSION

In this study, the Transition to Adulthood for the Moldovan 1941-1985 birth cohorts was analyzed. Rather than accepting that society uniformly moves from one pole (traditional TA) to another (modern TA), the TA patterns were classified by the events’ timing and the timings’ uniformity. The TA patterns were classified for two genders and two cohorts (1941-1960 and 1961-1985) separately, and for each subsample, four patterns were found, including early school, intermediate, and re-educational ones. The main novelty of this study is that it assumes that varying profiles existed in every birth cohort and that the profiles could have experienced varying trends, including (but not necessarily) modernization. The study showed that modernization happened indeed for some male profiles, but some other strata were unaffected by it. It also showed that the female pattern experienced traditionalization of their demographic patterns with one unsubstantial exception. At the same time, the composition of TA between the birth cohorts changed significantly over time, and the inter-gender differences in trends could have been hidden by such complementary compositional changes.

For both men and women, there was a noticeable group of people without a certain Transition to Adulthood pattern. For both genders, there was a stabilization of this pattern’s graduation timing at earlier ages, but the timing of their marriage and childbirth was still uncertain. For men, the timing of education declined for all profiles. The share of the education-intensive profiles also declined until the 1981-1985 birth cohort. At the same time, the timing of the demographic events grew for the traditional and the resuming education profile. The Intermediate profile was the most stable for men whose transition, on average, spanned for five years. For women, the inter-generational continuity was less pronounced. Still, one can observe an increase in the graduation timing for all profiles (except for the Unpredictable one) and a fall in the demographic events’ timing (except for those resuming education).

In some respects, the results of the study are in agreement with the previous studies’ results. It shows that the most traditional birth cohort in Moldova was those born in the 1960s. After that, the proportion of those graduating and marrying in the later ages grew. The Russian data also shows that the earlier generations had a catch-up education for both genders but especially for women (Mitrofanova, 2019; Zakharov, 2008). Both genders enjoyed re-education in Moldova as well, but it was more popular among men of the 1940s birth cohort, and the share of women participating in the programs fell drastically after 1960. It also suggests that there is modernization of TA in Moldova, although the matter of the degree of convergence is out of the scope of the paper.

In this study, there was an attempt to differentiate between those virtually unaffected by the 1990s transformation and the rest. This delimitation of the cohorts naturally affected the profiles’ calculations. It is plausible to suggest that those born in 1960 bear little difference from those born in 1961. However, the former were aggregated with those born in the 1940s and defined together with them, and the latter were aggregated with those born in the 1980s. This had a limited effect on the male profiles calculation, e.g., one can see that those finishing their education early started delaying it, but overall, one can also see continuity. It had a far more visible effect on the female profiles. While this is a natural part of the method, this result indicates that the aggregation proposed in this research should be revisited. One of the possible ways to do so is to subdivide them into ten-year groups. Note that some of the profiles had an unequal distribution over the years. If one divides the generations further, there is a risk the profiles become invisible.

REFERENCES

Billari, F. C., Hiekel, N., & Liefbroer, A. C. (2019). The Social Stratification of Choice in the Transition to Adulthood. European Sociological Review, 35(5), 599-615. https://doi.org/10.1093/esr/jcz025

Billari, F. C., & Liefbroer, A. C. (2010). Towards a new pattern of transition to adulthood? Advances in Life Course Research, 15(2-3), 59-75. https://doi.org/10.1016/j.alcr.2010.10.003

Crismaru, M. (2024). Experiences of Youth in Transition to Adulthood: The Models Developed in Base of Generation and Gender Survey. Economy and Sociology, 2, 65-75. https://doi.org/10.36004/nier.es.2023.2-07

Dennison, R. P. (2016). Transition to Adulthood. In: The Wiley Blackwell Encyclopedia ofFamily Studies (pp. 1-5). https://doi.org/10.1016/B978-0-323-60984-5.00062-7

Enneli, H. Ç., & Enneli, P. (2017). Reinforcement and erosion of traditional gender roles among young people in a poor metropolitan area of Turkey. Journal of Youth Studies, 20(3), 349-365. https://doi.org/10.1080/13676261.2016.1217320

Hajnal, J. (1965). European marriage patterns in perspective. In: D. Glass & D. Eversley (Eds.), Population in history: Essays in historical demography. Aldine Publishing Company. https://u.demog.berkeley.edu/~jrw/Biblio/Eprints/%20G-I/hajnal.1965.european.marriage.pdf  

Helske, S., & Helske, J. (2019). Mixture Hidden Markov Models for Sequence Data: The seqHMM Package in R. Journal of Statistical Software, 88(3), 1-32. https://doi.org/10.18637/jss.v088.i03     

Heymans, M. W., & Twisk, J. W. R. (2022). Handling missing data in clinical research. Journal of Clinical Epidemiology, 151, 185-188. https://doi.org/10.1016/j.jclinepi.2022.08.016

Lesnard, L., Cousteaux, A.-S., Chanvril, F., & Le Hay, V. (2016). Do Transitions to Adulthood Converge in Europe? An Optimal Matching Analysis of Work-Family Trajectories of Men and Women from 20 European Countries. European Sociological Review, 32(3), 355-369. https://doi.org/10.1093/esr/jcw003

Lesthaeghe, R. (2014). The second demographic transition: A concise overview of its development. Proceedings of the National Academy of Sciences, 111(51), 18112-18115. https://doi.org/10.1073/pnas.1420441111

Lesthaeghe, R. (2020). The second demographic transition, 1986-2020: Sub-replacement fertility and rising cohabitation-a global update. Genus, 76(1), 10. https://doi.org/10.1186/s41118-020-00077-4

Lesthaeghe, R., & Van de Kaa, D. J. (1986). Twee demografische transities. In: R. Lesthaeghe & D. J. Van de Kaa (Eds.), Bevolking Groei of Krimp? (Vol. 61, pp. 9-24). Van Loghum Slaterus.

GGS (2020). Republic of Moldova – Generations and Gender Study. Ministry of Labor and Social Protection of the Republic of Moldova, National Bureau of Statistics (study sample), UNFPA, NIDI-GGP (partner and distributor). [Dataset].     

Mitrofanova, E. (2019). Socio-Demographic Aspects of Entering Adulthood for Russians Born in 1930-1986. HSE University.

Mitrofanova, E. (2023). Calendars and sequences of events during the transition to adulthood among different generations in France, Estonia and Russia. Demographic Review, 10(1), 108-135. https://doi.org/10.17323/demreview.v10i1.17262

Mitrofanova, E., & Makarov, S. (2023). West and East: Convergence or divergence of Millennials’ transition to adulthood in four European countries. Population and Economics, 7(4), 68-90. https://doi.org/10.3897/popecon.7.e112452

Nadeekantha, H. A. D. D., Kavinga, H. W. B., Gunawardana, A., & Dissanayaka, D. M. P. V. (2023). A comparison of distance-based and model-based clustering methods. In: International Conference on Applied and Pure Sciences (p. 97). Faculty of Science, University of Kelaniya, Sri Lanka. https://colab.ws/articles/10.1109%2Fscse61872.2024.10550585#

Nelson, L. J. (2009). An examination of emerging adulthood in Romanian college students. International Journal of Behavioral Development, 33(5), 402-411. https://doi.org/10.1177/0165025409340093

Preud’homme, G., Duarte, K., Dalleau, K., Lacomblez, C., Bresso, E., Smaïl-Tabbone, M., Couceiro, M., Devignes, M.-D., Kobayashi, M., Huttin, O., Ferreira, J. P., Zannad, F., Rossignol, P., & Girerd, N. (2021). Head-to-head comparison of clustering methods for heterogeneous data: A simulation-driven benchmark. Scientific Reports, 11(1), 4202. https://doi.org/10.1038/s41598-021-83340-8

Raz-Yurovich, L. (2017). Vanguards, Laggards, and Isolates: Outliers and Demographic Research. SSRN Electronic Journal. https://doi.org/10.2139/ssrn.3050363

Rosenberg, J. M., & Van Lissa, C. (2018). TidyLPA: Easily Carry Out Latent Profile Analysis (LPA). Using Open-Source or Commercial Software. [Dataset]. https://doi.org/10.32614/CRAN.package.tidyLPA

Sobotka, T. (2008). The diverse faces of the Second Demographic Transition in Europe. Demographic Research, 19, 171-224. https://doi.org/10.4054/DemRes.2008.19.8

Spurk, D., Hirschi, A., Wang, M., Valero, D., & Kauffeld, S. (2020). Latent profile analysis: A review and “how to” guide of its application within vocational behavior research. Journal of Vocational Behavior, 120, 103445. https://doi.org/10.1016/j.jvb.2020.103445

Werding, M. (2014). Children are costly, but raising them may pay: The economic approach to fertility. Demographic Research, 30, 253-276. https://doi.org/10.4054/DemRes.2014.30.8

Zaidi, B., & Morgan, S. P. (2017). The Second Demographic Transition Theory: A Review and Appraisal. Annual Review of Sociology, 43(1), 473-492. https://doi.org/10.1146/annurev-soc-060116-053442

Zakharov, S. (2008). Russian Federation: From the first to second demographic transition. Demographic Research, 19, 907-972. https://doi.org/10.4054/DemRes.2008.19.24