2 Methodological framework: sampling frame, stratification variables, sample size and proportionality
To identify a restricted number of episodes that could well represent contemporary Italian TV seriality, the population had to be identified, that is, the complete number of occurrences (in our case: episodes) produced and distributed within a specific time span. The completed list of TV series produced was collected from the Internet Movie Database (IMDb, https://www.imdb.com/) by conducting an advanced search for TV series and TV miniseries, using 1980 as the lower time threshold. The data were obtained through a data mining process, followed by several pre-processing steps during the data-cleaning phase. First, the titles of the TV series that ended production before 2000 were excluded. Next, non-fiction and animated productions were manually filtered out, along with TV series that were under ten minutes in episode length or lacked sufficient information. TV series (from both linear and non-linear television) were then separated from web series. Two additional searches were performed using similar parameters—one excluding the Italian language and the other excluding the Italian country tag. The results were cross-referenced to capture any items potentially missing one of the two tags and were annotated accordingly. The final number of TV series identified is 487.
While IMDb represents a comprehensive repository for Italian TV series titles, covering the vast majority of productions, it exhibits significant gaps at the episode level. To estimate the full scope of episodes, we utilised additional sources, including video-on-demand (VOD) platforms, Wikipedia, the Italian website dedicated to audiovisual products, MYMovies (https://www.mymovies.it/), and the Italian newspaper La Repubblica (https://www.repubblica.it/). The final number of the episode population is 22,226.
However, the primary constraint in corpus construction derives not from cataloguing completeness but from the limited availability of complete episodes for viewing and analysis, particularly for minor, niche or discontinued series absent from current streaming platforms. This accessibility limitation necessitated the adoption of a sampling strategy that could maximize representativeness while working within practical constraints. Out of the many existing sampling strategies, the stratified random sampling technique was selected. The method consists of dividing a heterogeneous population into homogeneous sub-groups called “strata” and randomly sampling units from each stratum in proportion to the stratum’s size of the population (Singh and Masuku 2014: 4). However, the random nature of selection within strata requires additional methodological considerations when dealing with limited material accessibility, necessitating systematic procedures for episode substitution and stratum re-weighting as detailed in subsequent sections.
To stratify the episode population, we relied on years, macro-genres and macro-distributors as stratification variables. While years can capture longitudinal variation patterns, genre is a fundamental narrative and production variable (Altman 1999; Creeber 2023). In particular, genre is paramount for both audiences and industries: if audiences use genres to organize their personal preferences, viewing practices and activities as fans, broadcasters rely on them to produce shows, to self-define their thematic networks and in television schedules, placing genres within pre-established time slots (Mittell 2001: 4). Based on these economic and audience-oriented factors and along with editorial policies, distributors establish production models that orient the features of original TV series’ production in terms of narrative and aesthetic specificities. In Italy, for instance, RAI tends to have a socially committed offering for the general public,2 Mediaset is more prone to produce popular content, while Sky aspires to a more sophisticated, rebellious and quality offering (Barra and Scaglioni 2015). In addition to these players, OTT platforms established themselves as an alternative model, intensifying some of the features of linear pay TV, starting with its “international” scope. Netflix, the most prolific platform in the realm of original production, blended continuity and renewal in Italy, rediscovering especially teen dramas with a local sensibility (Barra 2023).
By conducting inevitable generalizations to ensure methodological manageability, the strata were derived from the combinations resulting from years (2000–2023), macro-genres (Comedy, Drama and Soap Opera) and macro-distributors (RAI, Mediaset and Other). This configuration results in 216 distinct strata (24 years × 3 macro-genres × 3 macro-distributors), which represented the maximum feasible subdivision—any further splitting would have made constructing and analyzing the corpus impractical.
The size of a sample needs to ensure two key elements: (i) Representativeness; (ii) Sufficient statistical power for descriptive and inferential analyses. Three elements determine representativeness: (i) The population size (22,226 in our case); (ii) The confidence level (or confidence coefficient), that is, the probability that the confidence interval constructed from the sample will contain the actual population parameter (usually set at 95%); (iii) The margin of error represents the maximum expected difference between the sample estimate and the actual population parameter (generally set at 5%).
The minimum significant sample size reflecting these three criteria is 378 episodes.3 The statistical power of a sample, by contrast, refers to the probability of detecting a statistically significant effect if it exists in the population. A commonly acceptable threshold is 0.80, which corresponds to a probability of 80% to avoid committing a false negative (type II error), that is, failing to reject a false null hypothesis (Araujo and Frøyland 2007: 307). The statistical power is estimated using G*Power, a free and easy-to-use software.4 By selecting linear regression as the intended statistical test, the statistical power for a sample of 378 occurrences amounts to 0.9031206 (i.e., 90%, error probability set at 5%), which suggests a high likelihood of detecting a real effect.
To enable the proportional allocation of episodes across strata, we computed the percentage distribution of episodes in the population across the 216 strata identified by multiplying the values from the three stratification variables. Based on these percentages and the provisional selected sample size (378 episodes), the number of episodes to be allocated to each stratum was calculated, yielding decimal values that required rounding to integers. To compensate for the distortions introduced by this necessary rounding process, a weighting coefficient was calculated for each stratum to maintain proportional representation in subsequent analyses. Every allocation number for a stratum was rounded to the nearest integer. However, if the computed allocation was greater than zero but lower than 0.5, the value was rounded up to 1 to ensure that every represented stratum contributed at least one episode to the sample.
3 Methodological challenges and mitigation procedures: overrepresentation risk, preliminary validation and corrective strategies
Stratifying a sample based on years, genre categories, and broadcasters can provide an important first methodological framework for studying Italian TV series. However, many other variables can remain unaccounted for. For instance, selecting sample episodes from accessible repositories rather than the entire population may overrepresent certain TV series and, simultaneously, underrepresent those that are absent. Likewise, long-running series like soap operas may be underrepresented due to their extended nature, which can result in inaccuracies when reporting data at the episode level in databases.
To shed light on the distribution of the TV series titles in the sample, we visualized the observed and expected occurrences of individual products through charts and frequency tables. Furthermore, we performed the Jensen–Shannon Divergence (JSD) test, which measures the similarity between two probability distributions from 0 (identical) to 1 (completely different) and is shown to be more effective than more commonly used tests like Chi-squared, as it performs well even with low-volume distributions (Dhinakaran 2023).5 The JSD obtained by our provisional sample indicated a moderate divergence between the population and sample distributions (JSD > 0.30), suggesting the need for optimization.
To minimize the divergence between the population and the sample while preserving the proportionality established during the stratified random sampling stage, an iterative optimization algorithm was implemented. The procedure combined two main strategies:
episodes from overrepresented products were systematically substituted with episodes from underrepresented series within the same stratum using a local search algorithm that iteratively tested episode reallocations to minimize JSD;
the sample size was increased through proportional allocation of additional episodes across strata based on their remaining capacity and population weights.
The optimization algorithm simultaneously adjusted both episode selection and weighting coefficients to maintain statistical rigor while accommodating practical constraints related to time allocation of resources and data accessibility. Accordingly, the final optimized sample size amounts to 431 episodes, and, following these computational procedures, the divergence between the population and sample distributions decreased to 0.18.
Based on the new sample size, we therefore re-computed the weighting coefficients of the strata and finally collected the episodes and their metadata. In cases where some episodes were not found in repositories, two main strategies were adopted:
If the missing episode was the only one present in its stratum, it was replaced with another episode from a stratum that shares at least two out of the three variables under consideration (i.e., year, macro-genre and macro-distributor);6
If the missing episode was not the only one in its stratum and metadata about the others were available, the weighting coefficient of its stratum was re-distributed across a smaller number of episodes.7
These two strategies had to be adopted for only 5 out of 431 episodes, a small number that does not risk introducing bias into the dataset.
A bootstrap analysis showed that the average JSD across 1,000 randomly resampled samples is 0.254 (95% confidence interval: 0.240–0.272), confirming that the optimized sample is significantly more representative of the population than the average of the random samples. Finally, the coverage of the sample was assessed by examining the representation of unique stratum-series combinations. The population contains 902 distinct combinations of strata and individual series (i.e., each series counted within each year-genre-distributor stratum where it appears).8 The sample represents 244 of these combinations, yielding a coverage of 27.05%. Although coverage is not complete, the optimization process enables us to maximize representativeness within the given constraints. Further reducing the JSD would likely have required increasing the sample size, thereby compromising the sample’s compactness, which was deliberately kept small to ensure its usability in qualitative analyses.
To assess the overall quality of the stratified sample, we developed a composite sample quality index following multi-criteria evaluation approaches commonly used in sampling methodology when multiple objectives must be balanced (Groves et al. 2011, Lohr 2008, Särndal et al. 1992). The index combines two key dimensions: proportional fidelity (measured as 1 – JSD, where higher values indicate better representativeness) and sample diversity (measured as coverage). The composite index is calculated as Q = (1 – JSD) × coverage, yielding a value of 0.222. This metric confirms a good balance between proportional fidelity and diversity of representation, providing a synthetic measure of sample quality that accounts for both statistical accuracy and heterogeneity coverage.
Furthermore, a post-hoc computation of the statistical power of the sample of 431 episodes calculated through G*Power for a linear regression test shows that it amounts to 0.9331787, increasing the power relative to the minimum sample size and achieving an extremely high probability to identify a statistically significant effect. Tab. 1 provides a comprehensive overview of the key statistical parameters and validation metrics that characterize the final stratified sample, summarizing the methodological process and its outcomes.
| Statistical Measure | Value | Notes |
|---|---|---|
| Population | ||
| Total TV series | 487 | Italian TV series identified |
| Total episodes | 22,226 | From 487 TV series (2000–2023) |
| Total series instances | 902 | Series counted across all years |
| Stratification variables | 3 | Year, macro-genre, macro-distributor |
| Number of strata | 216 | 24 years × 3 genres × 3 distributors |
| Sample Design | ||
| Confidence level | 95% | Standard threshold |
| Margin of error | 5% | Standard threshold |
| Minimum sample size | 378 episodes | Based on representativeness criteria |
| Final optimized sample size | 431 episodes | After optimization procedures |
| Validation Metrics | ||
| Initial JSD | >0.30 | Moderate divergence (pre-optimization) |
| Final JSD | 0.18 | Acceptable divergence (post-optimization) |
| Bootstrap JSD mean | 0.254 | 95% CI: 0.240–0.272 (1,000 iterations) |
| Sample coverage | 27.05% | 244 out of 902 combinations represented |
| Composite quality index (Q) | 0.222 | Q = (1 – JSD) × coverage |
| Statistical Power | ||
| Minimum sample power | 0.9031206 | α = 0.05, linear regression test |
| Final sample power | 0.9331787 | α = 0.05, linear regression test |
4 Final sample validation and descriptive statistics
To further shed light on whether the sample accurately represents the population, we computed descriptive statistics of both the population and the sample across the stratification variables (namely year, genre and distributor) and visualized the results through graphs.
As shown (Figs. 1a and 1b), the sample accurately reflects the distribution of the stratification variables of the population, with minimal shifts, such as a slightly higher weight given to comedy as a macro-genre and to Mediaset as a macro-distributor, further corroborating that the sample is representative of the population.
Moving beyond the population and sample comparison, what further emerges from these distributions is a peak in overall production from 2006 to 2008, a sharp decrease in 2009 and a re-increase from 2018, corroborating the tendencies observed by the OFI.9 As for genres, soap operas are unsurprisingly the most represented macro-genre in episode volume, while comedy and drama do not present substantial differences in terms of representation. Considering distributors, RAI is the most prolific broadcaster, followed by Mediaset and the other broadcasters, even though we can hypothesize that alternative distributors to RAI and Mediaset have increased their production recently. To verify this hypothesis and to further describe the dynamics of Italian TV serial production, Figs. 2a and 2b show the time trends of the episode sample regarding macro-genres and macro-distributors, respectively.
As Fig. 2a illustrates, although still by far the most represented genre in episode volume, soap operas have undergone a drop in production: after a peak of 80% in 2003, they now compose 50% of the episode volume produced yearly, a decline also confirmed by other studies (for example, Cardini 2017). After years of similar production levels for comedies and dramas, drama TV series have outpaced comedies in recent years. While the decrease in comedy production is due to the economic crisis and the fact that many recent productions by Mediaset—the most prolific broadcaster of comedies—failed to find an audience (Barra 2020), the recent rise of dramas could be due to OTT platforms’ inclination to invest in this macro-genre, especially under the form of teen dramas (Barra 2023).
As for macro-distributor (Fig. 2b), it is worth acknowledging a huge shift in production over time: if Mediaset was the broadcaster that produced the highest quantity of episodes until 2011 (exceptions being made for 2000, 2008 and 2009), its production decreased more and more in the following years, amounting to only 6% of the overall serial production volume in 2023. In contrast, RAI has constantly increased its original scripted production, accounting for 75% of the overall number of episodes in the last years. Likewise, other broadcasters have increased their original productions, stabilizing at around 20%. These tendencies align with the results shown in the latest available report produced by the OFI.10
5 Sample applications and research implications for Italian TV studies
The validated stratified sample of episodes of Italian TV series can be applied across a range of research domains. Given the broad scope of stratification variables used in its construction, the sample is particularly well-suited for research aimed at identifying trends and tendencies across the television landscape, rather than conducting precise or granular analyses. Nevertheless, its design makes it an asset not only for quantitative investigations but also for qualitative research: while these latter analyses often focus on a limited number of audiovisual productions, the representativeness of this sample ensures that selected episodes reflect the broader characteristics and developments of the TV series population. As such, the sample lends itself to a wide array of applications, some of which are listed below. For each research application, we also report examples of studies in that field and, where applicable, we privilege research works on Italian TV series. Hence, the main applications of the sample are:
Production Trends: The sample enables systematic analysis of evolving production practices, including the changing presence of specific professions in cast and crew credits such as technical specialization and creative hierarchies. This can reveal broader industry dynamics and shifts in labor organization within Italian television production (for studies addressing similar issues, see Re and Spalletta 2023; Rocchi et al. 2023).
Representation Studies: By capturing variation across time, genre and broadcasters, the sample supports investigations into how gender, sexual orientation, ethnicity, age, disability and other sociocultural patterns are represented on screen, which could enable assessment of equity, inclusion and diversity in Italian TV series (for example, Cattani and Innocenti 2024).
Stylistic, Aesthetic and Narratological Research: The stratified nature of the sample allows for comparative studies of visual style, narrative structure, genre conventions, directorial techniques and technological innovations, aiding in the identification of trends and mutations in television storytelling (illustrated in studies such as Cardwell 2021; Innocenti and Pescatore 2011, 2014).
Economic and Marketing Analysis: Researchers can use the sample to explore how economic imperatives, marketing strategies, and media policies have influenced content production, distribution, and circulation (for example, Carelli and Garofalo 2020; Scaglioni 2020).
Audience and Reception Studies: The episodes provide a basis for selecting audiovisual products to study audience engagement, viewer interpretation, reception patterns and fan activities across different demographic and cultural contexts (as exemplified by Avezzù 2019 and Crespo-Pereira and Juanatey-Boga 2016).
Linguistic, Discourse and Semiotic Analysis: The sample supports detailed analysis of language use in television—such as dialogue and translation patterns, character speech styles and narrative discourse (for example, Bruti and Ranzato 2019)—and can be extended to multimodal studies of communication in media texts, incorporating a semiotic perspective on how meaning is constructed through the interplay of verbal, visual and auditory signs.
Other researchers can utilize the sample for various methodological purposes: (i) as a baseline for comparative studies with other European national contexts; (ii) as a training set for developing machine learning algorithms applied to automatic analysis of audiovisual content and (iii) for longitudinal studies requiring standardized and methodologically consistent datasets.
The sample subset functionality further extends its research utility. Scholars can construct focused sub-samples by limiting temporal ranges (for example, pre- and post-streaming era comparisons), specific genres (for example, the evolution of crime drama), or particular distributors (e.g., public versus commercial broadcaster strategies), while maintaining statistical rigor through appropriate power analysis and significance testing. For example, a study examining the representation of social issues in post-2015 dramas could extract relevant episodes while recalculating statistical power for the reduced sample size.
The modular nature of the stratification design and the weighting system offers further flexibility for diverse research applications. The sample includes episode-duration metadata, enabling researchers to implement alternative weighting schemes tailored to specific analytical needs. For instance, duration-based weighting could reduce the episode-level overrepresentation of soap operas, providing more balanced genre comparisons in studies focused on narrative structures or production values. Similarly, duration weighting proves essential for representation studies employing voice recognition algorithms to analyze speaking time distribution between male and female characters, where raw episode counts would distort temporal analyses.
The open nature of the dataset also promotes interdisciplinary collaborations and fosters external validation of results through replication of analyses by independent research groups, thereby strengthening the robustness of scientific conclusions regarding contemporary Italian seriality.
As seen from the examples reported for every application, most studies are rooted in a foreign national television context, and research on Italian TV series remains limited. Therefore, we hope this sample will help stimulate more systematic investigations on domestic serial production in Italy.
6 Conclusions, limitations and future developments
This study proposed a methodological framework for constructing a stratified representative corpus of contemporary Italian TV seriality to enhance qualitative and quantitative media research. By adopting the rigorous methodological approach of stratified random sampling based on year, genre, and distributor as strata, and through multiple validation stages, we developed a sample of 431 episodes that reflect the broad characteristics and trends of Italian television serial production from 2000 to 2023. This approach supports not only comprehensive analysis but also ensures rigor and replicability in research on Italian TV series. By leveraging this sample, scholars from both quantitative and qualitative approaches, as well as from diverse academic fields, can conduct studies that span production trends, genre evolution, representation issues, and aesthetic shifts. Researchers can also customize their use by relying on sub-samples or specific metadata, such as episode duration. Research on the sample, in turn, may also contribute to a more nuanced understanding of Italian television as a cultural and industrial product, enabling comparisons with other national media dynamics. These comparisons help contextualize Italian TV within global media trends and foster a deeper understanding of how national television industries shape and are shaped by cultural, economic and technological forces.
Although methodologically robust, the sample construction has limitations. First, the sample was built based on previously available metadata. Future research could therefore expand the scope of the corpus by integrating additional TV series details based on emerging data repositories or by developing new tools for automated collection. Second, the sample was stratified based on the macro-variables of genres and distributors, for practical reasons related to time and data management. To increase the granularity of the research that the sample can support, future studies could enlarge the corpus to include data on specific genres and broadcasters, which could also help more accurately reflect underrepresented areas of television production.
One of the main advantages of the proposed methodological framework lies in its intrinsic updatability, which enables dynamic adaptation of the sample in response to the evolution of Italian serial production. This flexibility manifests through two complementary dimensions. First, the integration of new productions: with the continuous production of new Italian TV series, the sample can be systematically updated by incorporating episodes from subsequent years of production, maintaining the same stratification methodology and the same proportions between strata. This process does not require redistributing existing weighting coefficients, thereby preserving the methodological integrity of the original sample. Second, retrospective integration: the sample can benefit from the eventual ex post availability of previously inaccessible episodes, either through the emergence of new digital archives or through research agreements with broadcasters and producers. Similarly, the framework readily accommodates the integration of additional weighting variables—such as production budgets, audience ratings, or critical reception scores—whenever such data becomes available. This dual updating capacity makes the sample a particularly valuable longitudinal research tool for monitoring transformations in Italian seriality over time.
Future research applications demonstrate the dataset’s adaptability to emerging methodological approaches. Longitudinal studies tracking gender representation could benefit from regular sample updates to capture ongoing changes in the industry. Machine learning approaches to automatic content analysis could leverage the stratified metadata as training labels. At the same time, comparative European studies could use the Italian sample as a methodological template for constructing equivalent national datasets. Digital humanities projects examining the evolution of visual style could combine the episode sample with automated image analysis, using the stratification weights to ensure representative coverage across different production eras and aesthetic traditions.
Data availability statement
The stratified sample of 431 episodes of Italian TV series (2000–2023) is publicly available as open data on Zenodo.11 The dataset is provided in CSV format and contains metadata for each sampled episode, including stratification variables (year, macro-genre, macro-distributor), series information, and episode-level details. Additional episode-level information (e.g. cast and crew credits) will be progressively released within the “Italian TV Series” Zenodo community. Data are accessible and reusable under Creative Commons (CC BY) licensing terms, supporting reproducible research and facilitating further investigations into Italian television seriality.
References
Altman, Rick (1999). Film/Genre. Bloomington: Indiana University Press.
Araujo, Pedro and Livar Frøyland (2007). “Statistical Power and Analytical Quantification.” Journal of Chromatography B 847(2): 305–308. https://doi.org/10.1016/j.jchromb.2006.10.002
Avezzù, Giorgio (2019). “Il successo regionale della fiction italiana. La serialità generalista 2016-2018.” Cinergie – Il Cinema e le altre Arti 16: 163–180. https://doi.org/10.6092/issn.2280-9481/8992
Barra, Luca (2020). “The Italian Sitcom Journey: The Struggles and Failures of Italian Commercial Television’s Original Productions.” Simultanea 1(1): 1–12.
Barra, Luca (2023). “Dark Narratives or Sunny Stories?” In Streaming Video: Storytelling Across Borders, edited by Amanda D. Lotz and Ramon Lobato, 248–263. New York: New York University Press.
Barra, Luca and Massimo Scaglioni (2015). “Saints, Cops and Camorristi. Editorial Policies and Production Models of Italian TV Fiction.” SERIES – International Journal of TV Serial Narratives 1(1): 65–75. https://doi.org/10.6092/issn.2421-454X/5115
Bordwell, David, Janet Steiger, and Kristin Thompson (1985). The Classical Hollywood Cinema: Film Style and Mode of Production to 1960. New York: Columbia University Press.
Bruti, Silvia and Irene Ranzato (2019). “Italian Dialetti in Audiovisual Translation: Perspectives on Three Quality TV Series.” In Ragusa e Montalbano: Voci del Territorio in Traduzione Audiovisiva, edited by Massimo Sturiale, Giuseppe Traina, and Maurizio Zignale, 341–364. Fondazione Cesare e Doris Zipelli – Euno Edizioni.
Buonanno, Milly (2012). La Fiction Italiana: Narrazioni Televisive e Identità Nazionale. Bari: Laterza.
Cardini, Daniela (2017). “A Slippery Slope: The Rise and Fall of the Domestic Soap Opera in Italian Public and Commercial Television.” VIEW Journal of European Television History and Culture 6(11): 22–32. https://doi.org/10.18146/2213-0969.2017.jethc120
Cardini, Daniela and Paola Brembilla (2025). “La serialità generalista. Evoluzione e prospettive degli studi in Italia.” IMAGO. Studi di Cinema e Media 29: 215–226. https://doi.org/10.13134/2038-5536/1-2025/15
Cardwell, Sarah (2021). “A Sense of Moment: Appreciating Television Serials from Aesthetic and Cognitive Perspectives.” In Cognition, Emotion, and Aesthetics in Contemporary Serial Television, edited by Ted Nannicelli and Héctor J. Pérez, 285–308. New York: Routledge.
Carelli, Paolo and Damiano Garofalo (2020). “Transnational Circulation of European TV Series: National Models and Industrial Strategies for Scripted Pay Imports/Exports.” In A European Television Fiction Renaissance, edited by Luca Barra and Massimo Scaglioni, 56–67. New York: Routledge.
Cattani, Lorenzo and Veronica Innocenti (2024). “Genere e generi. Predire lo speaking time dei personaggi femminili nella fiction italiana.” In Forme di produzione nelle industrie creative e culturali. confini e significati, edited by Rebecca Paraciani and Lorenzo Cattani, 185–208. Roma: WriteUp Books.
Creeber, Glen (ed.) (2023). The Television Genre Book. New York: Bloomsbury Publishing.
Crespo-Pereira, Verónica and Óscar Juanatey-Boga (2017). “Spanish TV Series on Twitter: What Social Media Audiences Say.” In Media and Metamedia Management, edited by Francisco Campos Freire et al., 435–440. New York: Springer International Publishing.
Cutting, James E. (2021). Movies on Our Minds: The Evolution of Cinematic Engagement. Oxford: Oxford University Press.
Cutting, James E., et al. (2011). “Quicker, Faster, Darker: Changes in Hollywood Film Over 75 Years.” i-Perception 2(6): 569–576. https://doi.org/10.1068/i0441aap
D’Arma, Alessandro, Tim Raats, and Jeanette Steemers (2021). “Public Service Media in the Age of SVoDs: A Comparative Study of PSM Strategic Responses in Flanders, Italy and the UK.” Media, Culture & Society 43(4): 682–700. https://doi.org/10.1177/0163443720972909
Dhinakaran, Aparna (2023). “How to Understand and Use the Jensen-Shannon Divergence.” Medium. https://medium.com/data-science/how-to-understand-and-use-jensen-shannon-divergence-b10e11b03fd6 (last accessed 16-07-25).
Groves, Robert M., et al. (2011). Survey Methodology. Hoboken, NJ: John Wiley & Sons.
Iapalucci, Greta and Guglielmo Pescatore (2025). “Italian TV Series Sample.” Zenodo. https://doi.org/10.5281/zenodo.17191258 (last accessed 10-10-25).
Innocenti, Veronica and Guglielmo Pescatore (2008). Le nuove forme della serialità televisiva: Storia, linguaggio e temi. Bologna: Archetipolibri.
Innocenti, Veronica and Guglielmo Pescatore (2011). “Architettura dell’informazione nella serialità televisiva.” IMAGO. Studi di Cinema e Media 2(3): 135–144.
Innocenti, Veronica and Guglielmo Pescatore (2014). “Changing Series: Narrative Models and the Role of the Viewer in Contemporary Television Seriality.” Between 4(8): 1–15. https://doi.org/10.13125/2039-6597/4
Kang, Hyun (2021). “Sample Size Determination and Power Analysis Using the G*Power Software.” Journal of Educational Evaluation for Health Professions 18: 1–5. https://doi.org/10.3352/jeehp.2021.18.17
Littleton, Cynthia (2015). “FX Networks Chief John Landgraf: ‘There Is Simply Too Much Television.’” Variety. https://variety.com/2015/tv/news/tca-fx-networks-john-landgraf-wall-street-1201559191/ (last accessed 16-07-25).
Lohr, Sharon L. (2008). “Coverage and Sampling.” In International Handbook of Survey Methodology, edited by Edith D. de Leeuw, Joop J. Hox, and Don A. Dillman, 97–112. New York: Taylor & Francis.
Manganello, Jennifer, Amy Franzini, and Amy Jordan (2008). “Sampling Television Programs for Content Analysis of Sex on TV: How Many Episodes Are Enough?” Journal of Sex Research 45(1): 9–16. https://doi.org/10.1080/00224490701629514
Martina, Marta and Attilio Palmieri (2015). “Researching Television Serial Narratives in Italy: An Overview.” SERIES – International Journal of TV Serial Narratives 1(1): 89–102. https://doi.org/10.6092/issn.2421-454X/5117
Mittell, Jason (2001). “A Cultural Approach to Television Genre Theory.” Cinema Journal 40(3): 3–24. https://dx.doi.org/10.1353/cj.2001.0009
Mittell, Jason (2015). Complex TV: The Poetics of Contemporary Television Storytelling. New York: New York University Press.
Morse, Janice M. (2010). “Cherry Picking: Writing from Thin Data.” Qualitative Health Research 20(1): 3. https://doi.org/10.1177/1049732309354285
Osservatorio sulla Fiction Italiana (2018). “Bilancio della stagione 2017-2018: I dati dell’offerta Rai, Mediaset, Sky, Disney, Discovery, Netflix. Volume Orario.” Ricerche APA – Presente e Futuro dell’Audiovisivo. https://ricerche.apaonline.it/ricerca/i-dati-dellofferta-rai-mediaset-sky-disney-discovery-netflix/indicatori/volume-orario//#content (last accessed: 10-10-25).
Osservatorio sulla Fiction Italiana (2022). “Serie TV: Bilancio della stagione 2021-2022.” Ricerche APA – Presente e Futuro dell’Audiovisivo. https://ricerche.apaonline.it/ricerca/serie-tv-bilancio-della-stagione-2021-2022/ (last accessed: 10-10-25).
Osservatorio sulla Fiction Italiana (2022). “Serie TV: Bilancio della stagione 2021-2022. Volume Orario.” Ricerche APA – Presente e Futuro dell’Audiovisivo. https://ricerche.apaonline.it/ricerca/serie-tv-bilancio-della-stagione-2021-2022/indicatori/volume-orario/ (last accessed: 10-10-25).
Palmer, Nick (2025). “The Rise and the Fall of Peak TV.” Essence Mediacom. https://www.essencemediacom.com/thought-leadership/new-communications-economy-entertainment-special-report/the-rise-and-fall-of-peak-tv (last accessed 16-07-25).
Re, Valentina and Marica Spalletta (2023). “Unsuitable Jobs for Women: Women’s Behind-the-Scenes Employment and Female On-Screen Representation in Italian TV Crime Drama.” Comunicazioni Sociali 45(1): 82–97. https://doi.org/10.26350/001200_000177
Rocchi, Marta and Guglielmo Pescatore (2022). “Modeling Narrative Features in TV Series: Coding and Clustering Analysis.” Humanities and Social Sciences Communications 9(1): 1–11. https://doi.org/10.1057/s41599-022-01352-9
Rocchi, Marta, Lorenzo Cattani, and Guglielmo Pescatore (2023). “Gender Equality in European Netflix TV Series Production (2014–2019).” Mediascapes Journal 21: 188–204. https://rosa.uniroma1.it/rosa03/mediascapes/article/view/18426.
Salt, Barry (1974). “Statistical Style Analysis of Motion Pictures.” Film Quarterly 28(1): 13–22. https://doi.org/10.2307/1211438
Salt, Barry (2006). Moving into Pictures: More on Film History, Style, and Analysis. London: Starword.
Salt, Barry (2009). Film Style and Technology: History and Analysis (3rd ed.). London: Starword.
Särndal, Carl-Erik, Bengt Swensson, and Jan Wretman (1992). Model Assisted Survey Sampling. New York: Springer.
Scaglioni, Massimo (2020). “Made in Italy: The International Circulation of Italian Film and Series and the Role of Pay TV.” Cinergie – Il Cinema e le altre Arti 9(18): 17–24. https://doi.org/10.6092/issn.2280-9481/11169
Singh, Ajay S. and Micah B. Masuku (2014). “Sampling Techniques & Determination of Sample Size in Applied Statistics Research: An Overview.” International Journal of Economics, Commerce and Management 2(11): 1–22.
Valentini, Paola (2012). “Whodonit? RAI TV Fiction Production Between Detection and Giallo.” Cinéma & Cie 12(2): 25–38. https://riviste.unimi.it/index.php/cinemaetcie/article/view/16314.
The latest published report can be found at the following link: https://ricerche.apaonline.it/ricerca/serie-tv-bilancio-della-stagione-2021-2022/ (last accessed 10-10-25).↩︎
To know more about the TV serial production modes of the public service network, see Valentini (2012).↩︎
See Singh and Masuku (2014) for an overview of how sample size is calculated.↩︎
To learn more about how G*Power can be used to determine a sample size and its statistical power, see Kang (2021).↩︎
The Chi-squared test was not suitable for our analysis due to the presence of numerous strata with very low expected frequencies or zero values, which violate the test’s assumptions regarding minimum expected cell counts (typically ≥5). The JSD test proves more robust in handling such sparse distributions common in stratified sampling with multiple variables.↩︎
This occurred for only two episodes, both from Bradipo (2001–2002). Specifically, the series appeared with one episode in the stratum 2001_Comedy_Other and one episode in 2002_Comedy_Other. Accordingly, the first was replaced with an episode from a RAI comedy in 2001, and the second with an episode from a Mediaset comedy in 2002.↩︎
This occurred only for Un posto al sole (1996–) in three strata, from 2009, 2010 and 2011. For instance, in the stratum 2009_RAI_Soap Opera, the series had a weighting coefficient of 6.386, to be distributed across four episodes (single-episode weight 1.596). Since one episode was unavailable, its coefficient was transferred to another, doubling that episode’s weight and re-distributing the stratum coefficient over three episodes.↩︎
For example, Un posto al sole appears in each of the twenty-four years considered (2000–2023) and therefore accounts for 24 out of the 902 combinations.↩︎
See https://ricerche.apaonline.it/ricerca/i-dati-dellofferta-rai-mediaset-sky-disney-discovery-netflix/indicatori/volume-orario//#content (last accessed: 10-10-25) and https://ricerche.apaonline.it/ricerca/serie-tv-bilancio-della-stagione-2021-2022/indicatori/volume-orario//#content (last accessed: 10-10-25).↩︎
Available at: https://ricerche.apaonline.it/ricerca/serie-tv-bilancio-della-stagione-2021-2022/indicatori/volume-orario/#content (last accessed: 10-10-25).↩︎
The sample can be accessed through the following DOI: https://doi.org/10.5281/zenodo.17191258 (last accessed 10-10-25).↩︎