# What Shapes Children: A Defensible Influence Model, 1900–2020

A composite-index estimate of the relative share of formative influence on U.S. children (ages ~5–17) across seven channels — Parents, Church/Religion, Teachers/School, Peers, TV/Broadcast, Internet (Non-Algorithmic), and Algorithm/Influencers — by 5-year period from 1900 to 2020.

This is not a measurement. It is a transparent, reproducible model built from documented exposure data and explicitly-stated weighting parameters. The goal is to allow defensible discussion of how the channels shaping children have shifted over the past century, with full visibility into where the numbers come from and where they don't.

## Files in this project

| File | Purpose |
|---|---|
| **Interactive_Influence_Model.html** | Primary deliverable. Interactive chart with sliders for W (weight) values, peer/algorithm split control, presets, and per-year breakdown. Open in any browser. |
| **What_Shapes_Children_Chart.html** | Static chart with toggle between the original supplied values and the calibrated reconstruction. |
| **Methodology_Childhood_Influence_1900_2025.md** | Conceptual framework for the model: what's being measured, how cells are constructed, anchor years, falsifiability checks. |
| **Empirical_Sources_E_and_W.md** | Source-by-source backing for T (exposure) values and the research traditions backing W (weight) values. |
| **Empirical_Anchors_RAW.csv** | Every direct empirical measurement with year, value, unit, and citation. NO DATA cells explicitly marked. |
| **Reconstruction_Chart_vs_Empirical.md** | Coverage map showing which chart cells are empirically supported and which require modeling. |
| **Calibrated_Reconstruction.md** | The three adjustments (W tuning, peer/algorithm split, religious-instruction reallocation) that bring the model close to the original chart. |
| **Calibrated_Dataset.csv** | Side-by-side: original chart values vs. calibrated reconstruction, full 27-year series. |
| **Defensible_TandW_Assembly.md** | Tight assembly document: T and W values with citations, three scenarios for the 2025 Algorithm cell. |

## Methodology in brief

### The composite index

For each 5-year period and each channel, the formative-influence share is calculated as:

```
share = (T × W) / Σ(T_i × W_i)
```

Where T is hours per week of contact with the channel for a typical U.S. child 5–17, and W is the per-hour relational-primacy weight. Shares are normalized to sum to 100% per period.

The 100% sum is a heuristic that lets us discuss *relative* shifts over time; it is not a claim that the channels are perfectly substitutable.

### T (time exposure) values

T values are baked into the model and not user-editable. They are drawn from the empirical sources listed in `Empirical_Anchors_RAW.csv`. Post-1965 cells are anchored in direct measurements (Bianchi, Hofferth, Robinson, ATUS, Nielsen, Kaiser, Common Sense Media, Pew, Gallup). Pre-1965 cells are reconstructed from adjacent proxies — school enrollment (NCES), broadcast penetration (FCC), female labor-force participation (BLS), religious affiliation (Finke & Stark; Gallup post-1939) — with wider implicit error bars.

The current T values reflect three calibration moves:
- **Church T** is broadened from service-attendance hours to include the full religious framework: services + Sunday school + family devotions + parochial schooling + religious cultural context.
- **School T** counts effective formative time for enrolled students, not population-averaged hours.
- **Peers T (pre-1960)** reflects multi-age unsupervised play, sibling-dense households, walking-to-school, and one-room schoolhouse peer time — substantially higher than today's supervised, structured peer time.

### W (relational primacy) values

W values are user-adjustable in the interactive model. They represent the per-hour formative weight of each channel relative to passive broadcast (≈1.0 baseline). Default ("Literature default" preset):

| Channel | W | Anchoring research |
|---|---|---|
| Parents | 3.0 | Bowlby, *Attachment and Loss* (1969); Ainsworth et al., *Patterns of Attachment* (1978); Bronfenbrenner, *Ecology of Human Development* (1979); Groh et al. (2017) attachment meta-analysis |
| Church/Religion | 5.3 | Smith & Denton, *Soul Searching* (2005); King & Roeser (2009); Putnam, *Bowling Alone* (2000) religious social capital; Cnaan religious-socialization series |
| Teachers/School | 1.3 | Hattie, *Visible Learning* (2009) meta-analysis; Csikszentmihalyi & Larson, *Being Adolescent* (1984) on classroom disengagement |
| Peers | 2.4 | Steinberg & Monahan (2007); Brown (2004), *Handbook of Adolescent Psychology* |
| TV/Broadcast | 1.3 | Gerbner cultivation theory; Anderson et al. (2001) Recontact longitudinal study |
| Internet (Non-Algo) | 1.0 | Subset of broadcast; interactive but not engagement-optimized |
| Algorithm/Influencers | 2.8 | Haidt, *The Anxious Generation* (2024); Allcott et al. (2020) *AER*; Twenge et al. (2018); Bhargava & Velasquez (2021) attention-capture economics |

### Two modeling decisions worth flagging

**Peer/Algorithm split rule.** When a teen watches a TikTok her friend sent, that exposure is both algorithmic and peer-mediated. The interactive model lets you control this split (0–100%). At 0%, all algorithmic time stays in Algorithm. At 60%, most algorithmic exposure is reallocated to Peers (the chart's high-Peers narrative for the modern era depends on this).

**Religious-instruction reallocation.** Some fraction of parent time in religious families is religious teaching (grace, devotions, Bible reading, moral framing, parochial-school choice). The model treats this as Church formation rather than Parent formation — the tradition is the agent, the family is the bearer. This is a substantive theological/cultural claim the model makes, not a measurement. It's the reason early-20th-century Church can read at 30%+ rather than the 5–7% pure service-attendance hours would produce.

## Using the interactive model

Open `Interactive_Influence_Model.html` in any browser.

**Presets** load different W configurations:
- *Literature default* — values defended by the underlying research traditions (Algorithm 2020 ≈ 38%).
- *Original chart match* — high Church W, low School W, used to approximately reproduce the original supplied chart.
- *Media skeptic* — low W on all mediated channels, high peer split (Algorithm 2020 ≈ 15%).

**W sliders** (0–6) let you adjust each channel's per-hour weight live.

**Peer/Algorithm split slider** (0–100%) controls reallocation of mediated-peer time.

**Year detail panel** shows T, W, and resulting share for every channel at any selected year. Default 2020.

## Empirical sources by channel

### Parents
- Bianchi, S. M., Robinson, J. P., & Milkie, M. A. (2006). *Changing Rhythms of American Family Life*. Russell Sage Foundation.
- Hofferth, S. L., & Sandberg, J. F. (2001). Changes in American children's time, 1981–1997. *Journal of Marriage and Family*, 63.
- Hofferth, S. L. (2009). Changes in American children's time — 1997 to 2003. *Electronic International Journal of Time Use Research*. ([PMC](https://pmc.ncbi.nlm.nih.gov/articles/PMC2939468/))
- Robinson, J. P., & Godbey, G. (1997). *Time for Life: The Surprising Ways Americans Use Their Time*. Penn State Press.
- [BLS American Time Use Survey, 2003–present](https://www.bls.gov/news.release/atus.htm)

### Church/Religion
- [Gallup historical church attendance series, 1939–present](https://news.gallup.com/poll/642548/church-attendance-declined-religious-groups.aspx)
- Pew Research Center, *Religious Landscape Studies* (2007, 2014, 2024).
- Finke, R., & Stark, R. (2005). *The Churching of America, 1776–2005*. Rutgers University Press.
- Smith, C., & Denton, M. (2005). *Soul Searching: The Religious and Spiritual Lives of American Teenagers*. Oxford University Press.
- Cnaan, R., et al. (2002). *The Invisible Caring Hand: American Congregations and the Provision of Welfare*. NYU Press.

### Teachers/School
- [NCES Common Core of Data and historical statistics](https://nces.ed.gov/ccd/)
- Goldin, C., & Katz, L. (2008). *The Race Between Education and Technology*. Harvard.
- Hattie, J. (2009). *Visible Learning: A Synthesis of Over 800 Meta-Analyses Relating to Achievement*. Routledge.

### Peers
- Larson, R. W., & Verma, S. (1999). How children and adolescents spend time across the world. *Psychological Bulletin* 125(6). ([PubMed](https://pubmed.ncbi.nlm.nih.gov/10589300/))
- Csikszentmihalyi, M., & Larson, R. (1984). *Being Adolescent: Conflict and Growth in the Teenage Years*. Basic Books.
- Steinberg, L., & Monahan, K. C. (2007). Age differences in resistance to peer influence. *Developmental Psychology* 43(6).
- Brown, B. B. (2004). Adolescents' relationships with peers. In *Handbook of Adolescent Psychology* (2nd ed.).

### TV/Broadcast (includes radio pre-1948)
- [Television Bureau of Advertising — National TV Household Penetration Trends](https://www.tvb.org/wp-content/uploads/2022/10/National-TV-Household-Penetration-Trends.pdf)
- [ICPSR — Introduction of Television to the United States Media Market, 1946–1960](https://www.icpsr.umich.edu/web/ICPSR/studies/22720)
- Nielsen children's television viewing reports (1979+).
- Anderson, D. R., et al. (2001). Early childhood television viewing and adolescent behavior: The recontact study. *Monographs of the Society for Research in Child Development*, 66(1).
- [Measuring Children's Media Use in the Digital Age (PMC)](https://pmc.ncbi.nlm.nih.gov/articles/PMC2745155/)

### Internet (Non-Algorithmic)
- [Pew Research Center — Internet/Broadband Fact Sheet](https://www.pewresearch.org/internet/fact-sheet/internet-broadband/)
- Rideout, V. J., Foehr, U. G., & Roberts, D. F. (2010). *Generation M2: Media in the Lives of 8- to 18-Year-Olds*. Kaiser Family Foundation. ([PDF](https://files.eric.ed.gov/fulltext/ED527859.pdf))

### Algorithm/Influencers
- [Common Sense Media Census 2015, 2019, 2021](https://www.commonsensemedia.org/research/the-common-sense-census-media-use-by-tweens-and-teens-2021)
- [Pew Research — Teens, Social Media and Technology 2024](https://www.pewresearch.org/internet/2024/12/12/teens-social-media-and-technology-2024/)
- Haidt, J. (2024). *The Anxious Generation*. Penguin.
- Allcott, H., Braghieri, L., Eichmeyer, S., & Gentzkow, M. (2020). The welfare effects of social media. *American Economic Review* 110(3).
- Twenge, J. M., et al. (2018). Decreases in psychological well-being among American adolescents after 2012 and links to screen time. *Emotion* 18(6).
- Bhargava, V. R., & Velasquez, M. (2021). Ethics of the attention economy: The problem of social media addiction. *Business Ethics Quarterly*.

## Research traditions backing W (no single-number citations)

- **Attachment theory**: Bowlby (1969) *Attachment and Loss*; Ainsworth et al. (1978) *Patterns of Attachment*; Groh et al. (2017) meta-analysis on attachment-outcome links.
- **Ecological systems**: Bronfenbrenner (1979) *The Ecology of Human Development* — proximal vs. distal processes.
- **Social learning**: Bandura (1977) *Social Learning Theory*; the Bobo doll studies (1961, 1963).
- **Source credibility**: Hovland & Weiss (1951); Petty & Cacioppo (1986) Elaboration Likelihood Model.
- **Personal influence**: Katz & Lazarsfeld (1955) *Personal Influence* — two-step flow of communication.
- **Religious socialization**: National Study of Youth and Religion longitudinal series (Smith et al.); Putnam social capital research.
- **Algorithmic engagement**: Haidt (2024); Allcott et al. (2020); Twenge longitudinal screen-time series; recommender-system attention-capture literature.

## Honest limitations

### Five things the model cannot do

1. **It cannot measure formation outcomes.** Exposure × weight ≠ what children become. No longitudinal cohort raised on engagement-optimized feeds has yet reached adulthood. The link from algorithmic exposure to durable identity formation is hypothesized, not measured.

2. **It cannot eliminate self-report bias.** ATUS over-counts engaged parent time; CSM teen self-reports underestimate phone use; Gallup church attendance is over-reported by 15–25 points relative to actual attendance counts. All T values inherit these biases.

3. **It cannot produce quantitative W values from the literature.** No published study assigns numeric formative weight per hour to any channel. The W slider values are calibrated judgments anchored to research traditions that establish *direction* and *relative magnitude*, not numbers.

4. **It cannot model concurrent exposure.** A child watching TV with parents present is double-counted. Phone multitasking during family meals is treated as additive when it partly displaces family time.

5. **It cannot prove causation.** Even where exposure and outcomes correlate (e.g., teen social-media use and depression rates), the model does not establish that one causes the other.

### Cells with low empirical anchoring

| Period | Coverage |
|---|---|
| 1900–1935 | No direct child time-use data. Reconstructed from school enrollment, household composition, broadcast penetration, religious affiliation. Treat with ±10 point error bars. |
| 1940–1960 | Sparse anchors (Gallup begins 1939, TV penetration 1948, Robinson time-use 1965). Treat with ±7 point error bars. |
| 1965+ | Direct time-use measurements available for most cells. Treat with ±3 point error bars. |
| 2010–2020 | Multiple convergent measurements (CSM, Pew, ATUS, KFF). Highest confidence. |

## What's not in the model

### Channels we don't separately track

- **Siblings.** Distinct relational category, especially in pre-1960 large households where older siblings co-raised younger ones. Currently bundled into Peers.
- **Extended family.** Grandparents, aunts/uncles, multigenerational households. Significant pre-1960; effectively absent in modern model.
- **Books/Literature.** Substantial formative time pre-TV; still shapes vocabulary and worldview today. Not captured.
- **Music.** Has migrated from familial → broadcast → algorithmic without being treated as a distinct channel; currently absorbed into TV (radio era) or Algorithm (streaming era).
- **Workplace/chores.** Pre-child-labor-law era, working alongside adults shaped many children. After-school jobs still matter.
- **Self/solitude/play.** Solo reading, daydreaming, unstructured solo play. Developmentally important; not captured.

### Stratifications we don't break out

- **Age.** A 5-year-old's allocation is very different from a 15-year-old's. The model is for the broad 5–17 band.
- **Race/class/region.** Black church attendance significantly higher; immigrant family structures different; rural/urban gap material.
- **Religious tradition.** Mormon, Orthodox Jewish, evangelical, mainline Protestant, Catholic, Muslim, secular — radically different exposure profiles.
- **Geography.** U.S. only. Cross-national application requires re-estimating every series.

## How to defend the model in conversation

**Strongest defenses:**
- "It is a composite index built from documented proxies. The construction rule is published. You can disagree with the W settings and see the result live."
- "Post-1965 T values are anchored in direct time-use data — ATUS, Nielsen, Pew, Common Sense Media, Kaiser, Gallup. The shape of the curves matches independent series."
- "We aren't claiming algorithmic feeds *cause* a 40% share of formative influence. We are claiming that, under a documented allocation rule, they account for ~40% of the influence budget — and we've shown what assumptions would have to change to get to 20% or 50%."
- "The model embeds an explicit theological claim: that religious tradition is the formative agent and the family is the bearer. We surface that claim rather than hiding it."

**Things not to claim:**
- That this is survey data or measurement.
- That single-percentage-point precision is meaningful.
- That the U.S. estimates extend to other countries.
- That pre-1950 cells are anything other than reconstructions.
- That exposure × weight equals formation outcomes.

---

*Last updated: May 2026*