Polisci Bureaucracy Parties Voter Behaviour And Polarization

---
name: polisci-bureaucracy-parties-voter-behaviour-and-polarization
description: "Use for how politics actually operates day to day: bureaucracy and public administration including principal-agent problems, parties and party systems, voter behaviour and how weakly the informed-rational-voter model holds, public opinion and the measurement problems in polling, and polarization including what is actually measured and what is asserted."
---

# Political Science and Governance: Bureaucracy and Public Administration, Parties and Party Systems, Voter Behaviour, Public Opinion and Measurement, and Polarization

> **Part 4 of 6** of the *Political Science, Human Systems and Governance* reference (plugin `political-science-human-systems-governance`), covering §14–§18. Sibling skills: `polisci-what-the-field-establishes-power-the-state-and-theory` (§0–§4), `polisci-collective-action-commons-social-choice-and-voting` (§5–§8), `polisci-constitutional-design-electoral-systems-and-institutions` (§9–§13), `polisci-regimes-state-capacity-corruption-ir-and-field-problems` (§19–§25), `polisci-reference` (§26–§31). Section numbers are shared across the set; a reference written as §N → `skill` points into that sibling skill.
>
> **Currency:** The theory is stable. Two areas are live. See §26 → `polisci-reference` for current democracy measurement, and the scholarly dispute about whether those measurements are reliable.

> **⚠️ A note on how this document handles contested questions.** ⚠️ **Political science
> contains robust empirical findings, contested empirical findings, and normative
> disputes that evidence cannot settle. This document tries to say which is which** (§1 → `polisci-what-the-field-establishes-power-the-state-and-theory`),
> **and where a question is genuinely contested it presents the strongest version of more
> than one position rather than picking.** ⚠️ **Where I report a finding, I attribute it;
> where scholars disagree, I say so.**
>
> **Complements a civil/industrial engineering reference (institutions as accumulated
> failure logs), a pedagogy reference (evidence quality in social science), and an
> economics-adjacent treatment of collective action.**
>
> **⚠️ GOTCHA** boxes mark the results that are counterintuitive or routinely misstated.
>
> **The three ideas that organize this document:**
> 1. **⚠️ COLLECTIVE ACTION PROBLEMS are the discipline's deepest structure** (§5–§6 → `polisci-collective-action-commons-social-choice-and-voting`).
>    **Individually rational behaviour producing collectively bad outcomes explains
>    government, taxation, regulation, unions, cartels and treaties — and Ostrom's work
>    showed the standard remedies are not the only ones.**
> 2. **⚠️ INSTITUTIONS SHAPE BEHAVIOUR, and the effects are often perverse** (§9–§15 → `polisci-constitutional-design-electoral-systems-and-institutions`).
>    **Electoral rules, veto points and federal structures produce predictable
>    consequences that designers frequently did not intend.**
> 3. **⚠️ MEASUREMENT IS THE FIELD'S BINDING CONSTRAINT** (§25 → `polisci-regimes-state-capacity-corruption-ir-and-field-problems`, §26 → `polisci-reference`). **Most interesting
>    political concepts — democracy, power, legitimacy, polarization — are hard to measure,
>    and disputes that look substantive are often measurement disputes underneath.**

---

## §14. Bureaucracy and Public Administration

**⚠️ Weber's ideal type**: **hierarchy, written rules, specialization, meritocratic
appointment, impersonality.** ⚠️ **Designed as a REMEDY for patronage and arbitrariness —
its virtues and its pathologies come from the same source.**
**⚠️ The principal-agent problem** (§5 → `polisci-collective-action-commons-social-choice-and-voting`) ⚠️ **is unavoidable, because bureaucrats know more
than their political principals about their own domain — that expertise is the reason for
delegating and the reason control is hard.**
**⚠️ Control mechanisms**: ⚠️ **"police patrol" (continuous oversight) versus "fire alarm"
(react when interested parties complain — cheaper and biased toward organized
complainants, §5 → `polisci-collective-action-commons-social-choice-and-voting`'s Olson logic).**
**⚠️ Street-level bureaucracy (Lipsky)** is one of the most useful concepts here:
⚠️ **frontline staff exercising discretion under impossible caseloads effectively MAKE
policy, whatever the statute says.** **⚠️ Implementation is not a transmission belt.**
**⚠️ NEW PUBLIC MANAGEMENT** — ⚠️ **markets, targets, contracting, agencification — was the
dominant reform paradigm from the 1980s; the assessment is mixed, with documented gains in
some services and documented gaming of targets, fragmentation and hollowed-out capacity in
others.** ⚠️ **Later "public value" and "new public governance" frames were partly a
reaction.**
**⚠️ Goodhart's Law is the recurring failure**: ⚠️ **a measure that becomes a target ceases
to be a good measure.**

---

## §15. Parties and Party Systems

**⚠️ Party functions**: **aggregating interests, recruiting and screening candidates,
structuring choice, organizing government, and providing the brand that makes voting
cognitively cheap** (§16).
**⚠️ System types**: ⚠️ **dominant, two-party, moderate and polarized pluralism, and the
cleavage structures (Lipset & Rokkan) that produced them historically.**
**⚠️ Organizational evolution**: **cadre → mass → catch-all → ⚠️ CARTEL party (Katz &
Mair — parties as state-subsidized entities increasingly detached from members and
collusive with each other).**
**⚠️ Documented trends across established democracies**: ⚠️ **declining membership,
weakening party identification in some countries alongside strengthening PARTISAN ANIMOSITY
in others (§18), rising electoral volatility, and the growth of parties positioning
themselves against mainstream competitors.**
**⚠️ Populism** is a contested concept: ⚠️ **the most-used definition (Mudde) treats it as a
"thin" ideology dividing society into a pure people versus a corrupt elite, attachable to
either left or right.** **⚠️ Critics argue the term is applied selectively and does
normative work disguised as description — a criticism worth taking seriously when reading
any literature that uses it.**

---

# PART IV — BEHAVIOUR

## §16. ⚠️ Voter Behaviour

```
⚠️ WHAT IS FAIRLY ROBUST
   ⚠️ PARTY IDENTIFICATION is a strong, stable predictor — and
      the Michigan school treats it as closer to a SOCIAL IDENTITY
      than a running tally of policy agreement
   ⚠️ INCUMBENCY ADVANTAGE in many systems
   ⚠️ ECONOMIC VOTING exists but is ⚠️ heavily filtered by partisan
      perception — partisans disagree about economic FACTS
   ⚠️ TURNOUT correlates with age, education and income; ⚠️ and
      compulsory voting raises it enormously where used
   ⚠️ Most voters have limited political knowledge (Converse's
      finding on unconstrained belief systems), and ⚠️ use
      HEURISTICS — party label, endorsements, retrospection
⚠️ WHAT IS CONTESTED
   ⚠️ The size of persuasion and campaign effects (⚠️ field
      experiments generally find them SMALL, which surprised
      practitioners) · media effects on vote choice · the
      relative weight of economic versus cultural grievance
⚠️ THE PARADOX OF VOTING  ⚠️ a single vote almost never decides an
   election, so instrumental rationality struggles to explain
   turnout at all. ⚠️ The leading answers are duty, expressive
   benefit and identity — which is itself telling about what
   voting IS
```
**⚠️ Retrospective voting (Fiorina)** is a defensible low-information strategy: ⚠️ **judge
outcomes rather than policies.** ⚠️ **Its known failure mode is that voters punish
incumbents for things outside their control — drought, shark attacks, the world economy —
which weakens the accountability story it was meant to support.**

---

## §17. Public Opinion and Measurement

**⚠️ Polling relies on sampling theory that is sound and on response behaviour that
increasingly is not.** ⚠️ **Falling response rates mean modern polls depend heavily on
WEIGHTING and modelling assumptions, so two pollsters with the same raw data can publish
different numbers defensibly.**
**⚠️ Known problems**: ⚠️ **nonresponse bias (⚠️ and differential partisan nonresponse is a
live suspect in recent polling misses); question wording and order effects; social
desirability; and ⚠️ NON-ATTITUDES — respondents readily offer opinions on things they have
never considered, including fictitious legislation.**
**⚠️ The margin of error is a SAMPLING statistic only** — ⚠️ **it does not cover coverage
error, nonresponse, weighting choices or house effects, so the real uncertainty is wider
than the stated figure.**
**⚠️ Aggregation and forecasting** improve on single polls; ⚠️ **and probabilistic forecasts
are widely misread — a 30% outcome occurring is not a failed forecast.**
**⚠️ Deliberative polling and citizens' assemblies** measure something different: ⚠️ **what
opinion looks like AFTER information and deliberation, which often differs substantially
from raw polling.**

---

## §18. Polarization

**⚠️ The crucial distinction, because the two are frequently conflated:**
```
⚠️ IDEOLOGICAL POLARIZATION  ⚠️ how far apart people's POLICY
   positions are. ⚠️ Evidence for mass ideological divergence is
   comparatively modest in several democracies
⚠️ AFFECTIVE POLARIZATION  ⚠️ how much partisans DISLIKE and
   distrust the other side. ⚠️ This has risen markedly in a number
   of countries, and it is the better-evidenced trend
⚠️ ELITE POLARIZATION  ⚠️ legislators and parties have diverged
   more clearly than electorates in several cases — which raises
   the question of direction of causation
⚠️ SORTING vs POLARIZING  ⚠️ people may not have changed their
   views so much as ALIGNED their party with their other identities,
   producing sharper conflict without more extreme opinions
```
**⚠️ Proposed drivers, none established as dominant**: ⚠️ **media fragmentation and
selective exposure; social media (⚠️ where the evidence is genuinely mixed — several
large-scale studies find smaller effects than the public debate assumes); geographic and
social sorting; elite behaviour; and identity consolidation.**
**⚠️ The consequences that ARE reasonably documented**: ⚠️ **reduced willingness to compromise,
declining trust in institutions when they rule against one's side, and reduced social
interaction across party lines.**