Gary Goertz — Social Science Concepts and Measurement

336 published claims from this book.

• A complete definition supplies all necessary conditions that are jointly sufficient
• A concept can be realized by multiple substitutable means across cases.
• A concept can mix necessity and substitutability across its dimensions.
• A concept’s dimensions are ontologically its defining characteristics, not empirically found indicators
• A concept’s name can mislead about its actual theoretical balance between poles.
• A concept’s negation is logically distinct from its paired opposite.
• A conceptual framework that cannot see a hybrid makes it vanish from research.
• A country can dodge military modernization for any one of four independent reasons
• A focused positive concept makes its negative pole wildly heterogeneous, spoiling comparisons.
• A gender-inclusive welfare state requires childcare, maternity leave, and independent pensions.
• A good social-science definition states exactly which attributes are necessary and sufficient.
• A gray zone is a concept in its own right, not a diminished subtype.
• A household counts as deprived if it uses any one listed fuel.
• A latent variable causes its indicators, so indicators are not the construct itself.
• A linear middle of a wealth range differs from the categorized middle.
• A measure adequate for description becomes incomplete once used in causal hypotheses.
• A sample mean is not the same as a concept’s conceptual midpoint.
• A single failed necessary condition nullifies the entire aggregate score.
• A single failing criterion can veto an otherwise near-perfect whole.
• A single index can combine best-shot and summation aggregation within one structure.
• Abstract variables stay untethered because only variation, not empirical meaning, is required.
• Accurate welfare measurement requires counting all goods and services people consume
• Adding nonmonetary factors like rights breaks sum-based measurement approaches.
• Adding ordinal dummies into an interval index is more arbitrary than fuzzy-logic transformation.
• Adjectives attached to concepts add necessary dimensions conjoined by AND.
• Agenda success requires three independently necessary streams to converge at once.
• Aggregated multi-dimensional indices are causally mismeasured when they include causally irrelevant components.
• Aggregating heterogeneous indicators into one scale is a critical, under­examined step
• Aggregation rules can differ by concept level, so ignoring levels distorts measures
• Aggregation rules determine how severely conceptual tautology inflates correlations.
• Aggregation rules range from zero substitutability to full substitutability across dimensions.
• Aggregation rules valid at the data level are invalid at the semantic level.
• Agreement-based minimal definitions yield necessary but not sufficient conditions.
• An odd number of ordinal levels gives exactly one clear crossover point
• Any transformation of data is a claim about what the concept means
• Aristotelian true-or-false logic still shapes concept dichotomization in social science.
• Averaging components can make the whole weaker than its strongest part.
• Averaging dimensions severs the deterministic link between concept intension and empirical extension.
• Averaging every level of a concept means its structure was never thought through.
• Averaging independent measurements of one true value is optimal under classical statistics
• Averaging multidimensional concept scores contradicts the necessary-and-sufficient structure scholars define
• Averaging necessary dimensions signals conceptual confusion, not measurement success.
• Best-shot aggregation makes large values dominate small ones by design
• Big social concepts like well-being nest four or more hierarchical levels
• Binarizing a continuous measure is a distinct transformation, not mere information loss.
• Blacklist indices weaponize apparently technical thresholds for normative shaming
• Bureaucratic category cutpoints become the thresholds researchers use for fuzzy-set membership.
• Calling a feature essential implies conjunctive aggregation, not averaging
• Capability theorists treat well-being as constitutive of human being, not as indicators
• Categories invented to fill typological gaps vanish from later empirical use
• Causal inquiry determines which concept and measure are appropriate.
• Causal relations occur both between concepts and within a single concept.
• Classic welfare-state programs implicitly targeted a married male breadwinner.
• Clinical diagnostic categories routinely use required symptoms plus an m-of-n sufficiency rule.
• Coding amnesty as transitional justice smuggles a contested moral judgment into quantitative data.
• Coding manuals are the best source of a variable’s full conceptualization
• Coding typology rows and columns as 0 and 1 smuggles in an ordinal claim.
• Collapsed, occupied, and transitional states resist clean classification as regime types.
• Combining criteria with AND shrinks a category while OR expands it
• Completeness is a demanding standard that researchers are tempted to abandon
• Computing ratios automatically converts power variables into ratio-scale measures.
• Concept construction works like genetic engineering, combining parts with no guaranteed outcome
• Concept definitions relax required attributes as the number of attributes grows.
• Concept formation should be driven by semantics and theory, not by empirical data.
• Concept hybrids are named asymmetrically, defaulting to one pole of the pair.
• Concept membership should depend on a case’s own features, not the sample.
• Concept opposites are generally asymmetric rather than strict logical negations.
• Concept pairs like poverty–wealth are semantically asymmetric, not mirror opposites.
• Concept structure alone determines how many cases count as members
• Concept structures can differ across levels, but some combinations are rare.
• Concept structures can mix necessary conditions with family resemblance across levels.
• Concept structures operate on different logics at the secondary and indicator levels.
• Concept structures range from weakest link to best shot by substitutability
• Concept-as-ideal is a minor strand in the psychology of concepts.
• Concepts are built top-down while measurement aggregates bottom-up
• Concepts are equally core to causal theory, normative philosophy, and empirical description.
• Concepts are theories about what constitutes a phenomenon
• Concepts can be evaluated by how accurately they describe real-world phenomena.
• Concepts connect ontology to measurement through semantic mapping, not causal arrows
• Concepts defined relative to a historical period cannot track quality across periods
• Concepts have contestable cores and uncontested usual-suspect features.
• Concepts have distinct semantic and measurement levels that operate independently.
• Concepts occupy a cognitively preferred level of abstraction used in hypotheses
• Concepts should be conceptualized at the extreme pole of their continuum
• Concepts that resist adjectival form hide which end of the scale causally matters.
• Concepts that stop short produce empirical score distributions spiked at the top
• Conceptual change happens by adding previously excluded features to a definition.
• Conceptual gray zones are distinctive phenomena with their own hyphenated terminology.
• Conceptual hybrids get collapsed because analysis cannot handle many categories.
• Conceptual meaning, not measurement convention, should drive quantification
• Conceptual progress requires specifying underlying mental mechanisms, not just clearer definitions
• Conceptual stretching means dropping necessary defining attributes.
• Conceptualization and measurement cannot be separated in social science research
• Conceptualizing heterogeneous categories negatively is sometimes the better choice.
• Conflating a concept’s constituent dimensions with its data indicators causes measurement failure.
• Connecting all concept dimensions with AND shrinks extension toward zero
• Connecting concept dimensions with OR makes added intension increase extension.
• Constitutive features of causal variables must play a role in the mechanism
• Contested concepts get defined mainly by what they negate, not what they affirm
• Continuous concept measures raise the distinct questions of dimension weighting and substitutability
• Core social science concepts are inherently multidimensional and multilevel
• Corporatism is state-licensed monopoly representation of compulsory, noncompetitive interest categories.
• Correlations between measures overstate similarity because extremes dominate.
• Counting rules generate borderline cases just below their sufficiency threshold.
• Cumulative distributions calibrate fuzzy-set codings but must not be applied automatically
• Dashed indicator arrows deliberately signal a noncausal construction rather than cause
• Data-driven categories often measure the data rather than the intended concept.
• Declaring something part of humanness is an evaluative act, not descriptive.
• Declining to choose a mapping is itself a conceptual choice
• Defining requires prior knowledge of a thing’s properties before selecting which ones matter.
• Definitional completeness and logical structure are independent choices in concept building
• Definitions based on underlying structure explain more than those based on surface qualities.
• Definitions demand completeness while indicators reward redundancy
• Definitions of mental disorder explicitly exclude culturally expected grief and social deviance
• Democracy is multidimensional, encompassing contestation, participation, accountability, and civilian control.
• Democracy measures contradict the necessary-condition concept they claim to operationalize.
• Democracy measures contradict the necessary-condition definitions scholars endorse.
• Democracy requires elected officials to act free of unelected and external control.
• Demonstrating joint sufficiency requires studying cases where one cause is absent
• Deriving concept categories from data distributions is generally methodologically unsound
• Diagnostic categories like homosexuality and deafness reflect normative choices, not mere science
• Dichotomous poverty indicators function as weighted interval-level factors.
• Different democracy measures disagree most in the gray zone between regimes.
• Diminishing marginal returns justify logging GDP, encoding normative assumptions.
• Disciplinary defaults get hardwired until they become invisible
• Distinct freedoms link to each other causally and empirically, not constitutionally.
• Distinct social science concepts can share the same empirical referents, creating overlap zones.
• Drawing a sharp boundary through the competitive-authoritarian gray zone is deeply problematic.
• Each dimension enters the aggregate separately and cannot be swapped for another.
• Each single necessary condition is insufficient, but their joint presence produces the outcome.
• Easy cases inflate reliability while hybrid regimes expose measures’ true error.
• Economic utility and production functions collapse into the same three substitutability zones.
• Equal-percentile cutoffs produce deeply unequal measured categories
• Equal-sized categories usually signal unexamined rather than considered conceptualization.
• Equating concepts with variables erases the conceptual stage of measurement.
• Essentialism is logically equivalent to necessary-and-sufficient condition structure.
• Euclidean distance is just one aggregation choice among many in ideal point estimation
• Every choice of concept structure smuggles in a normative position
• Every complex concept structure reduces to a two-level, three-box building block
• Every moral theory implicitly commits to an aggregation rule over its values.
• Exhaustive typologies expand by inventing new types until every case fits
• Exponents above one on bounded scores punish weakness more than equal weighting does.
• Extension can shift purely from changing concept structure, not intension.
• Extreme categories have one configuration while gray-zone cases have many
• Family resemblance and necessary-condition concepts are two ends of one m-of-n spectrum.
• Family resemblance concepts operate on sufficiency alone, letting one feature substitute for another.
• Family resemblance concepts replace necessary conditions with substitutable sufficient ones
• Fixing indicator goalposts across decades is required for meaningful time comparisons.
• Flawed classification systems persist because sunk costs and absent alternatives outweigh their failure.
• Forcing patients into a single diagnosis drives inappropriate drug treatment.
• Fuzzy-set analysis rejects the classic objection that dichotomizing continuous data destroys information.
• Fuzzy-set calibration makes the steepness of the middle zone a choice, not a fact
• Global composite indices implicitly contain at least three structural levels
• Good indicators should be redundant measures of the same thing, not complete
• Good measurement means interrogating defaults against local substantive context
• Graded categories like World Bank income bands create partial membership across neighboring concepts.
• Gray zone concepts require first fixing the boundaries on either side.
• Gray-zone categories are defined by double negation, which naturally invites family-resemblance structure.
• Gray-zone categories can be scored by how many defining criteria a case meets.
• Gray-zone concepts are built by joining necessary conditions with negations of the two poles.
• Health is defined as more than the absence of disease.
• High correlations can conceal total disagreement in the middle of a scale.
• How dimensions and indicators combine is the decisive choice distinguishing measurement approaches
• How you aggregate two states’ democracy scores changes democratic peace results.
• Human concept use routinely violates necessary-and-sufficient-condition logic.
• Human sciences require fuzzy logic because concepts have unsharp boundaries
• Hyphenated gray-zone concepts pull meaning toward the middle rather than intensifying it.
• Ideal types occupy the unattainable maximum of a concept’s continuum, not a separate category.
• In effect-indicator models the latent variable is the cause and the indicator the effect
• In psychology and education, latent variables are standardly modeled as causing their indicators
• In substitutable concepts, one dimension can fully compensate for another.
• Independent researchers repeatedly rediscover the same two-level conceptual structure.
• Indicator lists can never be complete and should be redundant.
• Indicators are cheaply swappable while defining dimensions require normative defense
• Indicators of a single dimension can correlate poorly yet still be combined.
• Intensity-modifying adjectives operate as supersetting or subsetting operations
• Interactive concepts produce nonlinear outcomes that regression models systematically misrepresent.
• Interpreting an interaction term assumes zero is a meaningful value of the other variable.
• Knowledge of causal mechanisms should determine how concepts are defined.
• Latent variable models assume averaging logic, which never represents weakest-link concepts.
• Latent variable models build causal links that definitional frameworks reject
• Latent-variable models never settle whether indicators define or merely signal a concept.
• Linear interpolation makes a poor gray zone; its boundaries must be defined conceptually.
• Logging is standard practice yet its effect on semantic validity goes unexamined.
• Mandating mutually exclusive and exhaustive categories manufactures empty types and distorts real cases.
• Many core social science concepts are incomplete and require added criteria.
• Many core social-science concepts are defined by negation rather than positive attributes.
• Many historical regimes were simultaneously partly democratic and partly monarchical.
• Mathematical aggregation must be faithful to a concept’s semantics.
• Measurement and validity are causal concepts, not correlational ones
• Measurement is downstream of semantics, not the reverse.
• Measuring a quantity and defining it cannot be answered apart
• Medieval constitutionalism plus light war mobilization jointly produced European democracy
• Membership in ‘high test score’ is sharply asymmetric, saturating only in the top percentiles.
• Mental disorders often do not fit within a single diagnostic category.
• Methodological convenience, not theory, often determines how concepts get defined
• Min-max rescaling has become the default method for building global indicator indices.
• Minimal conceptualizations are incompatible with ideal types and should be rejected
• Minimal definitions built only on uncontested features are almost always incomplete.
• Mismatch between a concept’s definitional logic and its aggregation math signals conceptual confusion.
• Modifier adverbs shift degree of set membership by fixed mathematical inequalities
• Monitoring and sanctioning each have multiple substitutable routes, so no single institution is essential.
• Most mixed regimes are hybrids, but only some hybrids warrant a separate type label.
• Multi-word concepts hide unresolved questions about how their parts interact.
• Multiplying interval variables yields meaningless results because their origin is arbitrary.
• Mutual-exclusivity rules erase real cases that belong to two adjacent categories
• Mutually exclusive diagnosis harms patients by hiding causal links between overlapping disorders.
• Naming democracy a mere procedure does not strip its normative content
• Narrow diagnostic categories produce high comorbidity and heavy not-otherwise-specified use.
• Necessary conditions draw category boundaries while family resemblance structures organize what lies inside.
• Negated concepts have different causes and effects than their positive poles.
• Negating a middle category yields a comparison group too heterogeneous to interpret
• Nominal scales lack an underlying continuum that lets other scales convert.
• Nominal typologies often conceal a two-poled continuum with a gray zone.
• Nominal variables are not scales because no single continuum underlies them.
• Normative assumptions about family roles corrupted empirical welfare state concepts.
• Normatively loaded terminology corrupts datasets and downstream causal analysis.
• Nouns must be converted to adjectives before they can be measured.
• Numeric measurement includes dichotomous presence and absence coding
• Official definitions of health diverge sharply from the indicators actually measured.
• Once a threshold is crossed, further gains stop registering in the measurement
• One-of-n and n-of-n rules define opposite poles of conceptual inclusion
• Only necessary-condition concepts force added dimensions to shrink empirical coverage.
• Only two dimensions suffice to generate a concept’s full structural decisions
• Opposite concepts need separate definitions rather than logical negation.
• Ordered typologies distract from the continuous concepts that generate them
• Ordinal categories built on an underlying ratio variable reveal unequal spacing between groups
• Ordinary concepts are degree membership sets measured by fuzzy-logic thresholds
• Outsiders can legitimately identify real welfare thresholds in another society.
• Overlapping categories arise both conceptually and empirically, producing comorbidity by definition.
• Overlapping symptom criteria mean DSM mental disorders are not qualitatively distinct.
• Party, legislative, or reserved seats each independently count as a gender quota
• Policy attention requires problem, solution, and political context to converge simultaneously.
• Political activism, not clinical evidence, removed homosexuality from psychiatric diagnosis.
• Polity’s regime score implies zero is the conceptual midpoint between democracy and autocracy.
• Polity’s three democracy categories are arbitrary, driven by symmetry rather than substance.
• Poverty rankings shift with normative weight given to the poorest
• Problems with mutually exclusive and exhaustive typologies are structural, not case-specific.
• Psychiatric diagnosis rests on necessary conditions plus m-of-n symptom counts.
• Published analyses can abandon the methodology their own programmatic statements demand.
• Qualitative difference claims reduce to differences of degree on a continuum.
• Quality-of-life indices can rate a deaf life as worth less than a hearing one.
• Raising a classification threshold shrinks the population as a nested subset.
• Re-asking the same consumption item requires wording that flags one answer for exclusion
• Real-world categorizers create unequal, nonlinear categories rather than evenly spaced ones.
• Real-world indicators cut continua into unequal, non-linearly-spaced categories.
• Recoding a defining dimension to zero has indeterminate effects on extension.
• Redundancy in necessary components makes systems more stable.
• Redundant dimensions are functionally substitutes and therefore not necessary conditions
• Redundant typology cells collapse into an opposite pair that enables interval measurement
• Reforming a field’s terminology can unintentionally rewrite its substantive theory.
• Regional expertise shapes which features scholars add to a concept.
• Regressing human rights on democracy is definitionally circular, not causal testing.
• Researchers log data to fix skew, not to preserve conceptual meaning.
• Researchers rarely specify whether typology middle cells are ranked or tied.
• Researchers routinely treat high-level ordinal scales as interval measures.
• Reversing scale items produces contraries that do not exhaust all possible opinions
• Rigid mutually exclusive categories misdescribe people of mixed ethnic or religious background.
• Rule-driven category creation produces both discovery and dead categories
• Scale-type rules about permissible operations never crossed into political science and economics.
• Scholars agree on democracy but not on what autocracy means.
• Scholarship emphasized substitutable family resemblance while ignoring best-shot concepts
• Scholarship on intension and extension never considered OR-structured concepts.
• Science exposes pre-existing natural kinds and causal mechanisms
• Secondary concept dimensions can visibly shift democracy scores and required explanations.
• Secondary-level dimensions can causally produce the basic-level factors theories treat as causes.
• Secondary-level variables in two-level theories are causes, not mere measures, of the outcome.
• Semantics and normative goals, not data distribution, should drive standardization
• Semi-presidentialism requires a popularly elected president plus a premier answerable to parliament
• Set membership requires adjective-like concepts, not raw quantities.
• Setting measurement thresholds is a normative judgment about who deserves support.
• Shared dimensions between measured variables produce definitional, not causal, correlations.
• Small changes near a concept’s midpoint carry more meaning than changes at the extremes.
• Some concepts are measured directly while others require fallible symptom indicators
• Spending-based welfare state measures omit rights central to women’s welfare.
• Spreadsheet structure silently enforces categorical exclusivity on researchers.
• Standard poverty measurement begins by drawing a dichotomizing poverty line.
• Standardization across dimensions requires only a shared low-to-high ordinal continuum.
• Standardization can make autocracies the most extreme outliers by construction.
• Standardization choices can reverse which findings count as statistically significant.
• Standardizing scores makes one case’s value depend on other cases’ values.
• States missing exactly one necessary condition become de jure or de facto states.
• Statistical latent-variable models require correlated indicators, but concept definitions require separate, nonredundant attributes.
• Statistical means should not automatically define a concept’s membership midpoint.
• Statistical measurement originated in repeated observations of one physical quantity with random error.
• Statistical results can be driven entirely by the ignored comparison pole
• Statistical standardization serves rescaling, not semantic validity, so researchers should avoid it.
• Statistical typology methods silently bake in exhaustive, mutually exclusive category rules.
• Strict incommensurability makes defining any complex concept impossible.
• Substitutability between dimensions varies with their values, not just fixed parameters.
• Substitutable means to an outcome become causally relevant only when none is available.
• Subtype talk obscures that only a dimension’s 0 or 1 coding changes.
• Summing nominal or ordinal indicators is statistically invalid without an interval-scale argument.
• Swapping the comparison group can flip a coefficient’s sign or significance.
• The arithmetic mean gives moderate, not perfect, substitutability among components
• The democratic peace is really a democratic absence-of-war.
• The midpoint of a scale is rarely the real crossover for a concept
• The negation of a positive concept differs from its supposed opposite.
• The Polity index gives the 1900 United States a maximum democracy score
• The poor produce many goods themselves, so market-based poverty measures undercount them
• The scope of a regime concept is all state-years, including zero-membership cases.
• The World Bank’s income thresholds are arbitrary breakpoints on a continuum.
• Theoretical propositions require symbols with substantive empirical content.
• Theories of peace, wealth, and democracy do not explain their opposites.
• Theories use concepts, while data contain only indicators needing transformation.
• Theory, not everyday usage, should drive case selection in social science.
• Threshold placement matters more where population density is highest on the curve
• Threshold rules built from m-of-n criteria inevitably generate continua rather than sharp boundaries.
• Threshold rules in composite indices are often adopted without explicit justification.
• Thresholds form a continuum from linear to dichotomous transformations.
• Treating a partially ordered typology as nominal leaves its comparative hypotheses untested
• Turning a continuous measure into a concept requires setting a threshold.
• Two unrelated motivations, fuzzy semantics and data outliers, converge on the same transformation.
• Two-by-two typologies with tied middle cells form partially ordered scales.
• Two-way typologies dissolve into two interacting dimensions when analyzed.
• Typology orderings arise from either conceptual aggregation or causal outcome levels
• Under OR aggregation, adding defining features must expand a concept’s extension, not shrink it.
• Unmeasurable concepts like intelligence and ideology get measured via observable proxy indicators.
• Urban radicalism alone cannot produce social revolution in agrarian societies
• Utilitarianism yields linear poverty measures while Rawlsian ethics weights the poorest
• Valid ordinal scales require deciding level count, midpoint, and category spacing
• Variables in empirical analyses sit at the basic level, so variable means concept.
• Weakest link, mean, and best shot are zones on one substitutability continuum.
• Weighted multiplicative aggregation of concept dimensions is the Cobb-Douglas production function.
• Weighted sums in latent variable models silently convert ordinal indicators to interval.
• Welfare states are defined by sufficiency across multiple services, not one necessity.
• Well-being indices list necessary components but rarely specify sufficiency.
• What counts as an object’s essential properties depends on theoretical purpose
• What counts as an outlier depends on substantive theory, not statistics.
• When one measure is a subset of another, their disagreement is patterned, not random.
• Where a boundary is drawn in a classification determines which countries get aid.
• Whether concepts overlap in the data is an empirical question, not a design assumption.
• Which dimension drives a country’s composite score determines the correct causal explanation.
• With AND aggregation, adding necessary defining features can only shrink a concept’s extension.
• Zero’s meaning shifts silently when an interval scale gets dichotomized.
• Zeroing a defining dimension can shrink or expand a concept’s extension unpredictably.