Inferencing

Learner Variable · All learner variables
Evidence · 36 claims (26 show a difference, 6 mixed, 4 no difference) · 18 studies (10 causal, 4 review, 4 associational), q1–q4 · 5 of 18 report an effect size · 35 claims rest on one study

Description

Inferencing is the work a reader does to supply meaning a text does not state outright: linking one sentence to another, connecting the text to what the reader already knows or has experienced, and adding what is learned to existing knowledge. It supports reading comprehension, and the Learner Variability Navigator notes that it develops with age, moving from inferences drawn from within the text towards deeper inferences that draw on the reader's own knowledge. It is usually assessed through comprehension questions that ask for information the text implies rather than states; LVN lists no dedicated measure for it.

It matters most in middle and high school, where subject-specific texts assume background knowledge and expect readers to connect it to new information. It depends on prior knowledge and on working memory, especially for inferences that link ideas far apart in a long text, and it is part of what academic and critical literacy asks of a reader. It is a component of the comprehension side of Reading and Language.

Implications

Context

• Content-area reading in the secondary grades leans heavily on inferences that connect background knowledge to the text.
• Building background knowledge before reading, and asking questions that require readers to connect parts of a text, are the usual supports.

Target Learners

• Adolescent readers who decode and understand sentences but miss what a text implies.
• Readers with limited background knowledge of a topic, and readers with lower working memory spans, for inferences across long texts.

Target Learning Objectives

• Reading comprehension, especially of content-area and figurative texts.
• Following and evaluating arguments that leave steps implicit.

Claims

• Adding multiple design elements to the basic ITS structure promotes causal inference by limiting alternative interpretations, demonstrated with No Child Left Behind [+W] — instruction changes it
• Contrary to the science comprehension thesis, political polarization of data interpretation did not abate among the most numerate subjects but increased [+M] — learners who differ on it differ in outcomes
• Correct word inference in reading requires knowing roughly 95-98% of surrounding words, corresponding to about 3,000-5,000 word families [+W] — learners who differ on it differ in outcomes
• Subjects highest in numeracy draw valid causal inferences from data better than less numerate subjects when the data are politically uncharged (skin rash treatment framing) [+M] — learners who differ on it differ in outcomes
• For readers with misconceptions, the number of correct inferences generated during think-aloud reading predicts text recall [~W] — learners who differ on it differ in outcomes
• Incorrect inference generation is more common for elaborative than for coherence inference failures, indicating children are less aware which information is relevant for elaborative inferences [+M] — instruction changes it
• Less skilled comprehenders make fewer knowledge-based coherence and elaborative inferences than skilled comprehenders even when the relevant knowledge is equally available [+M] — instruction changes it
• Less skilled comprehenders' inference failures stem more from incorrect premise recall, whereas skilled comprehenders fail more at integrating recalled information [+M] — instruction changes it
• Low-SES students also generate stereotype-consistent inferences in the reverse direction, inferring SES indicators from memory performance information [+M] — instruction changes it
• Low-SES students, but not high-SES students, generate stereotype-consistent inferences linking targets' social class and memory ability [+M] — instruction changes it
• Low-skilled readers with strong background knowledge partially compensate for poor reading skill at the textbase level but not for inferencing [~M] — instruction changes it
• Skilled comprehenders recall more literal story content than less skilled comprehenders, but the groups do not differ significantly in interpreting novel similes [-M] — instruction changes it
• Students' SES and academic achievement level moderate their use of the SES stereotype [+M] — instruction changes it
• The comprehension-skill difference in inference making persists after controlling for memory for the text and delayed retention of the knowledge base [+M] — instruction changes it
• The frequency of associations produced during think-aloud reading predicts sentence reading times during silent reading [+W] — learners who differ on it differ in outcomes
• When explicitly required by direct questions, less skilled comprehenders' inference performance improves greatly while skilled comprehenders perform at ceiling [+M] — instruction changes it
• Inference-making performance improves with age between 6 and 10 years, with children performing better on global than local coherence inference questions [+M] — instruction changes it
• Prompting underspecified responses improves children's performance on both local and global coherence inference questions at all ages [+M] — instruction changes it
• Vocabulary fully mediates the effect of working memory on both local and global coherence inference making in 6- to 10-year-olds [+M] — instruction changes it
• Vocabulary is the only unique predictor of global coherence inference making at all three ages, 6 through 10 years [+M] — instruction changes it
• Vocabulary uniquely predicts local coherence inference making for 6- and 8-year-olds, but neither vocabulary nor working memory predicts it for 10-year-olds [~M] — instruction changes it
• Double-content examples had no significant effect on generating alternatives of action, drawing inferences, or taking multiple perspectives [-M] — instruction changes it
• ABM students engage more frequently in embodied reasoning strategies—first-to-third-person perspective shifts, causal inference making, and gestures—than video students, who rely more on third-person perspectives and information retrieval [+W] — instruction changes it
• First-to-third-person perspective shifting, causal inference making, and gesture use correlate positively with explanatory and transfer scores, while information retrieval and handwriting correlate negatively [+M] — learners who differ on it differ in outcomes
• Retrieval practice improved literal premise recall but did not improve deductive inference transfer compared to restudy [-W] — instruction changes it
• Initial skills moderate embodied training effects: arm-body training harms low-skilled students on dictation and legibility, and benefits high-skilled students on dictation [~M] — an instructional effect differs with it
• The training's effect on inference drawing was not limited to collaborative inferences but generally enhanced inference drawing of all types [+M] — instruction changes it
• Training improved inference drawing by increasing the efficiency of transforming pooled information into inferences rather than by increasing how often information was pooled [+M] — instruction changes it
• Training in collaboration strategies increases the frequency with which student dyads draw inferences during subsequent unscaffolded collaborative problem solving [+M] — instruction changes it
• Collaborative and individual inferences are less likely to be drawn than shared inferences during collaborative problem solving [+M] — instruction changes it
• Generating explanations (oral or written) outperforms retrieval practice on inference-based questions for the high-difficult text [+W] — instruction changes it
• Text difficulty moderates the effect of explaining modality: oral explaining outperforms written explaining on inference-based questions only for the high-difficult text [~M] — instruction changes it
• Narrative comprehension of bridging inferences remains stable across the adult age span [-W] — instruction changes it
• Linguistic inferencing uniquely predicts all three components of children's scientific reasoning over and above visuospatial inferencing [+W] — learners who differ on it differ in outcomes
• Visuospatial inferencing relates to hypothesis validation and experimentation but not to generating conclusions when linguistic abilities are controlled [~W] — learners who differ on it differ in outcomes
• Inference, but not theory of mind, is independently related to writing quality after accounting for transcription skills and working memory [+M] — learners who differ on it differ in outcomes

Measures

Related Learner Variables

• Writing Skills: inferring the links between parts of an argument supports understanding, and so building, arguments.
• Prior Knowledge: an inference reaches beyond the text into what the reader already knows, so background knowledge sets its limits.
• Academic and Critical Literacy: reading for bias and purpose, and reading within a discipline's conventions, both depend on connecting a text to outside knowledge.
• Access: readers who are deaf or hard of hearing may find it harder to draw correct inferences from text.
• Long-Term Memory: integrating what a text says into existing schemas in long-term memory is part of inferencing.
• Metacognition: awareness of one's reading goals and strategies goes with better inference-making and integration of ideas.
• Syntax: trouble with sentence structure can carry over into trouble inferring figurative meaning.
• Working Memory: inferences that link distant parts of a long text require holding information in mind.

Examples

Key Sources

• Digital Promise. Learner Variability Navigator: Inferencing (Building Literacy 7-12). lvp.digitalpromiseglobal.org. Licensed CC BY-NC 4.0.

Page metadata

Field	Value
type	learner-variable
id	inferencing
title	Inferencing
description	A reader's ability to fill in what a text leaves unsaid by connecting it with background knowledge, experience and other parts of the text.
status	review
canonical	True
generated	by claude/unspecified on 2026-10-16
trust	unverified — no one has confirmed this page's content

Frontmatter is the page's machine-readable record — see the Schema & Guide.