Media consumption and reading performance

Short-form videos have become a routine source of information, entertainment, and language exposure. On platforms such as TikTok, Instagram Reels, and YouTube Shorts, viewers often encounter speech-synchronized captions, rapidly changing text, and—in some cases—words presented one at a time in the same location on the screen.

These displays look surprisingly similar to Rapid Serial Visual Presentation (RSVP), a method long used in reading research. In RSVP, words or short phrases appear sequentially at a fixed location. Unlike conventional reading, readers cannot look ahead or easily return to earlier words. Information disappears as the sentence unfolds, increasing the need to retain and integrate it in memory.

My dissertation asks whether habitual short-form-video consumption is associated with how people comprehend sentences under these different presentation demands. More broadly, it asks a question that is easy to frame too simply: Do emerging media habits reduce reading ability, or do they change the kinds of reading environments to which people are best adapted?

Project status: This page describes a proposed dissertation study. The relationships below are hypotheses to be tested, not established effects of short-form media.

The central research question

The study examines how readers recover when their first interpretation of a sentence turns out to be wrong. It tests whether this recovery depends on:

• the amount of short-form media a person consumes;
• their verbal working-memory capacity;
• their vocabulary knowledge and broader language experience; and
• whether the sentence remains visible or disappears as they read.

The goal is not merely to ask whether high short-form-video use is “good” or “bad.” Instead, the study investigates when differences emerge and which mechanisms might explain them.

Why use ambiguous sentences?

Language comprehension is incremental: readers begin building an interpretation before reaching the end of a sentence. This usually makes reading efficient, but it can also lead the reader down a temporary “garden path.”

The experiment uses noun phrase/zero complement (NP/Z) ambiguities, a classic garden-path construction. Consider the following pair:

In the ambiguous version, readers may initially interpret the deer as the object of hunted. When they reach paced, that interpretation no longer works: the deer must instead be understood as the subject of paced. The reader therefore needs to detect the conflict, revise the initial structure, and update the sentence meaning. The comma in the unambiguous version prevents the initial misanalysis.

This construction is useful because successful recovery draws on several abilities at once. Readers must retain earlier words, notice the disambiguating information, revise a structure they have already built, and suppress an interpretation that may continue to feel plausible. A comprehension question such as “Did the man hunt the deer?” can reveal whether the initial interpretation still lingers after the sentence has been read.

Two ways of presenting the same sentence

The study compares two self-paced presentation formats. Participants press a key to reveal each new segment, so reading time can be measured at the disambiguating word and immediately afterward.

Cumulative linear presentation

Segments appear from left to right in their normal sentence positions and remain visible. Readers can look back at earlier material, making this condition closer to conventional reading.

While the man hunted   the deer   paced   in the zoo.

Centered presentation

Each segment appears in the same central location and replaces the previous segment. This approximates the visual experience of RSVP while remaining self-paced.

While the man hunted  →  the deer  →  paced  →  in the zoo.

The centered format removes rereading and parafoveal preview—the ability to obtain information from words just beyond the point of focus. Readers must therefore rely more heavily on information maintained in working memory. If the initial parse is wrong, revision may be especially difficult because the earlier sentence is no longer available for inspection.

Three dimensions of individual experience

Short-form media consumption

Participants will provide three weeks of device screen-time records, supplemented by a questionnaire about their broader media habits. The primary measure will be average daily time spent on short-form-video platforms. A secondary proportional measure will estimate how much of a participant’s total media environment consists of short-form content.

This distinction matters. One hour of short-form video may represent a small part of one person’s media use but nearly all of another person’s media exposure.

Reading span

A reading-span task will measure verbal working memory: the ability to process sentences while temporarily retaining other verbal information. This capacity may help readers preserve the developing sentence representation and revise it after encountering a disambiguating word.

Vocabulary knowledge

A 50-item vocabulary task will provide an indirect measure of language experience and lexical quality. Strong, readily accessible word knowledge can reduce lower-level processing demands, leaving more resources available for integrating and revising sentence meaning.

Two competing pathways

The proposal considers two explanations that may operate simultaneously.

1. Cognitive-linguistic mismatch

Greater short-form-video consumption may be associated with differences in attentional control or working memory. Short-form media may also provide a linguistic environment dominated by brief, highly contextualized, or structurally simple expressions. If so, some readers may have less practice maintaining and revising complex sentence structures.

Under this account, the challenge is best described as a mismatch, not necessarily a general loss of reading ability. Cognitive and linguistic habits reinforced by one media environment may be less well suited to tasks that require sustained integration and structural reanalysis.

2. Adaptation to dynamic text

The alternative is that repeated exposure to rapidly changing captions makes RSVP-like displays more familiar. Readers with more short-form-media experience might still show lower performance overall, yet experience a smaller additional cost when text is presented sequentially in the center of the screen.

This possibility complicates a simple “more media equals worse reading” account. A habit that is disadvantageous in one format could offer a limited benefit in another.

Study design

The planned study will recruit at least 80 English-speaking adults between 18 and 55 years of age. Participants will complete:

1. a media-use questionnaire and submit recent screen-time records;
2. a vocabulary assessment;
3. a web-based reading-span task; and
4. a sentence-reading task crossing ambiguity with presentation format.

The reading experiment contains 80 target sentences distributed across counterbalanced lists. Each participant will encounter ambiguous and unambiguous sentences in both cumulative and centered presentation formats.

The primary outcomes are:

• comprehension accuracy, especially whether readers retain the initial garden-path interpretation;
• reading time at the disambiguating region, where the initial interpretation becomes untenable; and
• spillover reading time, which captures processing that continues immediately afterward.

Mixed-effects models will be used to estimate how ambiguity, presentation format, and each individual-difference measure jointly predict reading time and accuracy while accounting for variation across readers and sentence items.

Predictions

The proposal makes four main predictions:

1. Ambiguity should reduce comprehension accuracy. This cost may be larger when earlier words disappear in the centered presentation format.
2. Reading span and vocabulary should support comprehension. Readers with greater verbal working-memory capacity and stronger lexical knowledge are expected to recover from temporary ambiguity more successfully.
3. Reading time can reveal different strategies. Longer reading at the disambiguating word may reflect greater sensitivity to the conflict and more complete reanalysis. A faster response is not necessarily better; it may sometimes indicate shallow or “good-enough” processing.
4. Short-form media may show both disadvantage and adaptation. Greater consumption may predict lower accuracy overall, but this negative association may become weaker in the RSVP-like centered condition if familiarity with dynamic text provides a small format-specific benefit.

The study will also test whether short-form media consumption, reading span, and vocabulary covary. Because the design is observational, any association will be interpreted cautiously: media use may predict reading performance without demonstrating that it caused the difference.

Why this project matters

This dissertation connects research traditions that rarely speak directly to one another. Communication and media psychology have examined the cognitive consequences of short-form-video use, while psycholinguistics has developed precise methods for studying ambiguity resolution, memory demands, and moment-to-moment reading. Bringing these approaches together makes it possible to ask how everyday media experience relates to specific language-processing mechanisms.

The project also treats short-form media as a potential source of language experience, not only as screen time. Existing language-background measures often ask about books, television, radio, and conversation but do not adequately capture dynamic captions and rapidly paced multimodal language. For many younger adults, these formats now constitute a substantial part of daily linguistic input.

Ultimately, the study aims to replace a broad question—“Does short-form media harm reading?”—with a more informative one: How do media habits, cognitive resources, language experience, and text presentation interact to shape comprehension? The answer may reveal both costs and adaptations, and may help reading research keep pace with the rapidly changing environments in which people encounter language.

Questions or ideas about this project are welcome. Please feel free to contact me.