The attitude filter hypothesis: how strong opinions influence text processing

Deshaies, Sarah-Elizabeth Marion-Antonia

Permalink

https://hdl.handle.net/2142/115578 Copy

Description

Title

The attitude filter hypothesis: how strong opinions influence text processing

Author(s)

Deshaies, Sarah-Elizabeth Marion-Antonia

Issue Date

2022-04-22

Director of Research (if dissertation) or Advisor (if thesis)

Christianson, Kiel

Doctoral Committee Chair(s)

Christianson, Kiel

Committee Member(s)

• Morrow, Dan
• Stern, Chadly
• Kern, Justin

Department of Study

Educational Psychology

Discipline

Educational Psychology

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Keyword(s)

• TL Effect
• Reading
• Eye-Tracking
• Opinions
• Attitudes
• Attitude Filter Hypothesis
• Transposed Letter Effect

Abstract

There has been a great deal of research surrounding attitudes or opinions, how they are created, how they influence behavior, and how they can be measured (Petty & Korsnick, 1995). However, what is not yet understood is how strong opinions influence text processing. Do strong opinions interact with lower-level orthographic or word order processing? Or do strong opinions influence post-interpretive processes interacting with just memory? The former will be addressed in this dissertation, the latter will be an open question for future investigation. In this dissertation, four experiments were conducted. Three were self-paced reading experiments whereby the participant read sentences expressing opinions about both political and apolitical topics one word at a time. The other was an eye-tracking experiment, in which participant’s eye movements were recorded as they read the same sentences presented in two of the self-paced reading experiments. In the experiments reported here, the goal was to determine the effect of strongly held opinions on the processing of low-level features of texts, specifically orthographic processing and word order processing. To examine orthographic processing, the Transposed Letter Confusability Effect was used to determine whether readers with strong opinions about the topics would be more or less likely to read misspelled words with transposed letters (e.g., insipre) or substituted letters (e.g., insqare) as if they were correctly spelled. To examine word order (syntactic processing), the Transposed Word Confusability Effect was used to determine whether readers with strong opinions about the topics would be more or less likely to notice word transpositions in the sentences. Across all experiments, there was evidence of strong opinions influencing text processing. Results from the self-paced word-by-word reading experiments in Chapters 3 and 4 replicated the classic transposed-letter effect, whereby reading times on the transposed-letter word was not significantly different from its correctly spelled Base, and that there was significantly increased reading time on the substituted-letter (SL) word. However, reading times on the SL word were reduced when participants strongly agreed or strongly disagreed with the statement, indicating failure to notice this normally disruptive spelling. The eye-tracking experiment in Chapter 5 confirmed these results, indicating that early word processing of the Base and the TL word were identical, but that later processing was influenced by the participant’s ability to create a concrete understanding of the text when the sentence contained a TL-word. It was not found that readers with strong opinions simply skipped these critical words in sentences they (dis)agreed with. The final experiment in Chapter 6 used the transposed word effect (Mirault, Snell, & Grainger, 2018), analogous to the transposed-letter effect, but at the word level. A study by Gilead, Sela, and Maril (2019) found evidence of scrutiny of text when the participant did not agree with an ungrammatical sentence. The results of Chapter 6 were mixed in their ability to replicate these prior findings. However, there was an effect of opinions influencing grammaticality judgements such that those who agreed with an item were less accurate when the words were transposed across a verb phrase boundary. Taken together, the results reported in this dissertation suggest that readers who hold strong opinions about topics expressed in texts process those texts in qualitatively different ways from readers who do not hold strong opinions. This may have an effect on how we model reading and on how we understand information processing in the social sphere (e.g., on social media, in the press, etc.) when topics elicit strong opinions. Future studies could explore this on the post-interpretive level of text processing as the results are also relevant to educational settings, where opinion-oriented reading could be explored, such that students could be asked to agree, disagree, or remain neutral. Then the “facts” that students obtain could be compared.

Graduation Semester

2022-05

Type of Resource

Thesis

Copyright and License Information

Copyright 2022 Sarah-Elizabeth Deshaies

Owning Collections