What's in a Title?

Abstract

I tell my students that if they turn in a manuscript for their semester-long research project with a title that includes “Factors affecting…” (or similar) I take 25% off their on their grade automatically. Is this harsh? Perhaps, but it gets them thinking about their title maybe for the first time. So what's in a title? The guidelines for the Journal currently do not say much about formulating an effective title except to keep it short and avoid abbreviations unless necessary. Yet, arguably the title is the most important element of a manuscript (Paiva et al. 2012). Is what makes a good title too intuitive, too specific to the individual article, or too personal to the authors for much to be said? Most authors are interested in having their articles cited, thus research on what title characteristics influence citation rate might just give us something to think about. Realizing that a well-written title serves several purposes is the first clue to a good title. The title attracts the reader's attention, informs the reader of the content, and serves as key words for database indexing and referencing purposes. Balancing these elements is not always simple. We could rely on computer algorithms that use Natural Language Processing for automatic titling as is done for webpages, blogs, and emails to provide informative and catchy titles (Lopez et al. 2014). But most authors still prefer to choose their own titles, sometimes even before they write the article. Maria Grant, Editor of Health Information and Libraries Journal, writes in her editorial, What makes a good title?, that article titles should be concise, informative, and where appropriate give details of the research design (Grant 2013). Keeping the title short is a long-time axiom (10 words or less in the Journal guidelines). But evidence for short titles rendering more citations is mixed (Hartley 2007, Jamali and Nikzad 2011, Paiva et al. 2012). For example, Subotic and Mukherjee (2014) compared 129 psychology journals in ScienceDirect's Top 25 Hottest with 129 articles not appearing on the Top 25 list and found articles with shorter titles had more citations, but the influence of journal impact factor was far greater than title length. They found titles associated with an above-median number of citations had roughly 10.4 words. In contrast, Habibzadeh and Yadollahie (2010) found longer titles with more words had higher citation rates in 2005 in 22 arbitrarily chosen English-language journals, and the association was even more pronounced for journals with high impact factors. High impact journals might require shorter titles, and title length is correlated with number of authors and length of articles (Yitzhaki 1994, 2002), so it can be difficult to know what the actual cause is. Nevertheless, when it was the human eye that had to search a multitude of titles, short titles were essential. In the age of digital searchers, longer titles may allow authors to specify more information that provides an advantage to the article with longer titles. But when full-text indexing becomes common, we may be brought full circle leaving the focus on other key elements of the title than length (Jamali and Nikzad 2011). We looked at word counts of titles of articles published in the Journal over the past 5 years (volumes 73–77, n = 960) and found the average word count was 11.5 words, just over the Journal guidelines of 10 words. There was some evidence for a consistent increase in the average number of words from 10.6 to 12.1 words over this brief time, but we found no differences among articles addressing taxonomic groups or subject matter (e.g., habitat, populations, economics, techniques, etc.). Research has shown that including a geographic region does not improve citation rate and might even limit it (Jacques and Sebire 2010, Paiva et al. 2012). This was attributed to readers avoiding articles narrowly focused on a geographic area. We found 35% of the articles published in the Journal listed a location in the title. Use of colons in titles of science articles has increased particularly among single-authored articles (Lewison and Hartley 2005). Colons in a title allow some freedom to combine 2 elements efficiently and titles with colons tend to be shorter. In a preference study, students and academics generally preferred titles with a colon than without (Hartley 2007). But in terms of citations, results again are mixed. Jacques and Sebire (2010) reported that the presence of a colon in the title was positively correlated with the number of citations, whereas Hartley (2007) found a colon did not affect the number of citations received. On the other hand, title content and approach matters. Paiva et al. (2012) examined titles from PLoS and Biomed Central journals and found a stronger association with citation rate for “result-based” than “method-describing” titles, indicating what the readers want to know are the results. Only 4% of the titles we surveyed in the Journal indicated a method or approach in the title. This is not to be confused with papers introducing a novel technique, which are often highly cited. Authors can take several approaches to inform the reader of the results. Jamali and Nikzad (2011) classified titles as 1) declarative, where titles indicate the main finding (e.g., Pre-commercial forest thinning alters abundance but not survival of snowshoe hares), 2) descriptive, where it describes the topic but not the finding (e.g., Targeted surveillance of raccoon rabies in Quebec, Canada), and 3) interrogative (e.g., Does aversive conditioning reduce human—black bear conflict?). They examined the association of types of titles with number of downloads and citations in over 2000 articles in 6 of the PLoS (Public Library of Science) journals. Unexpectedly, declarative titles that presented conclusions up front were both downloaded and cited less than descriptive titles; although the citation difference was not significant, they speculated that up to a point ambiguity may lead to downloading and reading the article, whereas the known result may not be what the reader is specifically seeking. Question-based titles have increased over the last 4 decades in the fields of physics, life sciences, and medicine (Ball 2009). Paiva et al. (2012) found titles with question marks were cited less than those without, whereas Jamali and Nikzad (2011) found question-based articles were downloaded significantly more often but not cited more than other types of titles. Question-based titles may evoke curiosity so that readers download an article, but article content influences citation (Gustavii 2008, Ball 2009). We found that titles of articles in the Journal are overwhelmingly what we considered descriptive (96%). Does humor attract readers? Sagi and Yechiam (2008) explored humorous or amusing titles in psychology articles and reported that citation rate increased slightly with increasing rank of humor, but in the extreme (ranks that were 2 standard deviations above the average) articles received fewer citations. Subotic and Mukherjee (2014) found title amusement level was weakly but positively correlated with downloads, but not citations. Sagi and Yechiam (2008) have argued that humorous titles may harm credibility, but Subotic and Mukherjee (2014) found no negative effect for less extreme ranks. One possibility is that humorous titles are perceived as communicating non-serious subject matter because readers of scientific papers have specific expectations, or that humor might signal low quality (Armstrong 1989). Some amount of title amusement might help in attracting a bigger audience leading to more downloads, but it seems likely other factors play a larger role in actually getting citations. Judging humor of a title is subjective but by our standards, we found few (<2%) articles published in the Journal used humor, and most titles bordered on being boring. So again, what's in a title? It may be unrealistic to think that simply changing the title of an article will improve its citation rate alone. In the end, a paper stands on its total merits: creativity and novelty in how the question is addressed, appropriate design, rigor in analysis, and quality of presentation. These are the elements on which I actually grade the student's papers in the end. However, optimizing the title is still an important task. A predator can't eat a prey until it has encountered it; a collegue cannot cite your paper until they have read it. The title attracts the reader's attention, informs the reader of content specifics, and ends before it confuses or loses us. Sound good? Then please, go beyond the mundane “Factors affecting…” —Evelyn Merrill Editor-in-Chief and —Anna Knipps Editorial Assistant

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 13 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.