Readability and political discourse: An analysis of press releases of Ghanaian political parties

Press releases have enormous influence on public opinion since the media sets agenda for public discussion. Thus, press releases are essential communication tool for political parties. For press releases to be effective, however, they must be readable. Therefore, the main objective of this study  was to evaluate the readability of news releases of the two major political parties (National Democratic Congress (NDC), New Patriotic Party (NPP)) in Ghana. Seventy press releases (32 were NDC releases whiles 38 were from the NPP) from the NDC and NPP were selected by convenience sampling. Flesch readability indexes were used to compute readability scores. Frequencies, measures of central tendencies, and one sample T-test using bootstrapping, were used to describe readability of the press releases. In addition, independent sample T-test was used to compare differences in readability between NPP and NDC news releases. The results revealed that news releases by the two political parties were generally ‘difficult’ to read, compared to standard readability of public documents. This was the case since a person must have attained, on the average, over 13 years of formal education in order to be able to read and understand the news releases of the NDC (X =13.29; σ=1.85) while it required about 12 years of formal education was required, on the average, to read and understand the news releases from the NPP (X =12.22; σ=2.63). In addition, the results showed that there was no significant difference in readability between NPP and NDC news releases. It is recommended that the political parties consider the readability of the news releases before disseminating them to the public. 
 
 Key words: Readability, political discourse, press releases, media.


INTRODUCTION
The adoption of democracy in Ghana, as in many countries, has opened the door for the formation of multiparty political organisations.The formation of multiple political parties underscores ideological differences that exist between such parties.Although political parties differ in their ideological perspectives, they share a commitment to constitutional means of advancing their objectives, attempting to convince a population as a whole of their correctness, and putting their policies to the test of periodic elections (McNair, 2011).
For parties, clearly, the smooth functioning of the process described earlier is dependent primarily on their ability to communicate with those who will vote for and legitimise them.Political parties have used the public relations and media management techniques such as news conferences, and leaks to communicate their stands on matters to the masses.
In Ghana, the use of press releases is particularly popular.Press releases have enormous influence since the media molds the public's view of the outside world.The way in which the press chooses to edit and portray the news releases equally affects the public's view of the activities of any political party (Johnson and Haythornthwaite, 1989).
Often, the ruling party gives press releases to present information on the activities of the government.This is aimed at informing the populace and portraying the government in favourable light.This enhances their chances of being retained in power.In other instances, some government officials may engage in activities that may dent the image of the ruling party.
Press releases are used under such circumstances, to correct, denounce, or defend the supposed misinformation.This tactic is particularly rife in Ghana.In the eyes of the party in power this is obviously the most vital and over-riding function of them all.Ministers and government officials of course have the advantage of controlling what is actually written in the releases.
During question time, for instance, Ministers have access to all the relevant information, whereas the information in the press releases, which helps to initiate the queries, is only that which the Ministers themselves have decided to include.Indeed, press releases can sometimes appear more as pieces of propaganda literature, particularly from the central government departments who have some of a government's more controversial policies to implement.
Without effective press releases, the government could easily lose credibility through misinformation.In instances where the government is actually guilty of charges such as misappropriation of public funds (and the like) by government officials, press releases could be used as a defensive technique or even to apologise.The ultimate aim is to remain in the "good books" of the populace.
In contrast, opposition parties make effort to be heard by giving out press releases to clarify their stands on issues of national importance.Often times, press releases are used to communicate supposed "rot" in government in order to put them in bad light with the populace.This contributes to opposition parties" chances of being considered for power during election years.Obviously then, the use of press release in Ghanaian political communication is a big deal.
For press releases to be effective, they are to "be retold as accurately as possible, preferably even verbatim, in news reporting" (van Dijk, 2006).It has been indicated elsewhere in the cited literature that the fate of press releases is determined through the goodwill of journalists, and only if they could accept the style and contents selected for publication by the company press release writer, would the press release be published.Despite the critical functions of press Gyasi 43 releases to political entities, it is known that often times, political entities do not employ professional public relation personnel to manage information flow from the parties to the media (Ashcroft, 1994).
For this reason, a lot of press release may be written in a complicated style, which will prevent media houses from publishing them.Even if the media houses choose to retell the press release, especially verbatim, the readability of the release will affect the understanding the masses will get.In short, the masses may be misinformed if the press release is not easy to read, and understand for that matter.
In fact, it has been indicated elsewhere that a large chunk of press releases annually get ignored through poor style and targeting (Whysall, 2005).For this reason, it is important to consider the readability of press releases before churning them out to the public.Since readability scores are based on basic writing style criteria such as length of sentences, words and paragraphs, and the percent of passive voice sentences, the scores provide an estimation of writing style quality.By considering the readability of news releases improve the chances of the release being published and deliberated in the media to serve the function for which they were released.
Despite the many readability studies that have been undertaken since the 1940"s, very few have been conducted on press releases (Warren and Morton, 1991).The closest "relative" of readability studies on press releases that have been undertaken is the newspaper readability.Hence, an attempt is made in this study to fill this gap.The following research questions will therefore guide this study: (1) How readable are the news releases of Ghanaian political parties are measured in terms of Flesch Readability Index?(2) Is there statistically significant difference in the readability of news releases of Ghanaian political parties?
(3) Is there statistically significant difference in the frequency of polysyllabic words in the news releases of Ghanaian political parties?

Definition of readability and readability indexes
Several studies have previously discussed readability and readability in details since the 1920"s.Hence, it may serve no additional purpose revisiting the subject all over again.Nevertheless, briefly defining what readability is and the tool to measure it (readability indexes) will give context to the discussions that follows thereof.
Readability is an attempt to relate the reading level of printed material to the "reading with understanding" level of the reader.A readability formula brings out calculations on a text, based primarily on sentence and word length, and results in a numerical score.However, other elements affect the understanding of what is being read which cannot be calculated or measured in this way.These include the motivation of the reader, the size and the layout of written material, previous knowledge of the subject, and the style of the author Zamanian and Heydari (2012).
Different readability formulas are likely to give different scores on the same piece of text, as they give different readability levels to various facets of the text, and may have been created and tested on different types of text.Therefore, it is important to employ the same formula if you are trying to compare the readability score or level of a number of different texts Zamanian and Heydari (2012).
According to Crossley et al. (2011), "traditional readability formulas are simple algorithms that measure text reliability based on the length of a sentence and word length".Readability formulas are used to predict reading ease but they do not help in assessing the extent to which the reader will understand the content or ideas in the text.Traditional readability formulas are for evaluating the readability of a text, usually by counting words, syllables and sentences.

Some popular readability indexes defined
Some of the popular and commonly used formulas include: The SMOG Readability formula, Flesch"s Grade level, J. Peter Kinkaid"s Flesch-Kinkaid index, Robert Gunning"s Fog index, and New Dale-Chall formula Below, two of these formulas (Flesch Reading Ease and Flesch -Kincaid Grade Level) used in this study are reviewed.
The Flesch Reading Ease Readability Formula is one of the oldest, and it is considered to be the most accurate of all the formulas.It was developed in 1948 by Rudolph Flesch who is a writer, and a reading consultant.It is a simple approach to evaluating the grade-level of readers.This formula is mostly used for academic text.It is largely used to assess the difficulty of a reading text written in English language.
According to Owu-Ewie (2014), instead of using grade levels, this formula uses a scale from 0 to 100; whereby 0 corresponds to the 12th grade (Senior High School 3) and 100 is also equivalent to 4th grade (Primary 4).This simply means that the higher the score the easier the passage to be read and the lower the score the more difficult the passage (Appendix A).
Flesch-Kincaid Grade Level Test is a related test which translates the Flesch Reading Ease Test scores to grade level.The formula was propounded by Peter J. Kincaid, and his team in 1975.It is mostly used in pedagogy.This formula is used to determine the readability level of a variety of educational materials especially books.This formula makes it easier for parents, teachers, and librarians to select suitable reading texts for their children/learners (Owu-Ewie, 2014) (Appendix B).
There are numerous other indexes, some which may not be as popular as the two described here.In fact, there are about 200 of such indexes.Different factors have necessitated and justified the formulation of new indexes.Examples of these indexes include Coleman-Liau index, automated readability index, and the PSK index.

Empirical studies on readability of press releases
Press releases in general, and in the context of political communication, has received little attention.However, many studies have dealt with the readability of the eventual destination of press releasesnewspapers.Warren and Morton (1991) have listed, in chronological order, nine studies on the readability of newspapers.On average, the readability of the nine newspapers were suitable for 12th graders (Warren and Morton, 1991).
Besides these, Fosu (2014) investigates the readability and comprehensibility of English language newspapers in Ghana.His study attempted to discover the extent to which Ghanaian readers find the language of the newspapers easy or difficult to comprehend.He established that the language used to communicate socio-political news to readers is complex and difficult for a significant proportion of readers across the educational categories of the country.
Similarly, Hart (1993) points out that many newspaper stories have readability scores of 13 or higher, and these scores exceed the reading ability of the average U.S. adult.Likewise, Bodle (1996) compared the readability levels of a sample of student newspapers with a sample of private-sector papers.He found that the private-sector dailies had a higher score than the student papers.The yawning gap in readability of news/press releases makes this study essential.

Research design
The study was quantitative.It employed the descriptive design because the aim of this paper was to describe readability of news releases (Blessing and Chakrabarti, 2009).Descriptive research helps to use numerical data to analyse associations using mathematically based methods.Descriptive research involves identification of attributes of a particular phenomenon based on an observational basis (Skovsmose and Borba, 2004).

Population for the study
The target population consisted of all press releases of the two most dominant political parties in Ghana (National Democratic Congress (NDC) and New Patriotic Party (NPP)).These two parties have essentially dominated Ghanaian politics.They have very large followers.Hence, they are often the parties to issue press releases.

Sampling procedure and sample size
Convenience sampling technique was employed as the sampling procedure for this study.A convenience sample is a sample that uses individuals or sample units that are readily available rather than those that are selected to be representative or selected via a probabilistic mechanism (Li, 2009).The choice of convenience sampling procedure was informed due to the fact that the sample frame is very large (all press releases of NPP and NDC).There is no evidence that all press releases by these parties have been accurately kept.Hence, not all press releases by these parties were available for inclusion if probability sampling procedure was to be used.In all, 70 press releases were obtained as the sample size and included in this study.Out of these 70, 32 were NDC releases whiles 38 were from the NPP.

Data collection
Electronic copies (PDF) of 70 press releases of the two political parties were downloaded from their official websites.The texts within each press release were then copied into an online readability calculator to calculate its readability (www.usingenglish.com).This online calculator was used because it is very accurate and popular.Microsoft"s Word processor was not used to compute the readability scores because of its inability to compute Flesch Reading Ease (FRE) beyond 12 (See Appendix B for raw scores).Flesch reading ease and Flesch-Kincaid grade level (FKGL) were used to calculate readability scores.The FRE and FKGL were used because it is among the most popular, reliable and recommended indices for grading the readability of written materials (College, 2016).

Data analysis
With the help of IBM Statistical Products and Services Solutions (SPSS) version 23.0, frequencies, means, and standard deviations were used to describe readability, and the number of years required to read the press releases of these two parties (Research question 1).A one sample T-test using bootstrapping technique was used to determine whether there were differences in the reading ease and the school years required to read, compared to standard scores.
Independent sample T-test was used to compare differences in readability between the NPP and NDC news releases (Research question 2 and 3).Eta-squared was estimated to measure the magnitude or the effect size of the significance when there were significant differences (Field, 2011;Pallant, 2013).
Independent sample t-test was employed although a nonprobability sampling technique was used (convenience sampling).This was the case because the choice of either using a parametric or a non-parametric analytical procedure depends on assumptions other than the sampling techniques employed).Indeed, several authors have indicated assumptions such as normality of distribution, and independence of observation (Field, 2011;Pallant, 2013;Tabachnick and Fidell, 2007).
Bootstrapping (using samples of 1,000) was employed in both the one sample and independent sample T-tests in order to ensure robust estimates of significant or p-value, standard errors and the confident intervals (IBM, 2013;Tabachnick and Fidell, 2013;Field, 2013) when t-test, was significant.

Research question 1: How readable are the news releases of Ghanaian political parties measured in terms of Flesch Readability Index?
Table 1 describes the readability of four news releases of NPP and NDC.The readability was measured using the Flesch readability indexes.
Based on the FKGL, a person must have attained, on the average, over 13 years of formal education in order to be able to read and understand the news releases of the NDC ( ̅ ).The least difficult news release by the NDC required 9th graders (JHS leavers) to be able to read.The most difficult news release required 16th graders (college graduates).
In comparison, about 12 years of formal education was required, on the average, to read and understand the news releases from the NPP ( ̅ ).The easiest to read news release by the NPP required about 8 years of education.This was relatively much lower than that of the NDC.This implies that the easiest news release by the NPP was much easier to read than the easiest news release by the NDC.On the other hand, NPP produced the single most difficult news release, which required over 18 years of education to be able to read compared to the NDC"s maximum of about 16.
Considering both news releases on the FRE scale, they were all "difficult" to read, on the average.The mean readability for both parties were within the ranges of 40, classified as "difficult".The easiest to read was graded as "standard", and was one of the news releases of the NPP.Both parties produced some news releases that were "very difficult" to read.
Recommended readability level of news items for the public is 7 to 8 years of education (Cutts, 2013).Anything above 12 is considered too hard for most people to read.The results thus indicate that the news releases of the two dominant political parties in Ghana are apparently beyond standard.
By the definition of the FRE and FKGL, it is deduced that these news releases discussed in this study are difficult to read because there is high proportion of polysyllabic words as well as long sentences.Such a conclusion will resonate with the findings of Adika (2012) and Sey (1973) who have asserted that educated Ghanaians often write with complex grammatical structures to the detriment of the readers, who may have difficulty comprehending or getting the import of the text.
Table 2 presents a one sample T-test to compare readability of the news releases to the "standard" level recommended elsewhere (FRE = 8; FKGL = 60) (Cutts, 2013) using the FRE and FKGL indexes.The one sample T-test conducted to compare readability NPP and NDC to standard levels found significant difference between these news releases and the standard levels The effect size for FRE was found to be "large" for both NPP (effect size = 0.79) and NDC (0.90).Similar findings were found for the FKGL scores.These findings corroborate what has been found by earlier researchers (Warren and Morton, 1991).
The findings from the one sample t-test confirmed that the readability of news releases by the two major political parties in Ghana were above the standard level for most of the populace to read.It implies that, the masses will find it difficult to be able to read and comprehend these news releases.Thus, the purpose of the news releases (that is to shape the masses thoughts) may not be realised since they are not likely to be able to read and comprehend the news releases.
The likelihood that the masses will understand, even in the event that the mass media publishes these releases, is doubtful.To put these findings into perspective, a comparison of these findings with recommendations in countries with higher rate of literacy will help.For example, In the UK and US, where majority of the populace are educated, it is indicated that the reading level of majority is at 8th grade (Cutts, 2013).
In Ghana, the education demography shows that less number of people are educated.Most probably then, the reading level of majority will fall below 8th grade.Hence, for these political parties to pitch their news releases on an average of around 13th graders defeat the purpose for which they are released.This is especially serious since majority of the so called "grassroots" sympathisers of each of these political parties are uneducated or very less educated individuals.
As a result, it is likely that majority of the grassroots will fall on commentaries by radio analyst of political talk shows to understand what is put in the public domain by these parties.Unfortunately, several of the political commentators, even so called moderators, of political talk shows have their own political leanings.Hence, the intent of news releases by any of these parties could easily be misconstrued and carried to the public domain.Interestingly, not all political hosts and commentators on radio have an advance schooling.This can further jeopardise the purpose for the news releases.
Research question 2: Is there statistically significant difference in the readability of news releases of Ghanaian political parties?
An independent sample T-test was conducted to compare readability of the news releases of the two political parties (Table 3).Preliminary analysis was performed to check for the assumptions of normality and equality of variance.Levene"s test of equality of variance was found to be insignificant for both the FRE and FKGL scores.Hence, equality of variance was assumed.The results showed that there was no significant difference in readability between NPP and NDC news releases.This implies that both NPP and NDC news releases were equally difficult to read.Fosu (2016) drew a similar conclusion in his study of editorials of NPP and NDC newspapers in Ghana.
This finding seems to suggest that the readability of the news releases is dependent probably on the authors and not the political affiliation per se.In other words, either both parties use non-professional public relations personnel or that the public relations personnel are themselves inconsiderate of their writings.The writings seem to reflect the general attitude of Ghanaians to write using flamboyant words to impress.

Conclusion
Findings from this study reveal that news releases by both the NPP and NDC are "difficult" to read.The average schooling required to be able to read was above standard age recommended for documents which have the whole populace as its target audience.The magnitude of the difference in readability of the news releases of both parties compared to standard recommended reading levels was large.
The readability and use of polysyllabic words of these news releases did not differ between political lines suggesting that the authors of these releases from both parties might have attained the same level and kind of education.If the same instructor groomed these, it might explain how come both are writing at the same level of difficulty.The implication to these conclusions is that a large number of grassroots sympathizers of both parties may not be able to read and understand the news releases because the news releases were written far above the reading comprehension level for the general public.Therefore, effective communication between the political parties and the masses are likely to be hampered.

CONFLICT OF INTERESTS
The author has not declared any conflicts of interests.

Table 1 .
Descriptive statistics of readability scores of news releases of NPP and NDC.

Table 2 .
One sample T-test for readability of news releases of NPP and NDC.

Table 3 .
Bootstrap for Independent Samples Test of readability of NPP and NDC news releases.
Appendix A. FRE and FKGL raw scores.