Corporate Communications: An International Journal Impression management and self-presentation dissimulation in Portuguese chairman's statements

Purpose – Drawn on social psychology theory of impression management, the present study tries to assess the way Portuguese managers build their narratives in Chairman’s Statement to manage stakeholders’ perceptions on corporate image, in a period of time of scarce resources. Design/methodology/approach – The paper’s theoretical framework draws on elements of social psychology theory of impression management developed by Leary and Kowalski (1990). Through the use of the two-component model of impression management (impression motivation and impression construction) the 45 Chairmen’s Statements of Portuguese non-finance companies were content analysed to understand how managers build their voluntary communication strategies. Findings – Results indicate that organizational outcome does not influence the adoption of impression management strategies. But public visibility and consumer proximity are crucial factors in explaining them. Larger companies with high consumer proximity present themselves in a favourable way, but consistent with an overall reading of the annual report. These companies show a higher level of verbosity, consistent to the argument of retrospective rationality. Originality – The present study goes beyond Merkl-Davies et al. (2011) work and obtains insightful knowledge on the influence of goal-relevance of impression in three different perspectives: company’s public visibility, company’s dependency from debtholders, and consumer proximity. Moreover, the analysis uses a period of scarce resources and a European Latin country, with no tradition in publishing Chairmen’s Statements, but that recently has changed its financial reporting practices from an institutional code-law logic to an institutional common-law logic. A research setting like this has not been studied hitherto.


Introduction
Impression management (IM) has been defined as the "attempt to control images that are projected in real or imagined social interactions" (Schlenkler, 1980, p.6). It is the "field of study within social psychology studying how individuals present themselves to others to be perceived favourable by others" (Hooghiemstra, 2000, p.6).
In financial reporting the adoption of impression management strategies embrace attempts "to control and manipulate the impression conveyed to users of accounting information" (Clatworthy and Jones, 2001, p.311). Literature on financial reporting quality has shown that these strategies may take the form of subliminal verbal/not verbal messages to manipulate the content and presentation of financial information. (Hooghiemstra, 2000;Merkl-Davies and Brennan, 2007).
The importance of this argument increases in the presence of: a) discretionary financial information, disclosed voluntarily; b) proximity of financial information to the auditor's report; c) recent financial scandals and the global financial crisis (GFC) (Neu et al., 1998;Merkl-Davies and Brennan, 2007;Ball, 2009).
From an economic perspective the relationship between IM strategies and negative organisational outcomes has been explained as a consequence of the managers' opportunistic behaviour to dissimulate investors' perceptions on company's performance (Merkl-Davies and Brennan, 2007). Drawn on a social psychology perspective, some studies have found that the adoption of IM strategies may be either a consequence of managers' opportunistic behaviour to consciously dissimulate corporate image, or a result of informational processes, through which managers engage in retrospective sense-making strategies by framing organizational outcomes, albeit in a favourable way (Aerts, 2001(Aerts, , 2005Merkl-Davies et al., 2011). In organizational outcomes, the concept of sense-making is the process by which people give meaning to experience. One function of sense-making is retrospection (the point in which time affects what people notice). Thus, for Weick (1993, p.635) "reality is an ongoing accomplishment that emerges from efforts to create order and make retrospective sense of what occurs". From a financial reporting context, retrospective sense-making is a description of chronological actions, facts, and events (retrospective framing) in order that they make sense in relation to one another and contextualize organizational outcomes (Merkl-Davies et al., 2011). Merkl-Davies et al. (2011) have explored two factors motivating IM strategies by UK listed companies: goal-relevance of impression (assessed by company's size) and value of desired goals (assessed by organisational outcomes). Drawn on social psychology theory of impression management, the present study tries to assess the way Portuguese managers build their narratives in Chairman's Statement (CS) to manage stakeholders' perceptions on corporate image. More precisely, it tries to understand how value desired goals (measured by negative organisational outcomes) influence the adoption of IM strategies to dissimulate stakeholders' perceptions on corporate performance. Moreover, the present study intends to go beyond Merkl-Davies et al. (2011) work and obtain insightful knowledge on the influence of goal-relevance of impression in two different perspectives: company's public visibility (assessed by company's size) and the relevance of stakeholders (assessed by company's dependency from debtholders and consumer proximity).
The CS included in consolidated annual reports for 2010 of Portuguese nonfinance companies were analysed. The setting of 2010 is interesting because it is a timeframing period not studied hitherto: it is a post-GFC period, and it corresponds to a period of time in which Portugal has been affected by a sovereign crisis. The recent financial scandals and the GFC have aware users for the need of higher quality in financial information (Ball, 2009). During these events managers may have engaged in IM strategies, because these events were related to periods of scarce resources, corporate image losses, which may trigger the adoption of IM strategies (Leary and Kowalsky, 1990). To restore corporate image companies can engage in IM strategies in financial reporting to manage stakeholders' perceptions on companies' performance, and attract vital resources to the viability of their business (Carter and Dukerich, 1998). Therefore, this setting will allow us to examine how managers used IM strategies to manage corporate image during this period of scarce resources.
The CS was selected because: a) it is widely read by investors (Courtis, 2004a;4 decision-making process and firm's value (Kaplan et al., 1990;Segars and Kohut, 2001); e) discloses elements of a CEO "mindset(s), aspirations, ideologies and strategic thinking" (Armenic and Craig, 2007, p. 26); f) often personifies the culture and personality of the company (Oliver, 2000;Costa et al., 2013); g) and are generally unaudited (Clatworhty and Jones, 2006). Thus, the potential to include IM strategies is huge.
Previous literature has focused on Anglo-Saxon companies. However, some studies have concluded on the influence of cultural aspects in managers motivations to engage in IM strategies (Merkl-Davies and Brennan, 2007). In Anglo-Saxon countries the financial reporting model is oriented toward full disclosure and transparency, and is focused on investors protection rights. By contrast, in Code-law countries (such as Portugal) financial reporting model is oriented toward legal compliance, with low disclosure levels, and is aimed at creditor protection (Ball et al., 2000;Meek and Thomas, 2004;Lopes and Rodrigues, 2007). Moreover, European Latin countries (such as Portugal) have little tradition in publishing a CS, basically because their corporate regulatory regime follows an institutional logic of code law. However, since 2005 there is been a progressive increase in its publication (Costa et al., 2013), basically due to a requirement in changing to an institutional logic of common law after the adoption of International Accounting Standards Board (IASB) standards (Costa et al., 2013;Guerreiro et al., 2012). Therefore, the present study intends to assess if the motivations of Portuguese managers in engaging in IM strategies are different from those managers in Anglo-Saxon companies.
Findings indicate that Portuguese companies adopt less IM strategies than UK listed companies. Organisational outcome does not impact on managers' motivations in adopting IM strategies. According to social psychology theory of impression management managers from larger companies adopt IM strategies to present themselves to others in a more favourable way, albeit consistent with an overall reading of the annual report. CS is commonly used to engage in retrospective sense-making by means of retrospective framing of organisational outcomes.
The remainder of the paper is as follows: first we document the literature review and develop hypotheses. Thereafter, we describe our research method, report results, and present conclusions.

Literature review and hypotheses development
Merkl-Davies and Brennan (2007) and Brennan et al. (2009) present an extensive literature review on IM strategies and identify seven discretionary narrative disclosure strategies carried out by disclosure choices and the presentation of information, by means of bias (conveying information in a very positive/negative way) and selectivity (omission or inclusion of some items of information). Table 1 show the definition for each discretionary narrative disclosure strategy.
(Insert table 1 about here) IM takes into consideration managers' rational opportunistic behavior to benefit from them by exploiting information asymmetries (Bowen et al., 2005). According to Table 2, prior literature has focused on the relationship between organizational outcomes and IM strategies included in the narrative sections of CS underpinned either on economic theories such as agency theory (Smith and Taffler, 1992;Abrahamson and Park, 1994;Courtis, 1995Courtis, , 1998) and signaling theory (Smith and Taffler, 1992), or on social psychological perspectives of attribution theory (Clatworthy and Jones, 2003).
(Insert Table 2 here) Merkl-Davies and Brennan (2007) acknowledge the existence of a wide and promising field of research on the motivations for IM through the use of social/political theories such as legitimacy theory (legitimacy/reputation threats), stakeholders theory (communicational process with relevant stakeholders), and institutional theory.
Literature on reputation risk management (Bebbington et al., 2008) and social psychology (Leary and Kowalski, 1990) also constitute promising and alternative sources for further studies on the motivations for IM.
Prior literature drawn on social psychology contends that IM entails selfpresentational dissimulation behaviours (Leary and Kowalsky, 1990) and enhancement behaviours (Merkl-Davies et al., 2011). Self-presentational dissimulation behaviours incorporate constructing inaccurate impressions of organisational outcomes, such as the obfuscation of negative organisational outcomes. They are control-protective and selfserving because they build "impressions at variance with an overall reading of the [annual] report" to maintain established public images or reputation that have been threatened (Stanton et al., 2004, p.57). Enhancement behaviours refer to the creation of accurate, albeit favourable, impressions of organisational outcomes. They are proactive, because they are "designed to enhance corporation's image (...) to build an image of the corporation that ingratiates it with its stakeholders" (Stanton et al., 2004, p.57).
Alternatively, Aerts (2005) contends that managers may use financial reporting narratives to retrospectively framing organisational outcomes (retrospective sensemaking behaviours). Merkl-Davies et al. (2011) have followed a social psychology perspective to explain motivations for IM in CS of UK listed companies. Drawn on social psychology theory of IM the present study analyses the adoption of IM strategies in CS of Portuguese non-finance companies, with no prior tradition in publishing CS, but that recently has changed its financial reporting practices from a code-law institutional logic to a common-law institutional logic (Guerreiro et al., 2012) and in which there is a requirement to audit any numbers included voluntarily in the management report, such as the CS. These particular aspects of Portuguese reality highlight the importance to In social psychology IM is a social concept, because it involves "the process by which people attempt to control the impressions others form of them" (Leary and Kowalshy, 1990, p.34). The concept of self is crucial as the images people have of themselves shape and are shaped by social interactions. The effectiveness of any IM strategy is determined by the social presence of outsider parties (relevant stakeholders) interacting each other through an accountability process (Leary and Kowalsky, 1990). Thus, IM entails an economic "rationale for corporate communication, which is controlled and managed and is influential and persuasive" incorporating psychological nuances (Stanton et al., 2004, p.58). Leary and Kowalski (1990) proposed a two-component model to explain IM: impression motivation and impression construction. Impression motivation takes into consideration the circumstances that determine the adoption of a specific IM strategy.
To achieve desired social and material outcomes, maintain self-esteem, and develop desired identities impression motivation is influenced by three factors: a) the value of the desired goals; b) the goal-relevance of impressions; c) and the existing discrepancy between the individual's current image and the image he wants to convey. Leary and Kowalsky (1990) contend that there is a positive association between the value of the desired goals and the motivation for IM. Consequently, "impression motivation is higher when desired resources are scarce", because the value of desired goals increases when resources availability goes down (Leary and Kowalsky, 1990, p.38). In the context of financial reporting, negative organisational outcomes can financial markets began to become suspicious about the ability of the country to fulfill its sovereign debt liabilities, risk premiums increased up to a point where access to capital markets was no longer an option and a debt default soon became imminent. Impression construction involves "not only choosing the kind of impression to create, but deciding precisely how they will go about doing so (such as deciding whether to create the desired impression via self-description, non-verbal behaviour, or props)" (Leary and Kowalski, 1990, pp.35-36). In financial reporting, impression construction, involves the creation of public images about certain actions or events consistent or not with management's self-concept of those facts or events.
Social psychology tries to "understand and explain how the thoughts, feelings and behaviors of individuals are influenced by the actual imagined or implied presence of other human beings" (Allport, 1954, p. 5). Verbal language is one psychological marker for human behavior, because it is "based on the assumption that the words people use convey psychological information over and above their literal meaning and independent of their semantic context" (Pennebaker et al., 2003, p. 550). Merkl-Davies et al. (2011), followed Newman's et al. (2003 structure to analyse the level of self-presentational dissimulation using six markers for verbal language, consistent with three IM strategies: rhetoric manipulation (number of words, first person pronouns, third person pronouns); thematic manipulation (positive emotion words and negative emotion words); and readability manipulation (words relating to underlying complex cognitive processes). The present study aims to assess how the value of the desired goals (assessed by negative organizational outcomes) influence the usage of eight psychological markers for self-presentational dissimulation behavior, consistent with four IM strategies: rhetoric manipulation (number of words, selfreferences, and references to others); readability manipulation (readability); thematic manipulation (positive and negative words); and choice of earnings number (qualitative references to financial performance indicators and quantitative references to financial performance indicators) (Clatworthy and Jones, 2006;Merkl-Davies et al., 2011;).

Number of words
Prior findings on the association between CS's verbosity and organizational outcomes were inconclusive. Segars and Kohut (2001) found that well-performing companies show higher levels of verbosity compared to poor-performing companies. Other studies did not find any association (Clatworthy and Jones, 2006;Merkl-Davies et al., 2011).
From a social psychological perspective it is expected that poor-performing companies would be less verbose than well-performing companies. Negative organizational outcomes promote self-presentational dissimulation behaviours. Consequently, narratives are shorter, because "lying is associated with fewer details, thus resulting in shorter communication" (Merkl-Davies et al., 2011, p. 323).
Hypothesis 1: Companies with negative organisational outcomes present fewer words in CS, than companies with positive organisational outcomes.

Self-references
Prior literature on the relationship between self-references and organizational outcomes have been inconclusive. Some studies have found positive associations (Thomas, 1997;Clatworthy and Jones, 2006) and an extensive use of first person pronouns (Hyland, 1998). Merkl-Davies et al. (2011) did not find any association between them.
Aristotle has proposed three persuasive audience appeals: ethos, pathos and logos. Ethos appeals to the authority and honesty of the speaker. Audience is more likely to be persuaded by a credible source because they are more reliable. Pathos relates to emotional appeals to alter the audience's judgment through metaphors, amplification, storytelling evoking strong emotion in the audience to motivate and persuade decision making. Logos use inductive or deductive reasoning to construct an argument including statistics, math, logic and objectivity (Kennedy, 1991). Ethos construction in CS is crucial because to assure a credible communication between management and audience, the message should be able to create an image of integrity, authority, and honesty of the management. Ethos construction can be accomplished through the use of self-references (e.g. first person pronouns). Self-references express personal beliefs, strength the writer's presence in the narrative, and align the writer with the message, creating a feeling of competence, responsibility, and authority (Hyland, 1998). Hyland (1998, pp. 235, 237) argues that "credibility [of the message] is obviously most easily gained on the strength on company (…). [However, the usage of self-references] may not always be appropriate, particularly in years when the company has not performed as hoped". Under these circumstances managers have to be more modest and conservative, in order to manage their reputations and maintain investors' confidence. Newman et al. (2003) indicate that narratives with self-presentational dissimulation strategies have fewer self-references.
Hypothesis 2: Companies with negative organisational outcomes use fewer self-references in CS, than companies with positive organisational outcomes.

References to others
Merkl- Davies et al. (2011) found that there is not any association between references to others and organisational outcomes, even after controlling results by size and industry.
Nevertheless, consistent with Newman et al. (2003) narratives with self-presentational dissimulation strategies present a fewer number of references to others.
Hypothesis 3: Companies with negative organisational outcomes use fewer references to others in CS, than companies with positive organisational outcomes.

Positive and negative words
In social psychology, verbal language conveys information about emotions, anxieties and frustrations. Consistent with the Pollyana hypothesis, managers tend to present themselves and company performance in the most favourable way, regardless of companies' organizational outcomes. (Hildebrandt and Snyder, 1981). However, Newman et al. (2003) argues that towards negative organisational outcomes managers adopt self-presentational dissimulation behaviours through a higher use of positive words (which is consistent with Pollyanna's hypothesis), and a fewer use of negative words arising from discomfort, guilt and frustration for the accomplished outcomes.

Readability
Consistent with agency theory, readability manipulation is intrinsically linked to the hypothesis of obfuscation of bad news (Courtis, 1998). But within social psychology readability can be related to other factors. Aerts (2001, p.7) has referred that attribution defensive behaviours in companies with negative organizational outcomes main entail "accounting bias, a tendency to explain negative performance results more in technical accounting terms, relating intermediary accounting effects, while positive performances are more accounted for in explicit cause-effect terminology whereby internal or external factors, not related to the financial reporting framework, are identified as causal influences". Negative organisational outcomes will be more salient to relevant stakeholders and expose management to their scrutiny. Therefore, management need to present more accounting explanations to contextualize those negative outcomes and legitimate themselves before stakeholders (Aerts, 2005). If the constructed public impression is consistent with management's self-concept of organizational actions and events, the referred contextualization will demand a higher detail, description and explanation. Therefore, like Aerts (2005) and Bloomfield (2008) argue, the level of readability of a narrative may be related to an informational process, rather than to the hypothesis of obfuscation of bad news. Contextualization of negative organisational outcomes can promote syntactical complexity, due to the inclusion of technical explanations containing technical terminology and complex syntactical structures.
On the other hand, in the presence of negative organisational outcomes if management adopt self-presentational dissimulation behaviours narratives will be easier to read (Newman et al., 2003). Narratives will contain fewer cognitive complexities: simpler sentence structure; fewer words of causality; fewer words demanding reflection.
Such an argument is consistent with the idea that "liars tell less complex stories" (Merkl-Davies et al., 2011, p. 322).
Hypothesis 5: Companies with negative organisational outcomes present more readable CS than companies with positive organisational outcomes.

Qualitative/Quantitative references to financial performance indicators
Prior literature found that companies with positive organisational outcomes disclose more information on performance indicators (Beattie and Jones, 2000; Clatworthy and Jones, 2006). On the other hand, it is expectable that in companies with negative organisational outcomes if managers adopt self-presentational dissimulation strategies they will report less information on performance indicators. Skinner (1994) found that good news were disclosed quantitatively, whereas bad news were disclosed qualitatively. impressions Leary and Kowalski (1990) contend that the factors that determine how relevant one's impressions are to fulfil the goals of social and material outcomes, self-esteem maintenance, and identity development are: company's public visibility and the relevance of stakeholders. Public visibility is a "function of both probability that one's behaviour will be observed by others who might see and learn about it" (Leary and Kowalski, 1990, p. 38). The more public visible companies are more incentives will managers have to adopt IM strategies because they will likely be scrutinized by their relevant stakeholders. Moreover, the more companies depend of some relevant stakeholders to access crucial resources, more incentives managers will have to engage in IM strategies. Therefore, larger companies depending upon relevant stakeholders to survive, have more incentives to engage in IM strategies.

Data collection
All CS were written in Portuguese and were content analysed to assess IM strategies adopted. Content analysis was performed manually, based on the methodology used by Clatworthy and Jones (2006) and Merkl-Davies et al. (2011), and followed a two-stage procedure. Firstly, an initial sample of CS was analysed by two independent coders. The prior coding helped establishing a set of dictionaries functioning as decision rules. To assure the reliability of the content analysis Scott's pi measure of the inter-rater reliability was used (Scott's pi=0.85) and its level was considered acceptable in the analysis of corporate disclosures (Lombard et al., 2002).
Secondly, the entire sample was content analysed by the first author. Table 3  rhetoric manipulation (number of words, self-references, and references to others), thematic manipulation (positive and negative words), readability manipulation (readability), and choice of earnings number (qualitative references to financial performance indicators and quantitative references to financial performance indicators).
(Insert Table 3 here) The variable "number of words" was measured by counting the number of words in CS (Clatworthy and Jones, 2006;Merkl-Davies et al., 2011).
The variable "self-references" was measured by the proportion of first person pronouns, references to the Group, and reference to company's name in CS (Clatworthy and Jones, 2006;Merkl-Davies et al., 2011). The variable "readability" was measured by the computation of the Flesh reading ease index (Aerts, 2005;Bloomfield, 2008). The readability analysis has its origins in social psychology and commonly uses the syntactical structure of narratives, in terms of sentence length and the number of syllables. There are several algorithms to measure readability using these two metrics: Fog, Flesch, Kwolek, Date-Chall, Lix, Fry, Cloze, Texture index, and Diction (Brennan et al., 2009). Flesh reading-ease index assesses the readability level of narratives counting the variables within a narrative and has been widely used in prior literature of IM due to its easy calculation through the use of a computer and its easy interpretation and comparison with other studies (Courtis, 1998;Linsley and Lawrence, 2007). The results from the Flesch reading ease index formula varies between the value "0" (indicating a low level of readability) and the value "100" (indicating a high level of readability).
Although Flesch reading ease formula has been developed for English language Cavique (2008, p. 62) argues that "since it does not depend on a dictionary it can be perfectly used on Portuguese language". Campbell et al. (2005) have concluded that comparative volumetric content analysis should be carried out by interrogating narratives in the same language. The present study has content analysed CS that were exclusively written in Portuguese language. However, for comparison purposes with prior Anglo-saxon literature, according to Campbell et al. (2005)  The variable "qualitative references to financial performance indicators" was measured by the number of qualitative references to financial performance indicators in CS (Clatworthy and Jones, 2006).
The variable "quantitative references to financial performance indicators" was measured by the number of quantitative (absolute or percent value) references to financial performance indicators in CS (Clatworthy and Jones, 2006). Impression motivation was measured by three variables that try to assess the value of desired goals (assessed by organisational outcomes) and the goal-relevance of impressions (assessed by company's public visibility and relevance of stakeholders) (Table 3, panel B). All of these variables were extracted from the companies' consolidated annual reports.
The variable "organisational outcome" was measured by a dummy variable assigning 1 if earnings before tax growth rate > 1, and 0 otherwise (Merkl-Davies et al., 2011).
The variable "public visibility" was proxied by company's size and company's listing status (Branco and Rodrigues, 2006, Oliveira et al., 2011a, 2011b. "Company's size" was measure by total assets (Branco and Rodrigues, 2008). "Company's listing status" was measured by a dummy variable assigning 1 if the company was listed on one or more regulated stock exchange markets, and 0 otherwise (Branco and Rodrigues, 2006;Oliveira et al., 2011a).
In Portugal banks dominate as a source of financing and financial reporting is aimed at creditor protection (Lopes and Rodrigues, 2007). Moreover, the more contact the company has with its stakeholders more motivated managers will be to try to control stakeholders' perception about them (Leary and Kowalski, 1990). Therefore, the variable "relevance of stakeholders" was proxied by company's dependence on debtholders and consumer proximity. "Company's dependence on debtholders" was measured by leverage (ratio of total debt to total assets). "Consumer proximity" was measured by a dummy variable assigning 1 if the company belongs to a high consumer proximity industry (consumer goods, consumer services, utilities, telecommunications, and oil and gas), and 0 otherwise (Branco and Rodrigues, 2008). Table 4 presents the descriptive statistics results. On average CS contain 1,117.84 words. Clatworthy and Jones (2006) and Merkl-Davies et al. (2011) found lower levels among UK listed companies. Portuguese companies do not use self-references (mean value=0.04), nor references to others (mean value=0.02) as IM strategies. These mean values are lower than those found among UK listed companies (Clatworthy and Jones, 2006;Merkl-Davies et al., 2011). These companies follow a financial reporting model focused in the capital markets and are held by a wide range of investors, whereas in Portugal companies are held by families (Lopes and Rodrigues, 2007). Thus, it was expectable to find lower information asymmetries among Portuguese companies and therefore lower levels of IM strategies adopted. Table 4 shows that on average companies use more positive words (mean value=0.07) than negative words (mean value=0.02). These findings are consistent with Merkl-Davies et al. (2011). The authors found that UK listed companies used IM strategies not as self-presentational dissimulation instruments. But present themselves in the most favourable way.

Descriptive analysis
(Insert Table 4 about here) The level of readability of CS is very low (mean value=6.46). The maximum value is around 25.9 (Table 4). This value is below the threshold of 30 established by the Flesh reading easy index ranking which corresponds to a very hard level of reading. This finding indicates that organisational outcomes may not be associated with the readability of CS, but with issues related to the informational process (Aerts, 2001(Aerts, , 2005. Portuguese companies disclose an average of 6.51 qualitative references to financial performance indicators. Quantitative references were monetary (mean value = 2.08) and percentage (mean value=2.43). Financial performance indicators most disclosed were: earnings (31 companies); sales (26 companies); profitability (14 companies); dividends per share (11 companies); leverage (8 companies); and cash-flow (6 companies). There were 4 companies that did not disclose any performance indicator. Clatworthy and Jones (2006) findings were quite different. Performance indicators most disclosed by UK listed companies were profit before taxes and dividends.
Leverage among Portuguese companies is high, indicating their strong dependence on debtholders. This is explained by the fact that in Portugal banks are the main sources of financing (Lopes and Rodrigues, 2007).

Hypotheses testing
The dependent and independent variables were tested for skewness and normal distribution. The variables "company's size" and "company's dependence on debtholders" were transformed through the computation of its natural logarithm to reduce skewness. The Kolmogorov-Smirnov Lilliefors test has shown that variables were not distributed normally. Consequently, hypotheses testing were performed using the following statistical tests: Mann-Whitney U test for two independent samples, Kruskal-Wallis test for k independent samples, and Spearman correlation.
Table 5 results indicate that IM strategies, such as rhetoric manipulation, thematic manipulation, readability manipulation, and choice of earnings number are not significantly different between companies with positive and negative organizational outcomes. Results do not support hypothesis H1 (number of words), H2 (selfreferences), H3 (references to others), H4a (positive words), H4b (negative words), H5 (readability), H6a (qualitative references to financial performance indicators) and H6b (quantitative references to financial performance indicators).
(Insert Table 5 about here) The IM strategies were not adopted to build a public image about organisational actions and events at variance with an overall reading of the annual report. Consistent with Merkl-Davies's et al. (2011) findings the IM strategies adopted by Portuguese nonfinance companies are not consistent with self-presentational dissimulation behaviours.
Negative organisational outcomes are not a trigger for self-presentational behaviour. Table 6 shows that there is a positive correlation, statistically significant, between size and the number of words (p-value<0.01), and the proportion of selfreferences in CS (p-value<0.05). And there is a positive correlation, statistically significant, between the company's listing status and the proportion of positive words in CS (p-value<0.05). These findings support H7a. Larger companies present higher levels of verbosity and self-references in the CS. Moreover, companies more publicly visible related to their listing status profile present a higher proportion of positive words in the CS. Results also document a negative correlation, statistically significant, between company's size and the proportion of negative words (p-value<0.05). This finding does not support H7a.
Note that the adoption of positive words is not significantly different between companies with positive and negative organizational outcomes (Table 5). Authors have also tested and verified size of sampled companies is not significantly different considering their listing status and organizational outcomes [1]. Merkl-Davies et al.
(2011) contend that managers can emphasise positive organizational outcomes regardless of their financial performance. Therefore, consistent with social psychology theory of IM publicly visible Portuguese non finance companies use CS to communicate a more favourable image (but consistent with managers self-concept of organisational actions and events) of themselves either through the use of more positive words or fewer negative words.
(Insert Table 6 here) Table 6 shows that there is not a statistically significant correlation between the several IM strategies and company's dependence on debtholders. This result does not support H7b. Company's dependence on debtholders is not relevant in choosing IM strategies. However, findings indicate a positive correlation, statistically significant, between consumer proximity and the number of words in the CS (p-value<0.01). This result support H7b. Authors have tested and verified that companies with higher consumer proximity were significantly larger than companies with lower consumer proximity [2]. Therefore, larger companies with higher consumer proximity present a higher level of words in CS.
Previous findings indicated that there are not significant differences in the number of words in the CS and its level of readability between companies with positive and negative organisational outcomes (Table 5), which is consistent with Aerts (2005) arguments. The levels of verbosity and readability are related to organisations' informational processes. Findings of Table 6 are also consistent with this argument.
Larger companies present higher verbosity because larger companies are more complex organisations, more exposed to business risks, and subject to a deeper scrutiny from relevant stakeholders. Therefore, managers need to present more detailed information to contextualize the description of organisational actions, facts, and events, which demands a higher verbosity. Managers "engage in retrospective sense-making by means of drawing together a series of events in order that they make sense in relation to one another (…) [manifesting] itself linguistically in the form of a more complex grammatical sentence structure and more causation and insight words" with consequences on the readability and complexity of the narrative (Merkl-Davies et al., 2011, p. 336). Our results indicate that larger companies engage in retrospective sensemaking through the use of more self-references. Self-references help them to build ethos, strengthening the manager's presence in the narrative, aligning the manager with the reader, and creating in the stakeholders mind an image of competency, honesty, integrity, responsibility and authority of the management (Hyland, 1998).
Non-parametric hypotheses tests were re-run taking into consideration the variables "consumer proximity" and "company's listing status". Results demonstrate that findings were not driven by these variables (Table 7).

Conclusions
Our results confirm that Portuguese non finance companies do not adopt IM strategies to present a self-presentational dissimulated impression. The levels of organizational outcomes (measured by earnings before tax growth rate) do not influence the adoption of IM strategies. that the publication of CS is increasing. However, its content does not follow a specific set of rules. But, the narratives in a CS can "lead people into the future" (Armenic and Craig, 2007, p. 25). It is important that users adopt a sceptical attitude towards the tone and content of CS. Thus, consistent with Falschlunger et al. (2015), we believe that to safeguard stakeholders' interests, the release of guidelines on the content and tone of CS narratives should be part of the international standard setters (such as IASB) agenda.
Revisiting Clatworthy and Jones (2006) this agenda should focused on the clarification CS (in which managers take responsibility for it and alert readers that the message is the Chairman's personal point of view) is a potential mechanism to achieve that desire.
Some limitations that should be noted are related to the subjectivity of content analysis.
The present study analyses only the information content of CS. Further research should consider the adoption of IM strategies in other sources of financial information such as press-releases, web-sites or other sections of annual reports (e.g. risk sections, relevant events section), and should analyse the influence of the GFC on the adoption of IM strategies.
1. Detailed analysis is available upon request from the authors 2. Detailed analysis is available upon request from the authors  (Courtis, 2004a, p. 292).

Rhetorical manipulation
To obfuscate bad news managers use persuasive language to constantly distort the narrative discourse in one or more ethical principles, such as: clarity, truthfulness, sincerity, and legitimacy (Yuthas et al., 2002).

Thematic manipulation
To conceal bad news managers do not report them or opt to present themselves in the most favourable way by emphasising good news or through the use of more positive words/themes, rather than negative words/themes (Merkl-Davies and Brennan, 2007;Merkl-Davies et al., 2011). Visual and structural manipulation To conceal bad news or to emphasise good news managers present information in different ways using visual emphasis, repetition, reinforcement of ideas, ordering, and location of information to create noise, emphasise ideas or direct readers attention (Merkl-Davies and Brennan, 2007). Performance comparisons To emphasise good news managers use selectivity by choosing benchmark earnings numbers and performance referents (Merkl-Davies and Brennan, 2007).

Choice of earnings number
To emphasise good news managers use selectivity by choosing specific earnings number and omitting others (Merkl-Davies and Brennan, 2007).

Performance attribution
Managers act in a self-serving manner, attributing positive organizational outcomes to themselves (entitlements) and negative organisational outcomes to external facts (excuses) (Merkl-Davies and Brennan, 2007;Aerts, 2001Aerts, , 2005.   Definition of variables: readability = Flesh reading ease index; size = total assets; company's dependence on debtholders = total liabilities/total assets; organizational outcome = 1 if company has an earnings before tax growth rate > 1; 0, otherwise; consumer proximity = 1 if company belongs to an industry with a high consumer proximity; 0, otherwise; company's listing status = 1 if company is listed in one or more stock exchange regulated markets; 0, otherwise.