Presentation on theme: "WHAT MAKES PEOPLE HAPPY?"— Presentation transcript:
1 WHAT MAKES PEOPLE HAPPY? Social Support & Emotional Intelligence as Predictors of Subjective WellbeingEmma Gallagher & Dianne Vella-BrodrickSuggested amendments are indicated in bold and underlined…………A presentation by Emma Gallagher, based on a recently completed thesis submitted as part of the Honours Degree in Psychology at Monash University, 2005.Thank you to Dr Dianne Vella-Brodrick, Monash University, for her supervision of the thesis.It is anticipated that the results of this study will be published in 2007.By the end of this presentation you will know the relative contributions ofsocial support and emotional intelligence as predictors of SWB.NB: For reviewers…please note that on some slides there are substantial notes (this section of the screen) and if the slides are printed with notes not all the information will be printed out, thus you will need to read from the screen for some slides. Thank you.
2 What we already know…Happiness is operationalised by Subjective wellbeing (SWB).SWB studies highlight factors that foster optimal psychological functioning.Three components of SWB: Positive Affect (PA), Negative Affect (NA), Satisfaction with Life (SWL).The main idea behind the study was to look at the gaps in the SWB research and try and explain variance beyond what we already know. In other words, what predicts SWB? So, what do we know?We know that happiness is operationalised in psychological research as SWB.Ideally, three components of SWB are measured to gain a compressive assessment of SWB.These components are: Positive Affect, Negative Affect and Satisfaction with Life.(Biswas-Diener Diener, & Tamir, 2004; Diener, Suh, Lucas, & Smith, 1999; Lucas, Diener, & Suh, 1996; Ryan & Deci, 2001).It has been found that people with both high positive affect and satisfaction with life and low negative affect have higher SWB, or are happier than people with the inverse (Keyes, Shmotkin, & Ryff, 2002; Myers & Diener, 1995; Ryan & Deci, 2001).
3 Sociodemographic variables & personality factors Sociodemographic variables 8-20% (Argyle, 2001; Diener et al., 1999).Personality factors consistently large proportion of variance.e.g. 34% of unique variance in a general community sample (Gannon & Ranzijn, 2005).Two variables we know a lot about in SWB research are Sociodemographic variables and personality factors …Sociodemographics 8-20% (Argyle, 2001; Diener et al., 1999).Personality factors ~ 34% (eg. Gannon & Ranzjin, 2005).Together they account for a substantial amount of variance in SWB, but there is still approximately 50% to be explained.
4 Social support Social support (SS) relates positively to SWB. Some believe SS is necessary for SWB (Baumeister & Leary, 1995; Diener & Oishi, 2005, Diener & Seligman, 2002).SS – thought to promote well-being through the provision of resources from one person to another which influence emotions, cognitions, and behaviours that in turn help people cope and enjoy life.Social SupportSocial support (SS) is one of the variables that has received a lot of attention in the search for predictive value in the SWB equation.It is generally accepted that SS relates positively to SWB, and some researchers think that SS is necessary for higher SWB (Baumeister & Leary, 1995; Diener & Oishi, 2005, Diener & Seligman, 2002).SS is thought to promote well-being through the provision of resources from one person to another, which influences emotions, cogitations, and behaviours that in turn help people cope and enjoy life. SS is thought to promote and maintain positive affect, while the lack of SS leads to the failure to regulate emotion manifesting in negative affect (Cohen, 1988, Cohen, Gottlieb, & Underwood, 2000).There are a lot of papers showing positive relationships between SS and SWB, and a Meta-analysis of 556 papers shows that SS can account for between 1% and 8% of variance in SWB (Oku, Stock, Haring, & Witter, 1984).The most robust empirical evidence pertaining to the positive relationship between SS and SWB comes from studies that control for other well-known predictive variables of SWB. Controlling for other variables facilitates the illumination of SS’s unique predictive on SWB(e.g.. Argyle & Lu, 1990).SWB studies that control for other well-known variables show clearer results for the main variable of interest, and this was an important research consideration in the present study.
5 Social support considerations Need to measure SS and all three components of SWB.Source of SS (significant other, family, friends) over type of support (material aid).Perception of SS over number of supports and/or receipt of support.SWB + source & perception of SS + sociodemographic variables + personality factorsAnother research consideration was the need to measure all three components of SWB at the one time, as a number of previous studies do not cover all three components. As highlighted earlier, ideally positive affect, negative affect and satisfaction with life should be measured for a comprehensive assessment of SWB.Other considerations ascertained from the literature were that:Source of SS is more important than type of supportWhere source of support refers to significant others, family, friendsWhere type refers to material aid, information or emotional supportAnd perception of SS seems to be more important than Number of SSs or Received supportWhere perceived SS refers to the beliefs people hold regarding the level and quality of supports available to them. It is thought that perception of SS is most salient precisely because it is about how an individual thinks about the support available to them and if it can be called upon when needed.Although there is clearly a lot of evidence relating to SS and SWB, gaps in the research where identified.We found no published study that investigated all three components of SWB whilst simultaneously taking into account source and perception of support, whilst also controlling for sociodemographic variables and personality factors.During our literature review we also found another gap in the research that we were very eager to explore.This gap was regarding the variance in SWB that Emotional Intelligence could explain.
6 Emotional intelligence Linked to both SWB and SS (Bar-On, 2005; Salovey, Bedell, Detweiler, & Mayer, 1999).Relatively ignored in SWB research.Eminent SWB researchers have suggested EI is ‘worth investigating’ for variance in SWB.Investigations of SWB/EI relationship are relatively new.Emotional Intelligence (EI) as a construct has been theoretically linked to both SWB and SS, yet seems to have been relatively ignored in psychological research (Bar-On, 2005; Salovey, Bedell, Detweiler, & Mayer, 1999). There are not many papers investigating the relationship between SWB and EI. This is relatively surprising given the main stream popularity of EI, and further surprising because Ed Diener, one of the pioneers of SWB research, along with his colleagues, suggest that the relationship between SWB and EI is ‘worth investigating’ (Diener, et al., 1999). The literature review for this study unearthed a paper by Bar-On (2005) purporting to be the first investigation of SWB and EI, which suggests that this is a very new area of investigation. However, on further investigation this postulation is not entirely true, there are other papers published before 2005.
7 Emotional intelligence EI was labelled and modelled by Salovery and Mayer (1990).Popularised by Goleman (1995).Broadly defined as: the cognitive ability to perceive, manage, and regulate emotions within one’s self, and others, in ways that maximise positive cognitive and behavioural outcomes that result in more beneficial life outcomes (Bar-On, 2005; Mayer & Salovey, 1997; Salovey et al., 1999; Salovey & Mayer, 1990).EI can be broadly defined as the cognitive ability to perceive, manage, and regulate emotions within one’s self, while also recognising and responding to other peoples emotions. People with higher EI are thought to posses a higher capacity to accurately perceive and reason with, and about, emotions in ways that maximise positive cognitive and behavioural outcomes, the consequence being more beneficial life outcomes, such as SWB (Bar-On, 2005; Mayer & Salovey, 1997; Salovey et al., 1999; Salovey & Mayer, 1990).EI was labelled and modelled by Salovery and Mayer (1990).Popularised by Goleman (1995).Although EI in its current manifestation is new, it is widely acknowledged that the idea of EI is not new (see Bar-On, 2005; Goleman, 1995).
8 Emotional intelligence Emerging evidence that EI can be taught, opposed to other SWB predictors such as personality (Emmerling & Goleman, 2003; Reshmi, 2006; Slaski & Cartwright, 2002; Stys & Brown, 2004).Utility of EI complicated as more than one model has been proposed.Conjecture regarding the discriminant validity of EI – now sufficient evidence showing discriminant validity (Bar-On, 2005; Brackett & Mayer, 2003; Ciarrochi, Chan & Bajgar, 2001; Ciarrochi, Chan, Caputi, 2000; Ciarrochi, Deane & Anderson, 2002; Gannon & Ranzijn, 2005; Lopes et al, 2004; Schette et al., 1998; Tett, Fox & Wang, 2005).The attraction of using EI to predict life outcomes lies in the emerging evidence that EI can be taught and developed, as opposed to measures of individual difference that are resistant to change, such as personality which is believed to be stable (Emmerling & Goleman, 2003; Reshmi, 2006; Slaski & Cartwright, 2002; Stys & Brown, 2004).However, the utility of EI is relatively complicated because more than one model has been proposed, and there is some conjecture surrounding the discriminate validity of EI. That is, do measures of EI tell researches anything that they can not already find out by using other measures, especially personality.There is sufficient evidence showing that EI does possess enough discriminate validity for it to be investigated for its predictive value in the SWB equation (Bar-On, 2005; Brackett & Mayer, 2003; Ciarrochi, Chan & Bajgar, 2001; Ciarrochi, Chan, Caputi, 2000; Ciarrochi, Deane & Anderson, 2002; Gannon & Ranzijn, 2005; Lopes et al, 2004; Schette et al., 1998; Tett, Fox & Wang, 2005).
9 Emotional intelligence EI has been shown to relate positively to SWL and PA.SWL influenced by how clearly people understand emotion.Controlling for well-known predictors of SWB provide clearer results for EI.Research need: use well–respected SWB measures; measure 3 x components of SWB; and control for well-established predictors of SWB.Of the few studies that look at the relationship between SWB and EI, EI has been shown to significantly predicted SWL and positive affect (Austin, Saklofske & Egan, 2005; Cairrochi &Scott, 2006; Gannon & Ranzijn, 2005; Gignac, 2006; Martinez-pons, ; Palmer, Donaldson & Stough, 2002).Researchers suggest that SWL is influenced by how clearly people understand emotionClearer results for EI’s utility in predicting life outcomes is gained when other well-known predictors of SWB are controlled (eg. Ciarrochi et al., 2000; Gannon & Ranzijn 2005; Saklofske, Austin & Minski, 2003).A number of limitations from previous studies provided the opportunity to add to the existing research. It was identified that new research investigating the relationship between EI and SWB needed to use well–respected reliable and valid measures of all three SWB components, and that well-established predictors of SWB such as personality and sociodemographic variables should be controlled to highlight EIs unique variance.
10 SWB, SS & EI Underlying processes of SS and EI seem similar. The moderator question is:Does the relationship between SS and SWB depend on an individuals’ level of EI?Interestingly, the underlying processes of SS and EI seem similar, and focus on effective emotion regulation.This similarity suggests that EI may act as a moderator in the well-established relationship between SS and SWB.Proposition: If both SS and EI act to regulate emotion it is possible that where EI is high, individuals self-regulate emotion in a highly effective way which promotes SWB, and they therefore may not need additional emotional regulation from SS to report high SWB. Conversely, where EI is low, self-regulation of emotion is not as effective, consequently SS is more important in eliciting a report of high SWB.Thus, the moderator question is: Does the relationship between SS and SWB depend on an individuals’ level of EI?Impetus for this idea was also drawn from findings that show EI acts as a moderator in stress related research (Ciarrochi et al., 2002; Slaski & Cartwright, 2002; Slaski & Cartwright, 2003).
11 HypothesesThat where sociodemographic variables and personality factors are controlled, SS from Significant Other, Family and Friends would significantly predict SWL, PA and NA. It was also hypothesised that where sociodemographic variables and personality factors are controlled, EI would significantly predict SWL, PA and NA. It was further hypothesised that EI would significantly influence the relationship between SS and SWB as measured by sources of support and SWL, PA and NA respectively, where the interaction effects between SS and EI would add significant variance to the prediction of SWB beyond main effects.Thus, there are three aims to our study1. To further investigate the predictive value of SS on SWB, taking into account: all three components of SWB; the need to control for sociodemographic and personality variables; and the need for further exploration of source and perception of support.2. To further explore the predictive value of EI on SWB.3. To explore the proposition that EI acts as a moderator in the well-established relationship between SS and SWB.It was hypothesised that where sociodemographic variables and personality factors are controlled, SS from Significant Other, Family and Friends would significantly predict SWL, PA and NA. It was also hypothesised that where sociodemographic variables and personality factors are controlled, EI would significantly predict SWL, PA and NA. It was further hypothesised that EI would significantly influence the relationship between SS and SWB as measured by sources of support and SWL, PA and NA respectively, where the interaction effects between SS and EI would add significant variance to the prediction of SWB beyond main effects.
12 Participants 267 adults 196 females, 71 males. general population: who volunteered after reading explanatory statement.Age (M=41.52 years, SD=14.28).52.8% in relationships.72.6% were tertiary educated.Income range $80-$2699 (gross) per week, modal income $1000 per week.The sample consisted of 267 adults, 196 female and 71 male (original sample size was 273, six cases were delete as >5% of the survey was not competed).The age range was years oldJust over half were in relationshipsThe majority were tertiary educatedAnd the income range was large – with the modal income being $1000
13 Plus: socio-dem questions Plus: Ballard’s short social desirability MeasuresResponse Type – all self reportPlus: socio-dem questionsAlpha(Original)This StudyItemexample# of Items5 point Likert Scale:very slightly or not at all to, extremelyPositive & Negative Affect Schedule – 2 subscales (PANAS)2010 e/sPA, (.88) .88NA, (.87) .89Indicate how you feel eg. interestedRef: Watson et al. (1988)7 point Likert Scale: Strongly Disagree to, Strongly AgreeI am satisfied with lifeRef: Diener et al. (1985)Satisfaction With Life Scale (SWLS)5(.87) .89Multidimensional Scale of Perceived Social Support - 3 subscales (MSPSS)124 e/s7 point Likert Scale: Very Strongly Disagree to, Very Strongly AgreeThere is a special person…when I am in needRef: Zimet et al. (1988)SO: (.91) .95FAM: (.87) .94FRI: (.85) .945 point Likert Scale: Strongly Disagree to, Strongly AgreeSchutte Emotional Intelligence Scale (EIS)Seven self-report measures were collated to form a questionnaire battery or kit. All the measures are relatively simple and self-explanatory.All measures used were well-respected and valid as critiqued within the psychological literature, and report good alpha levels in this study and in original studies.Three versions of the questionnaire were used to combat order effects.33(.90) .90I have control over my emotionsRef: Schutte et al. (1998)International Personality Item Pool – 5 subscales (IPIP)E: (.87) .88A: (.82) .75C: (.79) .84ES: (.86) .93II: (.84) .765010 e/s6 point Likert Scale: Very inaccurate to, Very accurateAm the life of the partyRef: Goldberg (1999)Plus: Ballard’s short social desirability
14 Procedure Questionnaire kits with posters. Public places. Explanatory statement.Anonymity and confidentiality.20 minutes.Australia Post or returns box.To elicit anonymous respondents, questionnaire kits were left in public places such as gyms, medical centers and libraries for ~ 8 weeks.The kits were placed with posters inviting people to participant.The first page of the questionnaire was an explanatory statement and assured anonymity and confidentiality.When completed participants were asked to place questionnaires in the accompanying returns box or mail it in the attached prepaid envelope to Monash University.There was no time limit for completion, it was indicated it would take 20 minutes to complete.1200 questionnaires were distributed and 273 returned = 4.4% return rate.
15 AnalysesMissing values and violations of the assumptions of multivariate analysis8 univariate outliers truncatedMean replacement where missing values were less than 5%N on pairwise analysesNo order effectsMain and interaction effects =Hierarchical Multiple Regression in SPSSPrior to the main analyses the data was screened and cleaned for missing values and the assumptions of multivariate analysis, including multicollinearity . To ascertain the relative contribution of SS and EI and their interaction effect on SWB Hierarchical Multiple Regression (HMR) analysis was employed.
16 HMR stepsBefore the HMR was conducted, bivariate correlations were run between all variables to test for assumptions of multicollinearity and to find which variables would be entered as control variables. Not all sociodemographic and personality variables were entered as controls. Only those that elicited a significant correlation to the DVs were entered at the first step as control variables.Thus, the variables entered at the first step of the HMR for each DV were slightly different, however each of steps 2,3, and 4 were the same.NB: Most of the bivariate correlations were moderate in strength. HRM was conducted according to Baron and Kenny (1986).
17 Satisfaction with Life Model as a whole predicted 44.1% (40.7% adjusted) R=.66, F(13, 216)=13.09, p<.0011 Sociodem & Personality 34%*2 SS %*3 EI %*4 SS x EI (interaction) %*- EI, and SSso x EI were significant at the last step*significant at .05Results for Satisfaction with LifeStep 1 Sociodemographics & Personality 34%*Step 2 Social support %*Step 3 EI %*Step 4 SS x EI %**significant at .05At the last step EI was significant, however none of the SS sources on their own were significant. The interaction effect between SS and EI was significant.
18 Positive AffectModel as a whole predicted 47.7% (44.9% adjusted) R=.69, F(13, 239)=16.78, p<.0011 Sociodem, Personality & Sdes 41.3%*2 SS (2%)3 EI %*4 SS x EI (interaction) + (1%)- EI significant in last step*significant at .05Results for Positive AffectStep 1 Sociodemographics, Personality & Social Desirability 41.3%*Step 2 SS (2%) not sigStep 3 EI %*Step 4 SS x EI (1%)not sigAt that last step EI was significant. Again, none of the individual sources of support were significant.
19 Negative AffectModel as a whole predicted 50.3% (47.6% adjusted) R=.71, F(13, 239)=18.60, p<.0011 Sociodem, Personality & Sdes 45%*2 SS %*3 EI (1%)4 SS x EI (interaction) (1%)- SS Friends x EI was significant at the last step*significant at .05Results for Negative AffectStep 1 Sociodemographics, Personality & Social desirability 45%*Step 2 SS %*Step 3 EI (1%) not sigStep 4 SS x EI (1%) not sigAt the last step the interaction between SS Friends x EI was significant, even though the step as a whole was not significant. NA was the only DV where individually one of the SS sources was significant in its own right, was Family.Overall the results did not yield consistent results for SS or EI across the three DVs.
20 Interactions SWL: SS Significant Other x EI NA: SS Friends x EI Interactions indicate that the relationship between SS and SWB is dependent on the level of EI reported.Interaction ResultsThe analyses showed two interactions: SWL and SS significant other x EI; NA and SS friends x EI.The presences of interactions indicate that the relationship between SS and SWB is dependent on the level of EI reported.To interpret the interaction EI was split into high and low EI to find where the interaction happens.
21 SWL SS Significant Other x EI SWL interactionWhere EI is high the level of perceived Significant Other SS does not interact to produce a differential level of SWL.Conversely, where EI is low, the level of perceived Significant Other SS does interact with EI to produce a higher level of SWL.EI low, b = t = p = .0008EI high, b = t = p =
22 NA SS Friends x EI EI low, b = -0.22 t = -2.79 p = .0028 NA interactionWhere EI is high the level of perceived Friends SS does not interact to produce a differential level of NA.Conversely, where EI is low, the level of perceived Friends SS does interact with EI to produce a lower level of NA.EI low, b = t = p = .0028EI high, b = t = .88 p =
23 Discussion The results partially support the hypotheses. SS predicted SWL and NA, but not PA.EI predicted SWL and PA, but not NA.2/9 interaction terms significant.Each model as a whole was significant.The results only partially support the hypotheses, even though the models as a whole significantly predict SWB.
24 DiscussionSmall amount of variance accounted for by SS, inconsistent with previous research.Possible explanation is stringent controls; however can be seen to provide clearer results for unique value of SS.Differential weights of SS sources across DVs; important consideration.The relatively small amount of variance SS accounted for was inconsistent with previous findings.A possible explanation for this finding is that our control variables were very robust. But we feel that this provides clearer results for both SS and EI in accordance to research considerations highlighted earlier.It was also interesting to find that the sources of SS did not yield consistent variance across each of the three SWB DVs. We feel that this shows that it is important to consider sources of support and all three components of SWB because they show differing results.
25 Discussion Adds support to the discriminant validity of EI. Predictive value of EI in SWB exciting, as EI thought to be subject to change.Significant interaction terms support suggestions that EI is important to SWB and SS.Regarding EI, this study adds to the growing support for the discriminant validity of EI.Although the incremental value of the significant results seem small they are important because EI was regressed after Sociodemographic variables, personality and social support – which are all well-established predictors of SWB.If EI can be repeatedly shown to predict SWB above personality, and it can be developed, its value in the SWB equation becomes exponentially important.The current study’s results support the suggestion that EI is important in the relationship between SS and SWB – due to the significant interaction terms found.
26 Discussion SS x EI interaction re: SWB not previously published. Where EI is high the level of perceived SS is inconsequential to the level of SWB reported. However, where EI is low the level of perceived SS and level of EI interact to produce a differential level of SWB, with higher reported SS having a relationship with higher SWB .The interaction findings are important because such interactions have not been previously published.Both the significant interaction terms elicit similar interpretation:the core point being that where EI is high the level of perceived SS is inconsequential to the level of SWB reported, however,where EI is low the level of perceived SS and level of EI interact to produce a differential level of SWB, with higher reported SS having a relationship with higher SWB.
27 DiscussionSS may not be necessary for everyone, specifically those with high EI.EI training could be considered as an alternative to formal SS.Mindful that these are ‘new’ results – with some limitations: gender and education split.The current findings are important because they suggest that a high level of perceived SS reported may not always be necessary in the SWB equation, as previously purported - perhaps SS is not necessary for everyone.In light of this, it is suggested that EI training could be considered an alternative to linking people into formal SS channels, where increases of SWB are concerned. This idea is put forth because the current results suggest that higher EI may act to protect individuals against reports of low SWB where they also report low perceived SS. In addition, it is argued that providing resources to enhance natural social networks are better than developing additional formal social support (Cutrona & Cole, 2000).Although these results are exciting, a number of limitations should be considered. The two prominent limitations are the gender and education split of the sample.
28 Discussion Valuable insight into SS and EI as predictors of SWB. Addressed limitations of earlier studies.Shows that continuous investigations into apparently robust relationships are warranted.The study provides a valuable insights into SS and EI as predictors of SWB.Addressed limitations of earlier studies by using all three components of SWB, using control variables, and using perception and source of support.Shows that continuous investigations into apparently robust relationships, such as the one between SWB and SS are warranted.
29 DiscussionProfessionals concerned with SWB need to consider main and interactions effects of known variables.This study shows that the relationship between SS and EI on SWB goes beyond main effects, and that it is important to explore the three components of SWB.A potential practical implication of the findings is that mental health professionals could consider not only main effects of known SWB variables but their interaction effects, which could help then allocate resources better.In conclusion, this study has shown that the relationship between SS and EI on SWB goes beyond main effects, and that the relationships between the three components of SWB found here add a level of complexity to current knowledge not before published.