Sex Differences in Well Being: Research by Julie Pallant - Coursework Example

? Sex Differences in well-being: A replication of Research Published by Julie Pallant Insert Pallant’s research was replicated; the focus was the importance of reporting effect size and the value of replicating research. The participants were males and females, mixed ages, recruited using opportunistic sampling. The sample quota was determined by running a power analysis. Significant results were found in both studies but there was a different effect size, which shows reporting effect size, has value as it supports the importance of replicating research and the reporting of effect size to provide a fuller picture of the results of research. Maybe this has implications for differentiating stress management programs according to gender. Introduction A result is considered statistically significant if it was not likely to happen by chance. This implies that a statistically significant test is one, which has assuring proof and can be explained. This research was done putting more focus on effect size and statistical significance. The report below gives a description of the investigation on sex difference and well-being while replicating the research done by Julie Pallant. Sample size is an important component of statistical significance in that larger samples have less mishap chances, a debate that is still ongoing. There is growing debate about the validity of significance testing for example Bakan(1966)concluded that “the test of statistical significance in psychological research may be taken as an instance of a kind of mindlessness in the conduct of research” (p .436). The level at which one can accept whether an event is statistically significant is known as the impact level or p-value. There are two forms of errors in statistical testing type I and type II .A type 1 error is also referred to as error of the first class. This is the wrong decision that occurs when a test discards a true or null hypothesis. The rate of the type I error is called the size of the test and denoted by the Greek letter ? (alpha) and it often equals to the significance level of a test. On the other hand, a type II error is the wrong decision that is made when a test accepts a false null hypothesis. The rate of the type II error is signified by the Greek letter ? (beta) and associated to the power of a test (which equals 1 ? ?). Neale and Liebert considered that a single study has a limited capacity for findings to be generalized. (Neale and Liebert, 1986 cited in Thompson, B), because there are too many substitute explanations which can be offered thus it establishes the need for research to be replicated, and justifies the current research (i.e. ours!). Fisher’s alpha level was set by arbitrary agreement, and that now there is a growing opinion that just refuting the null hypothesis is insufficient, “unsound, poor scientific strategy” (Meehl, 1978 cited in Dunleavy, Barr, Glenn and Miller: 29), There is greater emphasis on reporting effect size with Clark saying that statistical tests did not give researchers enough information (Clark cited in Carver 1978). Carver (1978) said alternative methods to statistical tests should be considered. The aim of this research is to replicate Pallant’s research and this will add additional data to either support or question Pallant’s finding of a genuine effect. Thus, the question this research sets to answer is, can Pallant’s well being research and results be replicated using effect size data from the original study? Thus, the following steps will explain the replication of Pallant’s research; Design Pallant’s research was the design for the current study, and as with Pallant’s study two measures were used, the Positive and Negative Affect Schedule (PANAS) and the Perceived Stress Scale (PSS). The study was a group differences design The independent variable (IV) was the sex of the participant. This had two levels – male and female. The dependent variables (DVs) were scores on self-report measures of positive effect, negative effect, and perceived stress. The quota system was used and was estimated by calculating effect size d from the means found in Pallant’s study for males and females on their levels of perceived stress and then performing a power analysis to estimate required sample size (Bakan, 1966). As for the participants, through the quota sampling we worked out the number of participants needed from Pallant’s study by calculating the effect size d from the mean found in Pallant’s study then performing a power analysis to estimate required sample size. This was teamwork in general, with everyone contributing in collecting the data. Based on this estimate, data from 137 males and 185 females was collected. Age ranged from 18 to 79 with a mean of 32.44 (SD = 12.85). Ethical guidelines were adhered to. The participants gave informed consent, their right to withdraw was explained and all participants were debriefed. All participants completed both the PANAS and the PSS and no one withdrew their consent. Materials In this research, we used the same measures as Pallant’s used in her research. One being the positive and negative affect schedule (PANAS) by Watson et al. 1988. This scale is made up of two mood scales, one measuring affirmative affect and the other measuring negative effect. Each item is rated on a 5-point degree ranging from 1 = very faintly or not at all to 5 = exceedingly to point out the degree to which the respondent has felt this way in the designated time frame. The scale has been used to measure effect at this moment, today, the past few days, the past week, the past few weeks, the past year, and generally (on average). The second scale used was the perceived stress scale (PSS) by Sheldon Cohen. The scale measures the extent to which circumstances in one’s life are considered stressful. There is also a quantity of direct reservations about present levels of experienced stress. The perceived stress scale had 10 questions. It should be noted that both the PANAS and PSS are Likert scales and both scales set out to measure the stress level in the lives of the participants and the daily lives effect in relation to stress (Elliot & Steven, p. 142 .2011). Procedure The data was collected by a team, which first set out to get the consent of the participants. The participants were briefed of what was expected of them and how much of their time the research would probably consume. Anyone not in agreement or did not want to do the research was allowed to return a blank questionnaires and that by filling the information on the questionnaires they gave us the researchers permission to use the information as we desired. The team then collected the data and fulfilled the quota needed for the research. The data was then put into SPSS. Results Pallant’s findings: 184 Males; 247 Females Wellbeing: F (3, 427) Wilks’ Lambda = 3.46, p = .016, ?2p = .024. Positive Affect: Males: M = 33.63 (6.99); Females: M = 33.63 (7.39). F (1, 429) < 0.01, p = .997, d < .01. Negative Affect: Males: M = 18.71 (6.90); Females: M = 19.91 (7.09). F (1, 429) = 3.09, p = .079, d = .17. Perceived Stress: Males: M = 25.79 (5.41); Females: M = 27.41 (6.09). F (1, 429) = 8.25, p = .004, d = .28. Given that there were three dependent variables designed to measure a construct of ‘wellbeing’, the data were inspected to determine whether they satisfied the assumptions for multivariate analysis of variance (MANOVA). There was an adequate sample size as there were more cases in each cell than there were DVs. The data were normally distributed and there were no univariate or multivariate outliers (Carver, 1978). There was a straight line relationship between each pair of DVs indicating linearity of relationship. There were only moderate relationships between the DVs; thus there was no concern for multicollinearity and singularity. Box’s M test indicated that there was homogeneity of variance-covariance matrices and Levene’s tests indicated that there was homogeneity of variance for all three DVs. Having satisfied these assumptions, a one-way MANOVA was performed on the data. Our findings: 137Males; 185 Females Wellbeing: F (3, 318) Wilks’ Lambda = 10.19, p Read More