Review of experiment data | My Assignment Tutor

Review of experiment data Please download the datasheet from BREO and open it in SPSS. You will see in the datasheet 9 variables. When the data was downloaded each participant was assigned a number from 1 to 124, so that each data point had a label for removal of outliers. Gender and age will be used for the participants’ section of your Method. This leaves 3 variables for each condition and therefore each multiple regression. ‘AppraisalD’ denotes the anticipatory appraisal value given to the stress task for the debilitating condition; this is a function of the demands rating divided by the resources rating, such that an appraisal > 1 reflects a threat state, and an appraisal < 1 reflects a challenge state. ‘ChangeHRD’ denotes the change in heart rate from baseline to post-stress task in the debilitating condition, representing the physiological arousal induced by the stress task. ‘CogPerfD’ denotes the N-Back task score for the debilitating condition (out of 50). It thus follows that ‘AppraisalE’, ‘ChangeHRE’ and ‘CogPerfE’ denote the same variables but for the enhancing condition. In the first regression model, we will use ‘AppraisalD’ and ‘ChangeHRD’ to predict ‘CogPerfD’. In the second regression model, we will use ‘AppraisalE’ and ‘ChangeHRE’ to predict ‘CogPerfE’. You may recall completing the appraisal scale twice: once before the video about stress, and once after. This is because we wanted to see whether the video was successful in changing your appraisals of stress. So we conducted a t-test. There was no significant change in appraisal after watching the stress appraisal video for the debilitating condition, t(68) = -.22, p = .824, 95%CI[-.08, .06]; or for the enhancing condition, t(68) = .77, p = .446, 95%CI[-.05, .11]. Assumption testing Multiple regressions are used to determine whether more than one independent variable can predict a dependent variable. In a regression model, these are also referred to as predictor and outcome variables, respectively. There are assumptions for each type of analysis, because the data being tested must meet the specification of the model that each analysis is designed to test. Each type of analysis is essentially a test of how well a set of data fits a predetermined model/pattern. In this case, we are running a multiple regression to test how well our class experiment data fits a predictive model. To run a multiple regression, there are many assumptions to check. The following 4 have been checked for you. Please write up these notes at the beginning of the results section as part of your written preamble. Linear relationship between independent (predictor) variables and dependent (outcome) variableNot normally distributedNon-zero variance was violated: it was apparent that two participants pressed M on every trial. These were removed5 extreme outliers removed Running the regression – SPSS instructions Before you run your regression, you need to have made a note of what is met and what is violated in your assumptions. Then you need to revisit your hypotheses so that you know what it is you’re testing when you open the regression window. Your dependent (outcome) variable is what you are predicting – the N-Back test scores; your independent (predictor) variables are the things you are manipulating during the experiment and are expecting to influence the dependent variable – stress appraisal and change in heart rate. You will be running two regressions – one for each condition (stress is debilitating / stress is enhancing). To bring up the regression window go to the Analyse tab at the top of the SPSS window, and select regression from the dropdown, then linear from the dropdown that follows. Move your independent (predictor) variables into the “Independent(s)” box, and your dependent (outcome) variable into the “Dependent” box. We are conducting an “enter method” regression. Under ‘statistics’ select the options for estimates and model fit (these are your basic regression, and should already be ticked), descriptives (these are the ones you report with the analysis in the results section as they’re specific to the regression model) and confidence intervals (for your inferential statistics). Then click continue and okay to run the regression. Reporting the regression – descriptive statistics This guide is intended as that alone, and you should refer to your research methods materials as well as the APA website for guidance on how to write up your results and your lab report. We are now adhering to the 7th edition of the APA publication guidelines. Once you’ve run your analyses, you are ready to pull all the statistics together to report the regressions. First, the descriptive statistics (mean, standard deviation and number of data points for each variable included in the regression). These are found in the first part of your output under the regression title. You can combine your descriptive statistics for both regressions into one table if you like, or keep them in separate tables, or report them in text, if you have the word count to allow for this. If you choose to present the statistics in a table, make sure to refer to it in text, and that the layout is concordant with APA 7. For example: Table 1 Descriptive statistics for variables included in the regression for the debilitating condition VariableNumberMeanStandard DeviationN-Back task score6940.756.82Stress Appraisal69.83.48Change in arousal696.009.16 Reporting the regression – inferential statistics The most important part of interpreting any analysis is to know where to get the numbers for your analytic statement. The analytic statement for a regression generally follows a statement of whether the model was significant or not, with the format: F(degrees of freedom, residual degrees of freedom) = F-value, p = model significance value, adj. R2 = effect size value For example: “Stress appraisal and change in physiological arousal did not significantly predict working memory performance, F(2, 66) = .29, p = .75, adj. R2 = -.021.” Your degrees of freedom, f-value and p-value for the model are all found in the (confusingly labelled) ANOVA box in your output. The adjusted R square value is given in the Model Summary box; this tells you the size of the predictive effect that the independent (predictor) variables had on your dependent (outcome) variable, as well as how much of the variance in your dependent (outcome) variable is explained by the model (as a percentage, so using the above statistics, stress appraisal and physiological arousal explains 2.1% of the variance in working memory performance). We use adj. R2 because it is modified to account for the number of independent (predictor) variables, and therefore a better model evaluator for multiple regressions. To go the next step, you can look at the Coefficients box, which gives you the individual regression coefficients and confidence intervals. This breaks down the regression model to show you the effect of each predictor in the model individually. The unstandardized coefficient (B) tells you by how much each variable changes with each unit increase in the dependent (outcome) variable, whilst the standardised coefficient (β) tells you how strong this effect is. That is to say that as stress appraisal increases by 1 unit change (towards a threat appraisal), the N-Back score increased by 1.31 (95%CI[-2.18, 4.79]). However, this was not significant (p = .457) with a negligible effect (β = .093). This information can be presented in a table or as a series of statements in-text as above. If you choose to present the data in a table, please make sure that the formatting is concordant with APA 7, and that you refer to it in text. Never copy tables from SPSS. Report all statistics to 2 decimal places, aside from significance values and effect sizes (3dp). Please refer to your research methods materials and APA 7 for the correct write-up style and format. Rinse and repeat for the enhancing condition! Don’t forget to write up your results in APA 7 


Leave a Reply

Your email address will not be published. Required fields are marked *