Guest Posted March 17, 2016 Report Share Posted March 17, 2016 Hello!! I am writing my psychology IA right now, but now I'm at the result section and I don't understand a thing! My teacher does not explain this well, and he doesn't seem to be certain about what he's doing. There was actually a few people last year who failed their psychology, mainly because of their IAs so I'm a bit terrified! I'm a HL student, so I will need inferential statistics but how does that actually work?? For example, my teacher said I should use the General Linear Model, but I don't know if I'm allowed to use that.. I haven't found anyone who have used it before, but it does calculated the level of significance. Also, do I need a graph or something similar for the inferential part? (I already have graphs in the descriptive statistics section of the essay) Thank you so much in advance! Reply Link to post Share on other sites More sharing options...
blindpet Posted March 17, 2016 Report Share Posted March 17, 2016 General Linear Model sounds more advanced than necessary for IB Psychology. Generally a simple t-test will suffice or Mann Whitney test. ANOVAs are something you learn in university Psychology programs. The purpose of inferential statistics is to provide a probability that the results you got could have occurred from chance. The smaller your probability (p level) the less likely it is your results occurred due to chance. You do not use graphs for inferential statistics p level reporting, you report the means, standard deviation and p level, then you use your descriptive statistics for the data visualization in a graph. 1 Reply Link to post Share on other sites More sharing options...
Guest Posted March 19, 2016 Report Share Posted March 19, 2016 On 2016-03-17 at 1:54 PM, blindpet said: General Linear Model sounds more advanced than necessary for IB Psychology. Generally a simple t-test will suffice or Mann Whitney test. ANOVAs are something you learn in university Psychology programs. The purpose of inferential statistics is to provide a probability that the results you got could have occurred from chance. The smaller your probability (p level) the less likely it is your results occurred due to chance. You do not use graphs for inferential statistics p level reporting, you report the means, standard deviation and p level, then you use your descriptive statistics for the data visualization in a graph. Thank you so much! It makes more sense to me now! Just another question - my teacher took my results from the experiment through a statistical programme called SPSS, should I in any way mention that in my essay, or should I just display the results and raw data? Reply Link to post Share on other sites More sharing options...
ellie Posted March 20, 2016 Report Share Posted March 20, 2016 On 3/19/2016 at 1:59 PM, Caramell said: Thank you so much! It makes more sense to me now! Just another question - my teacher took my results from the experiment through a statistical programme called SPSS, should I in any way mention that in my essay, or should I just display the results and raw data? What I did was to add a screenshot as an appendix with a footnote where I said that I used webpage x to calculate. This was done in a previous year by someone who got a 7 in their IA, so I think that should be fine. Reply Link to post Share on other sites More sharing options...
blindpet Posted March 20, 2016 Report Share Posted March 20, 2016 I honestly think the IB may be suspicious if students are using SPSS for statistics. That is a university level statistics software that I used in my bachelor's and master's program. Things may have changed though. If I were you I'd use a simple t-test like I have seen most Psych HL IAs use. 1 Reply Link to post Share on other sites More sharing options...
Guest Posted March 22, 2016 Report Share Posted March 22, 2016 Thank you so much! Now I'm just wondering how I will use the t-test and what variables that are considered then. (I researched if there was a correlation between attractiveness and trustworthiness and likeability.) Reply Link to post Share on other sites More sharing options...
blindpet Posted March 22, 2016 Report Share Posted March 22, 2016 @Caramell sounds like you looked at correlation which would be Spearman's rho testing (it is the Pearson's correlation equivalent for non-parametric data i.e. for ordinal data). Then again I do not know the details of your study, initially you mentioned the General Linear Model, perhaps your teacher used an ANCOVA. It would help if you laid out simply the design of the experiment and all the measures taken. Reply Link to post Share on other sites More sharing options...
Guest Posted March 23, 2016 Report Share Posted March 23, 2016 This is a part of the "procedure" section of my IA: " Each participant was presented with four different photographs separately, one from each of the four groups established in the pilot study (attractive male, attractive female ,unattractive male, unattractive female). The size of the photographs were measured to the 15x20 cm big. The participants then rated each photograph on likeability, trustworthiness and attractiveness using a 7-point scale (0 = not at all, 7 = very). The final question displayed all four previous photographs (see figure...). Together with the set of pictures, a hypothetical situation was presented. The participants were asked to choose which of the individuals on the photographs they would trust the most, if they would lend the person money with the promise to return it with interest. Every participant rated the same pictures and was presented with the same hypothetical situation." I'm not sure what measurements you were thinking of, but anyway - this is the calculated mean, median and mode values: Mean; Trustworthiness Mean; Likeability Person 1 (attractive male) 4.88 4.63 Person 2 (unattractive female) 4.5 3.88 Person 3 (unattractive male) 3.94 3.63 Person 4 (attractive female) 5.25 5.38 Trustworthiness Likeability Mode Median Mode Median Person 1 (attractive male) 4 5 5 5 Person 2 (unattractive female) 4 4 5 4 Person 3 (unattractive male) 3 4 3 3.5 Person 4 (attractive female) 5 5 5 5 And this is the (individual) result of the hypothetical situation (hence which of the four people they would trust the most): 4. 4. 4. 4. 4. 4. 4. 2. 4. 3. 1. 4. 4. 4. 2. 4. Reply Link to post Share on other sites More sharing options...
Guest Posted March 23, 2016 Report Share Posted March 23, 2016 I have now found a webpage where I was able to do a paired t-test, which I hope works. I'm just wondering exactly what numbers I should mention and if you think the results seems reasonable?? Reply Link to post Share on other sites More sharing options...
blindpet Posted March 23, 2016 Report Share Posted March 23, 2016 You didn't mention which design you used but it looks like it was repeated measures. Your data are ordinal so a t-test is not appropriate. If it was a repeated measures test (all participants rated all of the photos) then you should use a Wilcoxon t-test. If it was independent measures (between subjects), then you should use the Mann Whitney U test. These are just to look for differences between the groups (if you wanted to look for interactions then you would use the Kruskal Wallis or Friedman's test). You mentioned you were looking for a correlation in which case you should use Spearman's Rho since your data are ordinal (it is the non-parametric equivalent of the Pearson test). For Spearman's you would run the raw data and not the means, plot the data and report the correlation coefficient. Reply Link to post Share on other sites More sharing options...
Benji219 Posted February 11, 2020 Report Share Posted February 11, 2020 On 3/17/2016 at 7:54 AM, blindpet said: General Linear Model sounds more advanced than necessary for IB Psychology. Generally a simple t-test will suffice or Mann Whitney test. ANOVAs are something you learn in university Psychology programs. The purpose of inferential statistics is to provide a probability that the results you got could have occurred from chance. The smaller your probability (p level) the less likely it is your results occurred due to chance. You do not use graphs for inferential statistics p level reporting, you report the means, standard deviation and p level, then you use your descriptive statistics for the data visualization in a graph. I rejected my research in my descriptive and rejected my null hypothesis in my inferential data - I've doubled checked the math - what does this mean. Reply Link to post Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.