tag:blogger.com,1999:blog-8909074830238091680.comments2015-03-28T02:22:40.666+11:00Jeromy Anglim's Blog: Psychology and StatisticsJeromy Anglimnoreply@blogger.comBlogger781125tag:blogger.com,1999:blog-8909074830238091680.post-65083039596809526792014-10-20T06:00:00.662+11:002014-10-20T06:00:00.662+11:00Hi. I am so happy you wrote this article. I have b...Hi. I am so happy you wrote this article. I have been voraciously looking for some information regarding this. <br />I did a Master degree in organisational psychology via correspondence (without a graduate degree in it. I graduated as an engineer.). Would that have any value in Australia? I do not have any experience in this field.:(<br />Everywhere I read, I read about a requirement of 6 year course. <br />Can you throw some light on that.confused psychhttp://scribblography.wordpress.com/noreply@blogger.comtag:blogger.com,1999:blog-8909074830238091680.post-7903379996748447392014-09-18T13:01:22.329+10:002014-09-18T13:01:22.329+10:00Hi Christopher,
If you are using an established ...Hi Christopher, <br /><br />If you are using an established psychological scale like the IPIP, then you would generally just score the test as the test designers intended it to be scored. <br />This allows your results to be easily comparable to other studies that have used the same scale. <br />In such cases the exploratory factor analysis is generally just a guide to check that such an approach is reasonable.<br /><br />However, if you've modified the items a lot, then yes, you would need to perform exploratory factor analysis. In general, creating a scale is a big job, and I would advise against modifying personality tests. <br /><br />That said, you have adapted the test and perhaps you have good reasons for doing it.<br />In terms of resources, check out:<br /><br />The EFA lecture and tutorial as part of this course:<br />http://jeromyanglim.blogspot.com.au/2013/10/tutorials-answers-and-data-files-for.html<br /><br />And here are some general readings of EFA:<br />http://jeromyanglim.blogspot.com.au/2009/10/exploratory-factor-analysis-and-scale.html<br /><br /><br /><br /><br /><br /><br /><br /><br />Jeromy Anglimhttp://www.blogger.com/profile/12949204812496382042noreply@blogger.comtag:blogger.com,1999:blog-8909074830238091680.post-59502233469487780152014-09-16T04:05:00.632+10:002014-09-16T04:05:00.632+10:00Hello Jeromy!
Thanks for providing an interesting...Hello Jeromy!<br /><br />Thanks for providing an interesting blog! I’d appreciate it greatly if you could give me some advice or direct me to relevant resources concerning scale development.<br />Being a bit over ambitious 3 year student (psychology) I was/am interested to research the relationship between personality in first year students and their future academic performance (i.e. study pace, GPA, drop-out and motivation). I was afraid social desirability would influence the responses, as it was administered during their admissions to my university. Thus, I created my own item pool, inspired from IPIP scales. The questionnaire contains 160 items and is answered with a 6-point likert-scale.<br />I have now finished gathering my sample, resulting with approx. 500 respondents. The next procedure for me, which I sadly understand too little about, is to perform dimension reduction (I use SPSS):<br /><br />a) Decide which factor solution and rotation would be appropriate<br />b) Remove items with low factor loadings. Back to Step a.<br />c) Decide on the numbers of factors to extract<br />d) Convert raw scores into to new scores based on their factor loadings appropriate for further regression analyses to my dependent variables.<br /><br />1. Do you have any ideas or advice, which factor-solution & rotation would be appropriate for each step? <br />2. Which method would you recommend to compute factor scores? <br /><br />Greetings from Finland,<br />Christopher<br />Christopher Harjuhttp://www.blogger.com/profile/07596527347384472949noreply@blogger.comtag:blogger.com,1999:blog-8909074830238091680.post-79018844895836833512014-08-15T19:57:28.721+10:002014-08-15T19:57:28.721+10:00Hi Jeremy and fellow readers. Thank you Jeremy for...Hi Jeremy and fellow readers. Thank you Jeremy for your interesting and informative post. I really appreciate the content on your blog and appreciate the effort you put into making it available.<br /><br />Similar to Michael Reinhard, I work at a small private institution where I cannot justify the licensing fees for SPSS. I also teach statistics and research methods to undergraduate students and made the switch to R about 4/5 years ago. Here in South Africa, the history of Apartheid and an educational system struggling to come to grips with its own socio-political context, means that students come in to study with frustratingly low levels of mathematical and computer literacy. Getting math-phobic students to grasp statistical concepts is difficult enough, making them sit in front of a console and type in commands is just too intimidating for most (as suggested by jrkrideau). For this reason I teach them using the Rcmdr package. It has all the statistical functions that I need, the students get to see the commands that are run, and I can get them to play around with functions and arguments in the script window to gradually ease them into programming their own scripts. The advantage of this approach is that it provides an easy middle ground between the SPSS mental model and the R approach. Rcmdr makes it possible for students to see and edit their data in a tabular format, they can choose operations using an intuitive and helpful (it often prevents irrational analytical options from being chosen) menu interface, and yet they can still load additional packages, and run additional functions to add to the default outputs. In terms of teaching materials, in the past I have used Howell's text and provided the students with a DVD that has a number of tutorials that goes through the analysis in R we cover and interpretation of output using examples. This is the downfall in using GUIs, it is not easy to incorporate their usage in texts without taking up a lot of space, but the DVD approach is useful. I have ordered Field's text but haven't had a chance to read it yet, and I just downloaded Navarro's book and will have a look at that when I get a chance. I would definitely like to transition to an R based text some time in the future. I prefer to teach my students in R because it is affordable to them and creates opportunities for those that really want to grow their knowledge and skills (it provides them with an asset they can continue using after they leave the class and a propriety licence would have expired). Also, as this post suggests (but you focus more on stats than research) it allows for transference of learning between subjects. For example, in my psychometrics class I refer them to functions for DIF or other concepts they are learning to play around with at home with using R. By giving them an opportunity to play with these functions they improve their understanding of concepts and retention of new material. <br />The biggest challenge and disadvantage of using R as a teaching medium is articulation options for students. Where I teach (WC province in SA) students entering any one of the big state universities for post-grad will have to transition to SPSS for their post-graduate work, they will not have the opportunity to develop their 1st level statistical skills (e.g.: linear models and basic descriptives) or learn 2nd level statistical approaches (e.g.: modelling approaches such as PLS, SEM, and CFA or other multivariate techniques). Only those that are very motivated and competent, will be able to go on make use of stackoverflow, youtube videos, or workshops to develop this kind of competence.<br /><br />Kind Regards,<br />ConradC.S.Zygmonthttp://www.blogger.com/profile/09735050366969025470noreply@blogger.comtag:blogger.com,1999:blog-8909074830238091680.post-38870730180559214342014-08-10T00:01:33.738+10:002014-08-10T00:01:33.738+10:00The best latex APA is actually using biblatex, not...The best latex APA is actually using biblatex, not bibtex. biblatex uses your .bib file & also lets you use natbib citation styles.Joannahttp://www.blogger.com/profile/02968914847649268737noreply@blogger.comtag:blogger.com,1999:blog-8909074830238091680.post-65520863681001977102014-08-08T17:00:35.541+10:002014-08-08T17:00:35.541+10:00Better avoid table and use div..more ...Positionin...Better avoid table and use div..more ...<a href="http://www.corelangs.com/css/box/divindiv.html" rel="nofollow">Positioning Div</a><br /><br />linglingmaakihttp://www.blogger.com/profile/09477565569807599471noreply@blogger.comtag:blogger.com,1999:blog-8909074830238091680.post-24457663224213667662014-07-08T20:29:12.836+10:002014-07-08T20:29:12.836+10:00Hi Jeromy,
I have a data set containing 30 Binar...Hi Jeromy,<br /><br />I have a data set containing 30 Binary Variables (1,0) . I would like to perform SEM on these 30 Binary Variables against 4 Dependent Variables. Can you please explain the step by step procedure in R to perform the SEM analysis (R Code)Avinshhttp://www.blogger.com/profile/14272266272192325792noreply@blogger.comtag:blogger.com,1999:blog-8909074830238091680.post-11588539162649585572014-05-29T23:53:03.814+10:002014-05-29T23:53:03.814+10:00I am the first timer for using apacite, and got th...I am the first timer for using apacite, and got the following error message:<br /><br />!undefined control sequence. <br />1.1216 \st@rtbibchapter<br /><br />What's this? Thanks. Lihui Baihttp://www.blogger.com/profile/10038624117089266350noreply@blogger.comtag:blogger.com,1999:blog-8909074830238091680.post-50547165731347114682014-05-23T16:29:21.122+10:002014-05-23T16:29:21.122+10:00A great new course for beginners on youtube, has g...A great new course for beginners on youtube, has great content with short and crisp videos. Easy to follow. New content is getting added, So I suggest to subscribe.<br />https://www.youtube.com/watch?v=BGWVASxyow8&index=18&list=PLFAYD0dt5xCzTQHDhMPZwBoaAXWeVhZzgSelva Prabhakaranhttp://www.blogger.com/profile/05519694374996247829noreply@blogger.comtag:blogger.com,1999:blog-8909074830238091680.post-56990377886806817392014-05-06T20:38:28.865+10:002014-05-06T20:38:28.865+10:00how to find relative importance index using spss f...how to find relative importance index using spss for my likert scale data??Harrison Yanhttp://www.blogger.com/profile/12366111009273933160noreply@blogger.comtag:blogger.com,1999:blog-8909074830238091680.post-56099438351326313502014-05-02T15:27:53.752+10:002014-05-02T15:27:53.752+10:00I guess you could use the traditional graphics dev...I guess you could use the traditional graphics devices like png and pdf<br />http://stat.ethz.ch/R-manual/R-devel/library/grDevices/html/Devices.html<br /><br />Otherwise, I'm not sure. I guess you could scan the knitr options, but I haven't seen anything designed to do this: http://stat.ethz.ch/R-manual/R-devel/library/grDevices/html/Devices.htmlJeromy Anglimhttp://www.blogger.com/profile/12949204812496382042noreply@blogger.comtag:blogger.com,1999:blog-8909074830238091680.post-74402952848704416142014-05-02T13:08:43.370+10:002014-05-02T13:08:43.370+10:00Great cheat sheet for knitr newbies, thanks! I hav...Great cheat sheet for knitr newbies, thanks! I have a follow-up, though. Each chunk generates a .png, which default name is the name of the chunk. For example,<br />```{r ChunkName}<br /> p<-ggplot(ds, aes(x=varible1))<br /> p<-p+geom_bar()<br /> print(p)<br />```<br />will produce a .png named "ChunkName.png"<br /><br />If there are a several graphs to be produced by the same chunk, they will be automatically ordered. So<br />```{r ChunkName}<br />years <- 1997:1999<br />for (year in years){<br /> p<-ggplot(ds[ds$year==year,], aes(x=variable1))<br /> p<-p+geom_bar()<br /> print(p)<br />}<br />}<br />```<br />will produce 3 graphs: "ChunkName1.png", "ChunkName2.png" and "ChunkName3.png". However, I'd like to name the pngs myself, tying them to some automatic variable (like year, in this example), so that it produced something like<br />"ChunkName_1997.png", "ChunkName_1998.png" and "ChunkName_1999.png". <br /><br />Any ideas how to override the defaults? <br /><br /><br />Андрей Ковальhttp://www.blogger.com/profile/14264243871276423056noreply@blogger.comtag:blogger.com,1999:blog-8909074830238091680.post-16398006946768038312014-05-02T00:03:17.564+10:002014-05-02T00:03:17.564+10:00Dear Jeromy,
is there a chance that you could hav...Dear Jeromy,<br /><br />is there a chance that you could have a look at my question about modeling nominal variables in JAGS ?<br /><br />http://stats.stackexchange.com/questions/95929/bugs-jags-for-nominal-response-model<br /><br />thank you very much<br /><br />J.k1111http://www.blogger.com/profile/11449275035410259115noreply@blogger.comtag:blogger.com,1999:blog-8909074830238091680.post-84116919026177307512014-04-26T14:52:14.291+10:002014-04-26T14:52:14.291+10:00That's what this post is about. What do you wa...That's what this post is about. What do you want to know?Jeromy Anglimhttp://www.blogger.com/profile/12949204812496382042noreply@blogger.comtag:blogger.com,1999:blog-8909074830238091680.post-75939253518365646032014-04-26T14:51:15.047+10:002014-04-26T14:51:15.047+10:001. r means correlation; r12 means correlation betw...1. r means correlation; r12 means correlation between variable 1 and variable 2<br /><br />2. N means total sample size<br /><br />3 and 4. the symbol means square rootJeromy Anglimhttp://www.blogger.com/profile/12949204812496382042noreply@blogger.comtag:blogger.com,1999:blog-8909074830238091680.post-29638797646702761342014-04-25T09:16:35.793+10:002014-04-25T09:16:35.793+10:00Concerning Hotelling's T-test for "corre...Concerning Hotelling's T-test for "correlated correlations", I have some questions about the formula<br />1. r: what does "r" mean like r12?<br />2. df=N-3: What does "N" mean? Does N mean the sample size?<br />3. √ [N-3]: What does "√" mean? <br />4. √ 2: What does "√" mean?<br />Chun-Lan Changhttp://www.blogger.com/profile/14343630378681915325noreply@blogger.comtag:blogger.com,1999:blog-8909074830238091680.post-61859015692889431452014-04-24T19:11:39.611+10:002014-04-24T19:11:39.611+10:00If there are three quantitative variables: x, y, a...If there are three quantitative variables: x, y, and z, how to perform dependent correlation analysis? Thanks.Chun-Lan Changhttp://www.blogger.com/profile/14343630378681915325noreply@blogger.comtag:blogger.com,1999:blog-8909074830238091680.post-80327314929300406362014-04-15T18:28:19.303+10:002014-04-15T18:28:19.303+10:00Thanks for this article. I spent about an hour try...Thanks for this article. I spent about an hour trying to figure out scoreItems {psych} using >help(scoreItems) ... 10 minutes after reading this and I finished my script. Brilliant. Thanks again.<br /><br />@ste_nickstenick1989http://www.blogger.com/profile/04791920896372981288noreply@blogger.comtag:blogger.com,1999:blog-8909074830238091680.post-33638238211471711552014-03-28T22:14:44.033+11:002014-03-28T22:14:44.033+11:00Ahh i see.
I meant the subject couldn't perf...Ahh i see. <br /><br />I meant the subject couldn't perform the test ( single leg balance). She couldn't even hold it for a second. The actual tests gets over this by using a scale which gives points from 1- 100. So for unattempted task you get 0 points. But i don't have their scores hence. <br /><br />And you should write an E book. I will buy it. Anoopbalhttp://www.blogger.com/profile/17871276618033328611noreply@blogger.comtag:blogger.com,1999:blog-8909074830238091680.post-69427430088461570572014-03-28T12:37:42.187+11:002014-03-28T12:37:42.187+11:00the standard deviation of component variables is t...the standard deviation of component variables is the same with z-scores. Thus, composites of z-scores effectively weight components equally.<br /><br />I'm not sure what you mean by could not do a test. You could mean did not sit the test, which means you have missing data. Or you could mean failed the test, which presumably means you have an issue of how to score a test where people vary on both how well they did something and whether they could do it at all.<br /><br />With missing data, there are many approaches. One simple one, i just to take the mean score for the data you do have on the person as long as they still have a certain number of scores. This is equivalent to imputing the person mean to the missing data. That said, missing data analysis is a complex issue.<br /><br />If it's an issue of how to score a test where you have some that failed, then that is really going to require some domain specific knowledge. In general, if you want to put all these people on the one scale, then I think at the very least someone who can not do a test at all should generally get a score worse than those who could do it but did it poorly.Jeromy Anglimhttp://www.blogger.com/profile/12949204812496382042noreply@blogger.comtag:blogger.com,1999:blog-8909074830238091680.post-6344594076611047332014-03-28T12:30:29.131+11:002014-03-28T12:30:29.131+11:00Is it because the standard deviation is similar wi...Is it because the standard deviation is similar with z scores ? But when you combine scores by adding the standard deviations could be very different. <br /><br />Also what do you when someone couldn't do the test or scores a zero ( like balance). And this test is time based so if I give zero it means she did it very very good. What do you suggest in this case? Thank you so muchAnoopbalhttp://www.blogger.com/profile/17871276618033328611noreply@blogger.comtag:blogger.com,1999:blog-8909074830238091680.post-89140458551859609252014-03-28T11:43:40.642+11:002014-03-28T11:43:40.642+11:00You are going to get a different variable when you...You are going to get a different variable when you take a composite. Often it will be a more reliable variable. <br /><br />It's certainly possible to get a significant composite and for all the components to be non-significant. <br />There are several explanations for this.<br />The general statistical answer is that if several of the composites are close to significant and in the same direction (e.g., p=.09, p=.1, p=.15) then the combined variable may be enough to be statistically significant at .05.<br /><br />Equally, you could get the opposite where one or more of the components is significant but the composite is not.<br /><br />It's interesting to ponder whether the underlying effect size would be larger for composites or components. In general the increased reliability associated with composites and their more general nature, may lead to generally larger effect sizes. But this is certainly not guaranteed. <br /><br />Another nice thing about the composite is that it provides an overall test. When you have many component variables, you have issues of multiple testing (i.e., with associated increased risks of Type I errors).<br />Jeromy Anglimhttp://www.blogger.com/profile/12949204812496382042noreply@blogger.comtag:blogger.com,1999:blog-8909074830238091680.post-62501164341915715362014-03-28T02:30:31.876+11:002014-03-28T02:30:31.876+11:00I know composites gives an overall assessment of m...I know composites gives an overall assessment of multiple tests. But is there any statistical benefits to using composites?<br /><br />I have noticed in my composite scores, the composite turned out to be significant, but the individual components are not. I used the z scores to calculate composites of 3 variables.<br /><br />Why is this? Thank you so muchAnoopbalhttp://www.blogger.com/profile/17871276618033328611noreply@blogger.comtag:blogger.com,1999:blog-8909074830238091680.post-57102316649863974812014-03-27T17:24:59.071+11:002014-03-27T17:24:59.071+11:00Awesome article! If you haven't yet, you shoul...Awesome article! If you haven't yet, you should check out careerexchange.com.au , it's an Australian job board for allied health professionals, so it's a good place to look for jobs in psychology. It also has tips and tricks that are especially relevant to allied health professions. Psych Presshttp://www.blogger.com/profile/08795617839123104566noreply@blogger.comtag:blogger.com,1999:blog-8909074830238091680.post-18443374036269236942014-03-26T13:04:51.282+11:002014-03-26T13:04:51.282+11:00I'm not really sure what you are talking about...I'm not really sure what you are talking about. I know when people get reaction time data with many observations per participant that they sometimes extract individual quantiles to reflect aspects of each individual's reaction time distribution. This gives you more than just the mean which is particularly useful where the data is non-normal or contains outliers or where you where the variance is relevant and differs between people.Jeromy Anglimhttp://www.blogger.com/profile/12949204812496382042noreply@blogger.com