i'm writing my bachelor's dissertation now (BA, so you can guess i have no stats experience) and since im working with quantitative data, i suppose i felt pressured to learn some spss to do my analysis. for the past 2 weeks i've been watching youtube videos explaining how to run just about every test. well, halfway through analysing the data for my final experiment, i realised that it wasn't normally distributed and had a mini meltdown (i've only just wrapped my head around a pearson's correlation, and now i had to learn ANOTHER one?). but i can say that this is the clearest, most in-depth stats tutorial i've come across. thank you for explaining what everything means so beginners (like me) can understand easily!
@DrPC_statistics_guides4 жыл бұрын
Thanks!
@xxxmelissatacionxxx2 жыл бұрын
Very simple and straight to the point. Thanks!
@funphysicsathomebygisellea32912 жыл бұрын
Thank you for this video Prof. Christiansen. It helped me a lot!
@veen85262 жыл бұрын
Very helpful thank you
@donnarosem.sampong2362 жыл бұрын
Thank you for this video, Dr. Christiansen. This is the best explanation. By the way, may I know where I can get the data for this, please?
@donnarosem.sampong2362 жыл бұрын
For educational purposes only, if it is okay with you. Thank you so much.
@kossonouprunelle75762 жыл бұрын
Thank you for the presentation. I have both nominal and ordinal independent variables (4) with (2) ordinal dependent variables . Please which one should i choose between Spearman correlation and Chi-square test to analyze the association or relationship between the variables ? Or can we do both of them in the same study? Thanks
@jajejijoju01273 ай бұрын
Hello can you help me interpret the result of our test?😢 I dont understand and i dont know how to put it in a table
@Ritu-ey6nv3 жыл бұрын
Hello, thank you for your video. I am trying to find the relationship between Happiness (Ordinal variable with more than 5 categories) and Income (continuous variable). In this case, I am having trouble deciding if I should run Pearson correlation (by treating the happiness variable as continuous since it has more than 5 categories) or whether run a spearman's correlation? Also, while I tried to check the assumption for Pearson correlation (treating happiness as continuous), I found that Income is not normally distributed. And when I generated a scatterplot for these two variables, the result is not clear and somehow visually indicates no linear relationship. What should I Do???
@Laylabarud3 жыл бұрын
Hi Paul, I am trying to test whether the change in the trend of indirect maternal deaths in a specific country is statistically significant. I have four-year data sets but I am wondering what statistical test fits well here? Any help is highly appreciated. Thanks
@katy87912 жыл бұрын
Hello Nive video! As I have seen that Spearman rho correlation has 3 conditions and one of them is that the two variables should have a monotonic relationship... if the function of the scatterplot btw the 2 variables is not monotonic, can i still use spearman rho? Please help!
@train97753 жыл бұрын
Hi! I was wondering why the significance is not "highly significant" since it's 0.01 level like that of Pearson's?
@nelhuens75902 жыл бұрын
Is it possible to use this Spearman's correlation if the two variables are continuous (scale) and none are ordinal? Our data is not normal so we can't use Pearson's. Thank you!
@DrPC_statistics_guides2 жыл бұрын
Yes!
@musicevolution96823 жыл бұрын
Hi sir, which one should i use to know the relationship between knowledge-attitude, Attitude-practice, knowledge-practice. I saw in journal they use linear regression, some use spearman correlation coefficient so which one should i use. Or i can just put both in my manuscript
@DrPC_statistics_guides3 жыл бұрын
Regression is used when you have one specific outcome to predict from other variables. Correlation just looks at associations between two variables. If you are interested in the interrelationships between your three variables do a Correlation. If there is one of those variables you wish to predict then use a Regression
@izzahmora20923 жыл бұрын
hi. im trying to check whether there is an association of knowledge & attitude on preventing back discomfort. both of my variables are ordinal. what should i use?
@DrPC_statistics_guides3 жыл бұрын
Spearmans correlation unless your sample is small (less than 30) then use kendalls tau
@HoribleHistoriesRoks2 жыл бұрын
hi my spearman's is at -.21 and Pearsons is at -.177 does this mean that it isn't linear data and I have to use another test to see the relationship between them? the data are both nominal
@DrPC_statistics_guides2 жыл бұрын
if nominal data you cannot use a correlation. You need a chi square analysis
@edwardswazimhlongo4294 жыл бұрын
Good day, I'm trying to check if there is an association between gender and responses to perceptions about an invasive alien plant. I.e. if there's any relationship between gender and the perceived negative or positive impacts. Which statiscal test would be suitable for this?
@DrPC_statistics_guides4 жыл бұрын
Independent samples t test or mann Whitney test would probably be your best choice. Your IV will be gender. Videos for both can be found on my channel
@edwardswazimhlongo4294 жыл бұрын
@@DrPC_statistics_guides thank you is there a way to contact you directly if it is not too much to ask. I am kind of stuck. I just need you to have a glance at the data that's all it won't take too much time.
@skolen20104 жыл бұрын
I dont understand why did you write 38? I mean what is the reason for 40-2??
@DrPC_statistics_guides4 жыл бұрын
The degrees of freedom (the number that appears in the brakets) is calculated as sample size - 2. So in this case the N is 40 so we need to subtract 2 from that to give our degrees of freedom of 38. The correlation coefficients along with the df is how the p value is derived
@lexie85164 жыл бұрын
@@DrPC_statistics_guides Is this something that can vary? or does it have to be -2?
@DrPC_statistics_guides4 жыл бұрын
@@lexie8516 dfs in correlations are always N minus 2. However sometimes you may see people report N= in the brackets instead. However, generally speaking you should report the dfs of N-2
@pomaelo91674 жыл бұрын
Why do you have to minus 2 in the n? And what does negative correlation means, i mean they are still significant but...?
@DrPC_statistics_guides4 жыл бұрын
The df is often reported as N-2 (although sometimes people just give N). A minus correlation means as one variable increases the other decreases, i.e. a negative association. E.g if you had a data set were you found that as people got older they drank less alcohol there would be a negative correlation coefficient
@volcan_ism34409 ай бұрын
thanks !
@paulinauznanska93093 жыл бұрын
Where can I find data for this exercise?
@DrPC_statistics_guides3 жыл бұрын
Sorry I don't have the data uploaded for that exercise (and it's on my computer at work which is currently closed) apologies
@divyamandesa854 жыл бұрын
How do we correlate nominal and ordinal data?
@DrPC_statistics_guides4 жыл бұрын
Ordinal data can be correlated with Spearman's or Kendall's tau. However, because nominal categories are arbitrary (you can switch how you label them) you cannot really correlate them.