I'm in the dissertation phase of my PhD and not one professor has shown me how to do this, not even in my stats classes! Thank you SO much for this video!
@@nnodatulomah4312 Thanks for your comment. The video clearly explains how you can transform Likert scale dataset into continuous variables. I encourage you to watch the video again carefully. Thanks
Hello thank for the tutorial I have a problem Please I try to evaluate thé motivation of Drinking on some participant and for it a use the« DMQR-SF » which is a questionnary of 12 questions and each questions hav 5 items with point ( never = 1 pt, somestimzs =2 pt , half of times =3 pt, much of times =4 pt , everytimes =5 pt) , and each questions Belong to one of the 4 main reason Why thé participant drink alcohol ( expl: questions 1,5,8>social reason; question 3,6,10>Forget problem….etc at the end of questionnary we sum the point of participant in each main reason and thé main reason which has thé Highest score its thé reason Why thé participant drink alcool Now i try to do it on SPSS, i Créate variable for each main reason , i sum , i create another variable who show me the highest score between thé 4 with command » Max « but thé way is that this last variable dont show me Where thé highest score come from…i am force to go and see thé other 4 variable to see where it comme from it really not good because i have more than 1000 participant… it will be well if you have a solution who show me where thé highest score Côme from so that injust make a fréquency table I Hope you can help me Thank you in advance !
Greetings please. Your videos are commendable. thanks. I have a challenge. How can i perform the same statistical analysis on a data with missing values on the grouping Likert questionnaire items?
ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-zpnJmSVvgJA.html This video will help you to sort this out. Also, you can remove missing information before you analyze your preferred statistical analysis.
Hi Dr Okolie, thank you for the explicit tutorials. With your videos, I have evade enriching someone on something I can now do effortlessly. Please, I would like you to make a video on one-way ANOVA, to complement the once you have done already. Thank you.
Hi, thank you so much for the tutorials. It is very useful. Can I also ask you one thing. I thought it would be ordinal data instead of scale in the Measure section in SPSS. Can you explain a bit here. Thank you very much
Sir I have a question, what do I do if my questionnaire have two part likert scale questions (first part is disagree-agree scale and the second part is never-always scale) but all those will be used for one variable only. How do i check its reliability and how do I group it?? I would very much appreciate your help
Thank you. Whatever you call your rating scale (e.g., Strongly agree, always, never, not about me, etc.) doesn't matter. You can check the reliability coefficients following my tutorial. Also, you can group them to continuous variables as I did in the video. Best wishes.
Hello! I rank the A 1, B 2, C3... For the grades of school children and their habit of breakfast are 1 for every day, 2 for most days, 3 for rarely.... 5 never... This ranks are correct? it right to find the correlation...?
thank you once again for this impactful lecture, Dr Ugochukwu. please, i will like to ask if you have a video on how to answer research questions using regression analysis. if you do, kindly let me know and if you dont, could you do one?
Thanks for watching the tutorial video. Pls, like and share. Regarding your question, I feel that research questions are better answered using descriptive statistics. You can check my video in Likert Scale data analysis. Regression analysis is better conducted to test hypotheses. You can check my video tutorials on multiple regression analysis too. They're well detailed. Once again, thank you for finding the video helpful.
I have watchted several of your videos and have subscribed and shared accordingly. I must also say here that i actually benefited so much from one of your videos on regression analysis. However, i want to ask you again: Can i use regression to do data analysis when i have multiple independent sub-variables and multiple dependent sub-variables as well? or what is the best statistical tool to use in that situation? Waiting for your kind reply. Thanks@@Dr.Okolie
@@EdutObona Thank you for enjoying the videos. Yes, you can have dependent and independent variables with several sub-constructs. For example, the regression analysis videos showed that the independent variable (support) was decomposed into three sub-constructs of Parent, friends and teacher support. Same can also be done for dependent variable. You can have both variables decomposed into several other sub-constructs in one study as long as your motives are theoretically and practically justified.
Thanks for watching the tutorial and finding it helpful. You may also watch this just in case you want to conduct some sets of parametric tests: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-kEUf03LHiXI.html ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-_GbIBtg5yG0.html
Amazing thanks. I have a problem interpreting the data output. Do you provide private help about that? Like can I share my data output and help me write conclusions. Just thought I'd seek help since am watching tutorials for a week now (cry)
Your explanation was very clear and demonstrated your dedication. Thanks to your help, I was able to understand and complete the analysis on my own. Thank you so much!
Thank you so much Sir. Being a scholar of Agricultural Extension Education in India I have been struggling since the last few weeks for analysing likert scale data using SPSS as there is no course that gives insight to it. It's a life saving lecture video. I appreciate your efforts.
This video has really helped me so much during my research paper. i watched most of your videos on SPSS and i couldn't have done it without these videos. thank you
Thank you very much for finding the video helpful. Pls, share and like the video. You may also find this video ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-_GbIBtg5yG0.html helpful
I have a concern about how the instrument is calibrated as a measuring tool and zeroed. Again the instrument maximum scores it can measure. The illustration described in this video has a few issues of concerns, the undecided has code better than a statement response better than undecided so how do we reconcile this in terms of values assigned?
@@nnodatulomah4312 Thank you very much for expressing your concerns. Pls, use my contact on the tutorial to WhatsApp me. We can talk about your concerns for clarity.
That is my concern I flagged on another video by Dr Okolie: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-uHteOFO84MM.html The "undecided" seems to be causing some serious confusion leading to what I may call "wrong" interpretation and results. In my opinion, "undecided" should carry no weight on the scale, thus it would be better to either use "neither agree nor disagree" or remove it from the Likert scale. Otherwise the result are unreliable.
@@Luckisto Thank you for your comments. What is the meaning of "neither agree nor disagree" as you recommended in your writing? When somebody didn't agree or disagree on something, what does it mean?
Thank you very much sir on this. I really really appreciate. Please I have a question. After the whole grouping and stat analysis how to I perform the test of Hypothesis, like the actual test of the hypothesis? Is there a video please kindly attach link to the reply
Thank you very much for finding the video helpful. You can watch this for hypotheses testing ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-kEUf03LHiXI.html
Dear Prof., I have a question. Even after calculating the mean of individual items of likert scale variables, it's not normally distributed(I've found it in my dataset). So how can we run the parametric tests (Pearson correlation, Linear regression) with these mean scores, where the primary condition of running parametric test is having normally distributed variables? Please reply.
I will suggest a few things. First, you may transform the data with a function to force the data to be normally distributed. Again, if your sample size is relatively small, you may consider running a non-parametric analysis. Second, you may trace the actual respondent whose responses are causing the data not to be normally distributed, and then, remove it from the dataset. This option may apply if you have a larger dataset. I'll recommend you watch ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-TeL_Zke8Crk.html by Math Guy Zero. His tutorial explains how you can deal with a dataset that is not normally distributed. Hope this helps.
You're right sir. My dataset is small. That's why at first I performed Spearman correlation, but found no significant results. All of my hypothesis were rejected. Then after watching this video I performed Pearson correlation & linear regression on my likert scale variables, and found statistically significant results. 3 out of 4 hypothesis were accepted this time. Thanks a lot to you sir. My concern is that, how I'll justify it to my thesis? Is this analysis valid? I need your opinion sir, plz.
Thank you very much sir. I have one question If i have one more variable gender or qualifications of the respondent along with this data.and i want to run chi square test for this variable gender with all 10 items then how to do it?will you please explain this?