Cronbach's Alpha - SPSS (part 1)
HTML-код
- Опубликовано: 26 янв 2025
- Conduct and interpret an internal consistency reliability analysis through Cronbach's alpha, the corrected item-total correlations and the inter-item correlation matrix. Guidelines for identifying adequate levels of reliability are also provided.
how2stats.blogs...
Studying for my finals in Belgium, your videos have been so helpful, they are my favorites amongst the stats videos. thank you so much!!
wooh! how is life going now? It has been 5 years!
You are an angel! You will not believe if I tell you that you saved my coursework!
I always come here for a refresher on this subject. Thanks for all your videos and clear explanations
THANK YOU SO MUCH this walkthrough saved my ass on a Final paper I'm writing
LOL =))))))
Now, composite score variance is true score variance + error. So that is total variance. True score variance is only a fraction of total variance. In this case, 65%.
That's a complicated question. Most people would, but it is possible that the correlations between the items would be better represented by a tetrachoric correlation. Bruno Zumbo introduced ordinal Cronbach's alpha. It may apply in your case.
Thank you so much for this excellent video. It is clear and informative.
Thank you. I admire your knowledge and intellect. I recently subscribed to your channel. I work in statistics in education and usually if I am at home I might take a break by watching movies or music videos on youtube. I do have some time set aside time for formal studying in my field but I realised just how fast I can cover material outside of that time by watching your videos instead of music videos etc. I hope to watch all the videos on your channel soon. Your work is very well appreciated.
The only possible nominal scale you could run Cronbach's alpha on would be a dichotomous variable (e.g., yes/no). However, in such a case, the estimate of Cronbach's alpha would likely be an underestimate. If you are ambitious, you should check out "ordinal Cronbach's alpha" introduced by Bruno Zumbo. It in
What is the relationship between reliability and multicollinearity? I'm terrible at stats and I'm trying to figure out if there is any need to check for collinearity before conducting reliability analyses
does Cronbach's alpha work only when you have scale items like in questionnaires?? what if you have a writing test that is to be analysed by criteria when each criterion is scored (0/1/2)? can we calculate the reliability of this test using Cronbach's alpha??please answer me this is urgent
what is supposed to be done when a question asks to compute three new variables for the item and report the internal reliability of the scales. Thank you for help!
Hello! On SPSS how to check validity of 12 item questionnaire with a 7 point-likert scale against a 19 item questionnaire with a 4 point-likert scale? Do I just run a correlation test or do I need to standardize the scales? thank you :)
If Item deleted in reliability analysis, does it effect other analysis?
do i have to re-run compute mean, descriptive analysis, correlation pearson analysis, multiple regression analysis?
Yes, definitely. If you delete an item from the scale, you need to re-run all of your other analyses.
@@how2stats thank you so much for answering. Since my lecturer could not help much. 😂😁😁😁
Hey.
I have 10 questions where I have the responses from 10 respondents on a Likert Scale of 1 - 5.
Do I put the 'Questions' as columns and the responses for each question by all 10 respondents vertically for each question - OR do I put the 'Respondents' as columns and their responses below?
how about Objective type Question (i.e A,B,C,D)? can you advice how to do cronbach alpha when I provide 4 answers that sample can choose, but only one answer is correct. Should I use digit "1" for "true" and "0" for "false" or any other method?
Which one is the relaibility value cronbach's alpha or cronbach's alpha based on standardized item??? Plzzz guide
You would only consult the 'standardized' estimate if the variables you included in the analysis were z-scores, or any score that is based on z-scores (e.g., T-scores).
Thank you for your videos, honestly it is helping students and researchers in the world for their studies, i am phD info. system student in Malaysia, and i have a small problem in my data analysis, actually my model is about cyber criminology (4 constructs - 3 of them independents with reflective indicators, and the last one is dependent with formative indicators), the problem that 3 constructs are yes/no questions and the last is likert scale with 5 points, i want first to make the pilot test.
1. is it possible to test the reliability of them and how?
2. is it possible by smartpls? or spss?
i am confusing because no body give me the answer, the data was collected from another country (college students) and it is very difficult to recollect. please help me
waheeb abu-ulbeh You can use Cronbach's alpha with likert 5-point likert scale data in SPSS; I don't know about smartpls; when you have yes/no data, it's tougher; people would argue you should probably use polychoric correlations in that case, but SPSS doesn't do that; most people just run Cronbach's in SPSS when dichotmously scored data; alpha will typically be underestimated in that case, but at least it's an estimate.
Can you please elaborate how we can check the test of significance?
That figures!
So the Cronbach Aplha will test the consistency of questions.
I just did it and I get it.
Thanks!
Most people try to achieve alpha .70 or higher, but there is nothing magical about that number. You can still use a composite with alpha .50, but the chances of detecting an effect are lower. If I were you, I would have a good look at your item inter-correlations. Are there several items that should not be in the scale? Or maybe the scale is just not very good and needs more work.
4:20 - 65% of the variability in the composite score is true score variance. Could you explain this a bit more please? Are you referring to true score as in true score in test theory ie. true score = observed + error. In that case are u saying that 65% of the variance that we observe in the total score is ACTUALLY variance in the true score ?
Hi, do you have any resources on how to write up a results section for a lab report on a cronbach's alpha and justify a deletion of an item? Specifically for a dichotomous scale. Thank you.
You can also recode correct to one, incorrect to zero and use alpha then.
Great! Is it possible to have the databaseused in the presentation?
Hi, just a quick question, do you have to have a range of responses on a scale to be able to employ Cronbach's alpha or can you conduct it successfully on a questionnaire where the responses were either correct or incorrect? Thanks!
hi, want to ask whether nominal scale required to run reliability test or not? if yes, is it using the same way?
What does this warning message mean, and how do you deal with it?
The determinant of the covariance matrix is zero or approximately zero. Statistics based on its inverse matrix cannot be computed and they are displayed as system missing values.
When I ran the alpha test for each individual section of my survey all alpha scores were good or excellent, and the inter-item correlations were good too (virtually all above .3, though a handful just below in one section of the survey). Also, the alpha score if item deleted were all good too (except for just one) -- if any one were removed the alpha would drop.
Any advice?
hi im a news subscriber! if i may ask, do you have videos or tutorial or blogs about measuring cronbach alpha using spss but with SME? Im currently working on my thesis ryt now and i dont have much resources due to pandemic. My study is about consumer behavior using Theory of Planned behavior. Glad if you could help. Ty!😄
What does SME stand for?
@@how2stats Structural Equation Model sir
I have a very important question to me. I am writing a master theses in marketing. I am using 7 point Likert scales. Lets say, I want to interpret a 10-item seven-point Likert scale (after I check Cronbach's alpha). How in SPSS can I get the average score of all the 10 items, that correspond to one characteristic of the consumers? I can get the means of all 10 items separately. But I want to evaluate the sample in terms of their total mean score of the characteristic. Thank you! Monika
Hi when do we use Cronbach alpha or Intraclass correlation? and are there differences in the interpretation of these two? Thank you in advance.
Shane Manat People seem to use ICCs when examining inter-rater reliability and Cronbach's when examining internal consistency of items/subtests. As a general statement, they will give you the same estimate (when certain conditions are satisfied).
Thank you very much.
can i use this if i do item analysis in a 50-item test questions using multiple choice questions (4 choices per item)??/
Yes, the 4-choices per item is not an issue. However, Cronbach's assumes that the items all measure the same dimension (and only one dimension). So, if the 50 item questionnaire measures several susbcales, you'd want to estimate the reliability of the subscales.
Hello, quick question, what should I do if I have items for example "I do NOT xy" wich should score negatively on the scale?
can you use cronbach alpha if your scales are different e.g one is 1-4 likert and one is 1-2 likert
Yes, no problem.
I'm so thankful for your videos! so helpful!
what about the scales with reverse items?
You have to reverse score them, before calculating Cronbach's alpha.
You would be the questions as columns and respondents as rows.
Does Google sheets run Cronbach's alpha tests?
I don't think so. But it can calculate Pearson correlations, so, if you calculate the average inter-correlation between your items, you can use the formula I show at 30 seconds in this video to estimate Cronbah's alpha: ruclips.net/video/EdCdTzpZrVI/видео.html
excuse me, i wanna ask you if my reliability test number is always under 0.3. what should i do. Do i need delete some numbers?
Hello, is this two tailed
what if there are reversed questions(with negative meaning) in our scale???what could we do??shall we use them too in the reliability analysis?
You must recode the reverse items before using alpha test.
Hi, this may sound like a bit of a stupid question (massive newbie) but do you run this analysis after you have conducted your initial exploratory factor analysis?
nope, what I do is run cronbach alpha first, then run correlation matrix, KMO index, Bartlett's test of sphericity, and then finally PCA.
Can i use Crobach's alpha to check my physics Multiple choice questions' reliability?
you would need all of your class mates' scores as well; you can't calculate reliability for test scores based only on one person's responses.
What happens if some need to be reverse coded please?
You definitely have to reverse code negatively keyed items, before conducting Cronbach's alpha. I'll have to make a video on how to do that. Stay tuned.
Thank you for your video. It's really helpful. I have a questions here.. Can I use cronbach's alpha for scale with 'yes' or 'no' answer?
Yes, in the sense that many people do. But it most likely would be more appropriate to conduct the analysis based on a tetrachoric correlation matrix, rather than Pearson correlations. You can't do that in SPSS (at least not easily), so you're pretty much stuck with Cronbach's alpha as demonstrated in the video.
if my cronbach's alpha is 0.4 what does mean 0.5 , 0.6 , 0.7 , 0.8 ?
Hani Albarni It' not high enough, it should be around 0,8
Crobach's Alpha Level of Internal consistency
α ≥ 0.9 Excellent
0.8 ≤ α < 0.9 Good
0.7 ≤ α < 0.8 Acceptable
0.6 ≤ α < 0.7 Questionable
0.5 ≤ α < 0.6 Poor
α < .5 Unacceptable
can anybody tell me how to calculate the scores for a Mental toughness 48 inventory (MTQ48)???
Thanks very much for your help! I was a bit confused in the beginning as my scales don't have that (EOT) but I managed to do it anyway! :)
I know im asking randomly but does any of you know of a trick to log back into an instagram account??
I stupidly lost the password. I would love any tricks you can offer me!
@Cruz Jefferson instablaster =)
@Mason Leonardo I really appreciate your reply. I found the site thru google and Im waiting for the hacking stuff now.
I see it takes quite some time so I will reply here later when my account password hopefully is recovered.
@Mason Leonardo it did the trick and I now got access to my account again. Im so happy!
Thank you so much you saved my ass :D
@Cruz Jefferson Glad I could help :)
You videos are great! Could you please post an ROC curves tutorial?? I'm having so much trouble with this!
Please, can you make another video on how to conduct validity test using SPSS, please. Thanks in advance
thank you
kindly where can i download a database like this to practice
Thank you very much for the video, it is very informative indeed.
This is great! Very thorough :)
Im mad doing research right now. Can somebody teach me. 😭😭
Same.. I'm New to research also
Super useful! thanks very much
Thank you great explanations
Great stuff, thank you!!
Thanks so much in deed...
Awesome explanation
Really thank you
thanks a lot great help...
great help, thank you very much ;)
Thanks! Now im Cronbachs alphaing the shit out of my data!
Thank you
thanks you!
mantap gan
it is pronounced CRONBACH not cronback! Please and Thank you!
shutdown
cayalah...syamim comel
You sound like Dave Grohl
Hello and thanks. In a control group experiment, would one ever conduct Cronbach's for a DV separately for each condition? If so, why?
+31415926535equalspi There is an analysis known as 'reliability generalisation', which tests the hypothesis that Cronbach's alpha is equal across groups. However, in practice, researchers estimate Cronbach's alpha for the whole sample (groups collapsed) and simply assume that the overall Cronbach's alpha estimate applies to all of the groups (e.g., treatment, control, etc).
Thank you very much for your response. Very helpful.
thank you very much!
thank you!