Visit the channel for more on psychology, psychometrics, psychological statistics, and research methodology. Subscribe to provide 'positive reinforcement' and press the 'bell' icon to express your 'unconditional' support!
Thank you very much. Very well explained. Congrats. My question is about a real case when you have maybe 5 sections in the questionnaire, do you calculate the reliability coefficient for each section. What about follow up questions (open-ended questions)?
Hi Phil John. Thanks for watching. Yes, if you have different domains/factors/sections in your test, you should compute different internal consistency reliability coefficient for each. For example, if you have 5 domains/sections, then 5 different reliability coefficients. And if your test is not a knowledge or speed test, and has Liker scale, I would recommend you to go for Cronbach's Alpha reliability. In case of dichotomous items, KR-20 is an option. However, in all cases, Split-Half reliability will be a good addition for the psychometric properties of your scale.
And yeah, about the follow-up questions, you can report those as qualitative inferences which would be a valuable qualitative addition to your scale. Also, these can be used as feedback too for further refinement of the tool.
Hi sir, in the video you mentioned that the SBC>0.7 is acceptable, else it is not. May I know is it a Table of Spearman-Brown Coefficient to be referred?
Hello dear, No, there's no table to refer for Spearman-Brown Coefficient and it's not a correlation coefficient to see it's significance value by refering to critical values table. Generally, Spearman-Brown Coefficient value above 0.7 indicates an acceptable internal consistency reliability for a test. If the SBC value is > 0.8, it is considered as good international consistency reliability. SBC > 0.7 is the most frequent recommendation including the recent study by Mokkink et al., 2017.
@Shivanand Thorat Hello Thorat ! Nice video👍 I wish to know when do we use split half reliability is it when we have a small sample less than 30 so instead of using the normal cronbach alpha we use split half in that case?
Hi Katy, split-half reliability is generally used when there are large number of items in the test or if the test is timed. Split-half has nothing to do with the sample size as such. But it's always desirable to have sample size greater than 30. Hope this helps. Thanks for watching. Consider subscribing to the channel if you find the content helpful.
@@ShivanandThorat Hello Thorat🙋🏻♀️ thanks for your reply🙂 i wish to know what did you mean by test is timed? U mean repeated? Also, am I allowed to use cronbach alpha if i don't have another sample like a pilot sample?
By the term timed test I mean that the test has a time limit to complete. Cronbach alpha can be used even if you don't have pilot testing done before. In fact it can be used as a tool for item analysis. Please watch my another video on item analysis using Excel and SPSS to know how to use it for item analysis. In case of queries, feel free to write to me on shivanandthorat.yt@gmail.com
@@ShivanandThorat Thanks a lot Thorat 🙂 so if i am doing a pretest and a posttest in my research is it enough to show that cronbach alpha is 0.7 for the pretest or i should also do that for the posttest? Also, thanks for your contact details... what about shivanandthoratresearch@gmail.com can i contact you on that mail too?
No problem Katy. And you can contact me on shivanandthoratresearch@gmail.com. You can use any set that is pretest or posttest, compute cronbach for the one with larger sample size. If sample size is equal then go for any (pre or post)
There is as such no table for split-half reliability values. But generally any value above 0.7 is considered as acceptable and any value above 0.8 is considered as good. I hope this helps.
Hello madam, I didn't get your question. Accurate as in we are using software, values will be accurate only unless you make any error in the formula that is employed in the Excel or SPSS. You can compute manually and check if the values appearing in the software are accurate or not
I'll make it in near future. Thanks for watching. Interpretation of the split-half reliability coefficient (Spearman-Brown Coefficient) is identical to other internal consistency reliability coefficients. SBC > 0.7 is acceptable, SBC > 0.8 is good. Higher the coefficient, better it is.
You mean first 16 and second 17 items? There are 33 items to be split into two halves. So you can use odd even split which will result into 16 and 17 items in two halves. And then find correlation and follow further procedure as shown in the video. SPSS provide two Spearman-Brown Coefficients, one for equal halves, the other for unequal halves. You you will interpret unequal halves in this case. Hope this helps. Consider subscribing to the channel if you find content helpful. Best, Shivanand Thorat
lots of thanks......you provided very useful and benefial ways of odd even split half method which no any tutorial taught....
I am glad it helped you. Thank you for watching, subscribe the channel if you find content helpful
Very Informative.
It would be very fortunate to follow this man for psychology academics !!!
Very much useful !!
Very helpful!
Visit the channel for more on psychology, psychometrics, psychological statistics, and research methodology.
Subscribe to provide 'positive reinforcement' and press the 'bell' icon to express your 'unconditional' support!
This really saved me!
Thank you very much :)
Thanks for watching. Subscription to the channel would be amazing.
this is very helpful, thank u
Glad it was helpful!
Nicely explained... Thank u
Glad you find it helpful. Thanks for watching.
Good one brother 🔥
Very helpful! Thankyou!
Thanks for watching. Consider subscribing to the channel if you find content helpful
Thank you very much. Very well explained. Congrats.
My question is about a real case when you have maybe 5 sections in the questionnaire, do you calculate the reliability coefficient for each section. What about follow up questions (open-ended questions)?
Hi Phil John. Thanks for watching.
Yes, if you have different domains/factors/sections in your test, you should compute different internal consistency reliability coefficient for each. For example, if you have 5 domains/sections, then 5 different reliability coefficients. And if your test is not a knowledge or speed test, and has Liker scale, I would recommend you to go for Cronbach's Alpha reliability. In case of dichotomous items, KR-20 is an option. However, in all cases, Split-Half reliability will be a good addition for the psychometric properties of your scale.
Wow, some people are so available for others. God bless you, Sir!
May I have your email if you don't mind?
And yeah, about the follow-up questions, you can report those as qualitative inferences which would be a valuable qualitative addition to your scale. Also, these can be used as feedback too for further refinement of the tool.
@@philjohnk.7129 Mention not John. shivanandthoratresearch@gmail.com is my mail ID. Feel free to write.
Thanku so much n god bless u
Thanks for watching
Hi sir, in the video you mentioned that the SBC>0.7 is acceptable, else it is not. May I know is it a Table of Spearman-Brown Coefficient to be referred?
Hello dear,
No, there's no table to refer for Spearman-Brown Coefficient and it's not a correlation coefficient to see it's significance value by refering to critical values table.
Generally, Spearman-Brown Coefficient value above 0.7 indicates an acceptable internal consistency reliability for a test.
If the SBC value is > 0.8, it is considered as good international consistency reliability.
SBC > 0.7 is the most frequent recommendation including the recent study by Mokkink et al., 2017.
Nice explanation
Very helpful 👍👌
Thank you so much!!
Sir will to tell me through manual that how we check the reliability of likert scale through split half method
Hi Preeti,
Check psychometric properties section in the manual for reliability of the scale. Not all scale manuals mention split-half reliability
Really helpful! :)
Glad you found this useful, thanks for watching
@Shivanand Thorat Hello Thorat ! Nice video👍 I wish to know when do we use split half reliability is it when we have a small sample less than 30 so instead of using the normal cronbach alpha we use split half in that case?
Hi Katy, split-half reliability is generally used when there are large number of items in the test or if the test is timed. Split-half has nothing to do with the sample size as such. But it's always desirable to have sample size greater than 30. Hope this helps.
Thanks for watching. Consider subscribing to the channel if you find the content helpful.
@@ShivanandThorat Hello Thorat🙋🏻♀️ thanks for your reply🙂 i wish to know what did you mean by test is timed? U mean repeated?
Also, am I allowed to use cronbach alpha if i don't have another sample like a pilot sample?
By the term timed test I mean that the test has a time limit to complete.
Cronbach alpha can be used even if you don't have pilot testing done before. In fact it can be used as a tool for item analysis. Please watch my another video on item analysis using Excel and SPSS to know how to use it for item analysis.
In case of queries, feel free to write to me on shivanandthorat.yt@gmail.com
@@ShivanandThorat Thanks a lot Thorat 🙂 so if i am doing a pretest and a posttest in my research is it enough to show that cronbach alpha is 0.7 for the pretest or i should also do that for the posttest?
Also, thanks for your contact details... what about shivanandthoratresearch@gmail.com can i contact you on that mail too?
No problem Katy. And you can contact me on shivanandthoratresearch@gmail.com. You can use any set that is pretest or posttest, compute cronbach for the one with larger sample size. If sample size is equal then go for any (pre or post)
How to calculate reliability index test using split half methods if given 8 items and total score in percentage
Is there any table for reliability value
There is as such no table for split-half reliability values. But generally any value above 0.7 is considered as acceptable and any value above 0.8 is considered as good. I hope this helps.
How we can know the reliability value is accurate or not
Hello madam, I didn't get your question. Accurate as in we are using software, values will be accurate only unless you make any error in the formula that is employed in the Excel or SPSS. You can compute manually and check if the values appearing in the software are accurate or not
Can you make another video how to interpret the result?
I'll make it in near future. Thanks for watching.
Interpretation of the split-half reliability coefficient (Spearman-Brown Coefficient) is identical to other internal consistency reliability coefficients. SBC > 0.7 is acceptable, SBC > 0.8 is good. Higher the coefficient, better it is.
Namaste sir is SPSS necessary to calculate split half odd even reliability ?
No. You can do it on Excel
Thanks a lot sir ,if I have any doubts will ask later will do in excel
Sure!
If we have 33 questions for 10 respondents , can we seperate first 11 and second 10 for average value and there by correlation?
You mean first 16 and second 17 items? There are 33 items to be split into two halves. So you can use odd even split which will result into 16 and 17 items in two halves. And then find correlation and follow further procedure as shown in the video. SPSS provide two Spearman-Brown Coefficients, one for equal halves, the other for unequal halves. You you will interpret unequal halves in this case.
Hope this helps. Consider subscribing to the channel if you find content helpful.
Best,
Shivanand Thorat
thanks idol, bisita ka dito minsan samen. pameryenda ako
Could you please write in English?
Sir, I saw your address, but another video came in, and now I can find back your e-mail. Sorry, I'm not really gifted in these technologies.
No problem. Technology is developing exponentially. We all learn with our need and pace 😊
Actually I'm asking your email again
Oops, I thought you found it back. Anyway. It's shivanandthoratresearch@gmail.com
I sent you an email