Statistically Testing Validity of a questionnaire Using Pearson's Correlation Coefficient. (Part 1)
ฝัง
- เผยแพร่เมื่อ 12 ก.ย. 2024
- The validity is probably the best way to judge the theoretical soundness of an instrument. In this video, I have presented one very simple way of testing the validity of the items of the questionnaire using Pearson's Correlation Coefficient. This is the first video of validity testing. In the coming videos, I will be introducing you all with the other statistical methods of testing the validity of the questionnaire.
For more videos on SEM with AMOS kindly visit below-mentioned links
How to transfer data from a questionnaire to the SPSS file?
• How to transfer data f...
Structural Equation Modeling (SEM) and path analysis using AMOS
• Structural Equation Mo...
How to perform Confirmatory Factor Analysis (CFA) Using AMOS?
• How to perform Confirm...
How to check the reliability of a scale/questionnaire using SPSS?
• How to check the relia...
For a theoretical explanation Kindly Visit Dr. Seema Khan's Educational Lectures
Characteristics of a Good Test / Tool
www.youtube.co....
www.youtube.co....
Reliability - Types of Reliability (Part 1)
www.youtube.co....
www.youtube.co....
www.youtube.co....
www.youtube.co....
www.youtube.co....
Validity-Types of Validity
www.youtube.co....
www.youtube.co....
you can reach me at dradilzia@gmail.com
#SPSS #Datafile #Questionnaire #dataentry #SPSSfile
#reliability #SPSS #construct #Chronbach's #dradilzia@gmail.com #validity #pearson correlation #adilzia
~-~~-~~~-~~-~
Please watch: "One way ANOVA and Hypothesis test using SPSS. "
• One way ANOVA and Hypo...
~-~~-~~~-~~-~
Simple and understandable explanation. Great teacher.
Thank you for explaining understandably regarding the statistically testing validity of a questionnaire.
Your explanation to validate all the items of a questionnaire is really communicable in a very easy manner. Appreciate your CLEAR CONCEPTS as well as your COMMUNICATION SKILLS.
Its great that you liked the explanation.
Thank you so much for the content it is helping a lot of students around the world your work is much appreciated
Excellent Explanation Dr. Adil Zia. You made it clear to me within 8 minutes which my guide could not do it in 1 year.
Sir, I have got a question to ask you, let's say if my obtained value is greater than critical value, but significance is more than .05, in this case, is that particular variable is valid?
If It is not valid first of all there is no question of significance level. First check the validity and then look for significant value.
Thank you for this helpful video! Please guide me on how to handle negative statements while checking the validity and If we have 5 factors in a single construct then how to check the validity and how to report it.
Thank you very much, Dr Zia, It was really simple to understand.
Glad to hear that
Thank you so much for your video. It really helped me a lot in my dissertation for quantitative methodology and analysis.
This is very clear and very helpful. Thanking you. Appreciated. Great job.
1st time I understood statistics so easily!😂💐♥️
I am happy to hear that. Keep learning.
thanks for this valuable video it helped me a lot, please keep uploading a video regarding how to assess the reliability of the questionnaire as well thank you
Sure I will upload more such videos. These days I am busy with a project. Till that time you can read my latest publication www.emerald.com/insight/content/doi/10.1108/IJQRM-09-2021-0335/full/html
Hi sir, good explanation. how can we make the table of critical values??
how about the Cronbach alpha to test the validity of the test?
Very informational 👌
Hi Dr. Adil
May I know what type of validity for this statistical test that you are showing? Construct validity, criterion validity or content validity?
This is Convergent validity. the other methods are the Rho and AVE calculations. I hope this answers your question.
Very useful video Sir. I have a small question. What type of validity (content/criterion/construct validity) is measured for a questionnaire by Product Moment Pearson Correlation in SPSS?
It is construct validity
It's a very useful video. Thank you a lot for preparing it for us. When the calculated values for Pearson's correlation coefficient are greater than the critical values, the question/item is valid. Sir, could you please share relevant references with us?
There are a lot of references you can search on Google scholars relevant to your context. the one I am sharing for your convenience is doi.org/10.1177/1460458220975
@@adilzia Thank you for sharing the reference. But I could not open the link. It would be highly appreciated if you could please share the once again.
@@adilzia Thank you sir. But the link is not working.
@@md.golamkibria9397 Validity and reliability testing of the
Indonesian version of the eHealth
Literacy Scale during the COVID-19
pandemic.
DOI: 10.1177/1460458220975466
@@adilzia Thank you very much Sir. In the meantime, I also found a similar article. For the benefit of readers, I have added the article here. Translation validity and reliability of the Indonesian version of the 5-item International Index of Erectile Function (IIEF-5) (doi:10.5152/tud.2021.21185).
Thanks, sir, I have learned a lot from your video, But I have a query. I have five different variables My sample size is 50 for pilot testing. The df is 50-2=48, I have computed the value at 0.2787. Is that correct? In my spss output table, it reflects two asterisk signs in the total column(**), and at the end of the table the footnote is ** correlation is significant at the 0.01 level(2- tailed)
Very well explained
Very nice explanation sir. Sir, I just want to know is validity of questionnaire can be measure only if it contains scale questions? or it can be done for all type of close end questions
Thanks a lot. Yes, validity can be checked for all types of closed ended questionnaire.
@@adilziaThank you very much sir
Thanks for the sharing sir. How do we go about if we have negative items. Should I reverse code and then do the validity analysis? Please advice, thanks.
Sir, may i ask.. where do i find the literature or reference that explain the same thing just like you did at the video ? About the convergent validity. It is hard to find, the literature that explain the same thing like you did. Thanks sir. So much appreciate.
you can search for reference papers on web of science or Scopus databases. good luck
@@adilzia First of all, thank you so much sir for the respond. really fast respond. But still, i cant find any literature that do the same procedure as you did at the video. Would you recommend me some literature specifically explain about the same thing just like you did at the video sir? I would really appreciate it. Much love for you sir. Ur video did me so much help.
@@jordainrt7392 you can get some understanding by reading these articles link.springer.com/article/10.3758/s13428-012-0289-7
journals.lww.com/anesthesia-analgesia/Fulltext/2018/05000/Correlation_Coefficients__Appropriate_Use_and.50.aspx
@@adilzia Oh my God. Thank you so much sir. I really appreciate the help that you gave. WISHING YOU & YOUR FAMILY A VERY HAPPY & HEALTHY LIFE. Amen. (Sorry sir if my english was broken)
I understand there are different types of validity. Which type have you tested with this approach?
statistical
Thanks.. Sir insightful video and easy to understand... I have a little query how you have calculated or from where critical pearson table was generated with which to compare SPSS correlation outputs? Thanks
I will make a video to calculate the critical value. for the time being, you can use the available critical value table.
Its very clear sir. I am working on a research these days. After this step, can I go directly to clear collected data sir ?
Dear @vuragi, there are different processes to be followed for adapting a scale to collect data. You have to check the reliability and validity of the scale. for validity, you have to check the discriminant and convergent validity. For the different types of validity, see my other videos.
@@adilzia thank you so much sir. I'll follow your lessons.
Thank you Dr Adil Zia, just one question if this approach can be used for a large sample of say 500 respondents?
yes, u can. but now there are more advanced techniques to check the reliability and validity of the scale. u can check my other videos for that.
Hi sir, it is so nice explanation. A lot of doubts cleared. Thnks a lot.. one doubt pending, can you suggest any article for reference, which should be included in my bibliography to support the process of validation using correlation.
Ok i will send u soon. Right now i am busy in a project.
@@adilziahi sir. Could you please share it?
Hi Dr. Zia, thanks for the video. I have read the comments as well, but still i cannot find my critical value for 406 samples. I understood to N-2=404, but how can i calculate the ctirical value when 404 is not shown on the table?
The corresponding critical correlation value
for a significance level of α=0.95, for a two-tailed test is 0.003
Aslamalikum sir, thank you so much for making these knowledgeable videos they help us a lot. sir
i have a question that i am making a questionnaire by myself i have used Likert scale but after scoring i dont know what to do should i use composite of score and get the result or any other way
والله اني صبرت عليك صبر عظيم 😩🤍
thank you very much its really helpful
Dear Sir, it is really a nice explanation, i am a research scholar. How can I get this exact this table for df
dear @shamsmukhtar1591, U can listen to this video again to understand the concept of DF. I have explained this here. good luck.
hi what is this validity called? what type is this?
Hello Dr. Zla, my question is in validating a test questionnaire, are the respondents here the experts to the research topic or the respondents set for the study, like the students?
Dear @leonorapoe9218 this example is just for illustrative purposes. This is not real data.
Hello sir. Must you do the questionnaire validity testing based on data you collected from a pilot survey? Or is there no issue with checking for questionnaire validity based on data from your actual target participants?
See, you should go for pilot testing first and apply the EFA test, then u should go for CFA. But practically if u have already collected the data, u can select a small portion from the data and consider it as the pilot testing sample.
@@adilzia Thank you sir. I greatly appreciate it.
Hello, Good morning, Dr. Adil Zia. May I know what is the meaning of the result that I get for the significance value is more than 0.05, but Pearson's value based on the critical value table is acceptable? Can I consider this a valid questionnaire question or not?
Did u find the answer? Plz share
Is this should be used for each individual items (question) and under what type of validity this comes under?
yes item wise. pubmed.ncbi.nlm.nih.gov/35118967/
check this paper for details.
If not Valid means that calculated value less than tabulated value Then next step will be ? means whats effect will fall?
The next step will be to drop that item of the questionnaire and re-compute the validity after dropping that item. I hope this answers your question.
Very useful for me. Sir my significant value of two tailed test in total column is
yes, highly significant.
Hi! is it the convergent validity of the questionnaire per factor or for overall items in the questionnaire? Thank you
Same question here
Should we conduct validity test only for the liberty scale question or for all questions sir?
Good day sir, I have a question. My sample size is 135. I want to know the critical value specified for it. However, the df is 135-2= 133 and there is no 133 in the table. How can I know the critical value? Can you suggest me a study regarding it? Or can you help me know what the critical value of 133 (2-tailed) is? I would really appreciate if you could help me with this. Thank you so much!
Same issue , if u know plz let me know also
I would also appreciate if you could guide us on the same issue. My sample is 800 and I don't know where to find that comprehensive table.
Sir, it's good to learn, I want to know what is the critical value for DF=118?
Number of Degrees of Freedom (df) = 116; The corresponding critical correlation value is 0.006.
@@adilzia Sir df is 118 not 116. 2ndly if I want to cite it, from where I can check the critical value for 118 to cite?
@@asifhaider8542 yes for df 118 the critical value is the same as 0.006 I will share the reference later or u can search for it on google scholars
Hi, what will happen to the items that are lower than the critical value? Will it be removed from the questionnaire?
yes
Thank you for this video - I have tested my questions and found that one of the questions is not significant. It also has a correlation value that is below the critical value. Since the question is not valid - must I delete that question? Can I still retain that question? It is just one question out of 10 that I have. The other nine questions are all valid. What is your advise?
Yes, you can include that question also in the questionnaire but you have to justify the inclusion in the questionnaire. You have to present the supporting literature for a lower value than the critical value.
@@adilzia Thank you for your invaluable feedback.
Dear Dr.Zia, i would like to ask..what happen if there is low and even negative correlation. I read that it's still possible the questionaire still valid even the correlation is low or negative since it has a negative correlation ? please help
U have to search the context of your research. In the literature review, u can solve this issue. It all depends upon ur methodology. what is ur objective and how u handle the discussion sections. if u can find some citations related to your research, all is good.
Hi sir, Well explained thanks. Could you please clarify how to perform test-retest validity of the questionnaire using correlation
Hello Nathya. The test and retest method is used in psychology and in education. this method is used when you have the excess to the same respondents again and again. For example, if you want to test the validity of your instrument then you have to collect the data from the same respondents again and again.
in books, you will find the example of a weighing machine that if you want to test the validity of your instrument using test and retest method and you know that your wt is 50 kg and it shows 50 kg again and again it's a valid instrument. I hope this answers your question. for the rest watch my videos on this topic.
thank you so much I could understand easily
what is the method of validity test at 02:10 called? I can't find any academic article that use this method to test validity. Can you give me the source?
There are a lot of references you can search on Google scholars relevant to your context. the one I am sharing for your convenience is doi.org/10.1177/1460458220975
@@adilzia Hi, I tried to open the link, but it says "DOI Not Found". Thank you for your reply tho
pubmed.ncbi.nlm.nih.gov/35118967/
Will you explain about testing of criteria-related validity of a test?
Observe that in this case, the null hypothesis is rejected if ∣r∣>rc
@drZla Is there a scholarly source I can use to support my use if this test? Thanks in advance.
Can we change the sample size? If the SPSS sheets record 100data and if I want to test 40. Is it possible to do?
yes it is possible.
@@adilzia sir plz do let me know how? Plz upload the video. I have to proceed with the final submission of thesis couldn't find the way of how to do.
Thank u .. but , is this way we can test validity for all type of questioner, likrt scale, test questioneer?? Plz help me
This is one of the ways. U can decide based on the variables and the model u are using.
How do you get Critical Values for Pearson's Correlation Coefficient where the DF is more than 100 on the critical value table?
Hello. I have 5 independent variables and 1 dependent variable. Each variable has 5 items and a total of question 30. So my question, should I validate each of them based on construct or all in just how you did in the video. Thank you
There are different requirement of validity. You can check item wise as well as the construct level of validity. You have to justify your decision.
simple and understandable. TQ
Thank you sir for this helpful video. please, I want to ask, what if the DF is more than the DF available in the Critical Table, what am I to do? Secondly, am I to transform the variables based on each construct in the questionnaire or transform all the questions in the questionnaire together without paying attention to their different constructs, as you did in the video?
In majority of cases we Consider a questionnaire as one entity.?but it solely depends upon the researchers.
Secondly for DF I have provided the detailed explanation. Kindly see it again.
Good evening sir, do we have to include demographic questions also in total questions
No we don’t
What if the obtained value is not greater than critical value?? Still can we use pearson correaltion??
You need to listen to the video again carefully to get the answer.
sir very helpful video. sir, I want to know what is the critical value for Df 254 as my sample size is 256
The corresponding critical correlation value
for a significance level of α=0.95, for a two-tailed test is 0.004.
Thankyou sir
Sir but I've searched the table for values. The table has only 250 and 300 values : 250 =0.1236 and 300 = 0.1128. Sir, the critical value should lie between this value.
Sir please correct me if I am wrong or not.
for more clarity kindly read this real-statistics.com/correlation/one-sample-hypothesis-testing-correlation/correlation-testing-via-t-test/
@@pallavisharma524
Hi Dr. Zia, thanks for the video. I have read the comments as well, but still i cannot find my critical value for 204 samples. I understood to N-2=202, but how can i calculate the ctirical value when 202 is not shown on the table?
The corresponding critical correlation value for a significance level of alpha = 0.95, for a two-tailed test = 0.004.
I will try to make a video for the calculation of critical value.
Sir if there are 27 questions of different variable then how to check validity?
Kindly elaborate your query. What type of variables are you talking about?
Hello Sir! This is a great video. However, my sample size is 114, which makes the DF 112. Although I could not find the value in the critical value table. How can I calculate this? Thank you.
when n = 114, the Number of Degrees of Freedom is 112. The corresponding critical correlation value
for a significance level of α=0.95, for a two-tailed test is 0.006
Thank you for the video. I have a question what if some of your questions are insignificant is there any technique through which we can make them significant or what else we do??
Yes. if some of the items of the questionnaire are insignificant, you can drop them. if you increase the sample size you might get some more items as significant. there are other ways to test the validity of the questionnaire. see my other videos for that.
Greetings sir, I had query that what if few of the obtained value is less than the critical value .. what should we do after that? should we drop the item?
Hello. There are two options. One either you can increase the size of sample second, you can drop the item.
With a higher sample size you can get the desired value.
Hello, thank you again very clear, please how canwe get the critical value for a DF = 322? THANK YOU
Nice explanation
Hello Sir. Thank you for your video.
I have a data with 110 participants. Thus, the df is 108. How can I check the critical value for this? I am not getting it in the table.
Please reply
for a sample of 110, DF is 108. The corresponding critical correlation value
for a significance level of α=0.95, for a two-tailed test, is 0.006.
Sir, you mean we need to conduct initial survey in order to get the data? How many people?
The sample size depends upon many criterion.
Most acceptable method is to have 10 times the items of a questionnaire. Example If ur questionnaire has 10 items then 100 will be the sample size.
if my sample is 160 than how I find the critical value in the table and how I will identify the degree of freedom as you did not mentioned how you select the degree of freedom you just show the ppt
when n = 160, the Number of Degrees of Freedom is 158. The corresponding critical correlation value
for a significance level of α=0.95, for a two-tailed test is 0.005
How do we check the values for degrees of freedem for 127. where can i find the values?
in your question, u told me that DF is 127 which means that the sample size is 129. The corresponding critical correlation value r_cr for a significance level of alpha = 0.05, for a two-tailed test, is 0.173
Where can we get the data in the data sheet? From the questionnaire? So we need to have an initial interview? 40 people? Is it ok if i have more than 60 items?
I have replied your question in another reply. Kindly see that for your question. We this sample size is not adequate for your questionnaire.
Hello Sir please my Sample size is 116 and so my DF is 114. How to calculate my pearson critical value ?
for the sample size of 116. The corresponding critical correlation value for a significance level of alpha = 0.95, for a two-tailed test = 0.006.
I will try to make a video for the calculation of critical value.
Observe that in this case, the null hypothesis is rejected if ∣r∣>rc=0.006|r| > r_c = 0.006∣r∣>rc=0.006.
Hello sir, thank you for this content, I had a question, If I have 3 different variable and each variables have 4 different questions do I only sum up the questions related to one variable or do I sum up each and every questions ?
Dear @Ashutosh, you have to handle each variable one by one. in compute variables add all the questions of that variable divided by the number of questions. For example, if variable 1 has 3 questions then Variable 1 = (Q1+Q2+Q3)/3. I hope this answers your query.
@@adilzia Thank you, Doc, This answers my question
Hello sir. May I ask? If the item has a low critical value, how should I improve it?
You have to increase the sample size. second, you have to improve the items on the scale. third, you can drop some of the items of the scale.
@@adilzia Thank you. 🙏🙏🙏🙏
Thank you very much for the explanations.
I am trying to validate a translated version of a questionnaire intended to measure the quality of life. How many respondents should I have? The questionnaire is designed for patients diagnosed with a certain disease. Is there a mathematical formula that I should use? Is there a minimum number of respondents accepted?
The basic thumb rule is to have a sample size 10 times the total number of questions. Otherwise there are other ways also to find out the sample size.
@@adilzia Thank you very much for your response. Is there an official formula for calculating this number? I must be able to explain the reasoning for my decision.
Dr Adil... may I know if my data is not normally distributed, should i then use spearman correlation and do this same validity test in the same way?
if you are planning to apply a non-parametric test then you continue with it otherwise if you are planning to apply parametric tests then try to make your data normal.
can you please share the critical values of pearson's correation coefficient ?
Kindly check the video at 2:22 minutes. the critical value calculation depends upon the sample size. It is explained in detail here. But still, if you face any problem then you can contact me. I hope this answers your question.
@@adilzia sir in our test we have a total of 47 items. these items are divided into three categories. and i have to find the validity of all parts and total items. the sample size is 50. first part contains 23 items, 2nd part 12 items and 3rd part 12 items. how do i find the reliability? and i am also confused about what number of parameters we have..
@@tanmayiswarnkar Surely i will make one video on testing the reliability using Cronbach Alpha very soon. For your case select all your items of the questionnaire and see the value. This is called composite reliability. Then select items of the question for each dimension or factor one by one. This will be called Construct reliability. All values should be more than 0.6. I hope this answers your question.
@@tanmayiswarnkar Surely i will make a video on testing the reliability using SPSS Software. first of all, select all the items of the questionnaire and see the value of Chronbach Alpha. This is called Composite Reliability. Then select the items of the questionnaire dimension wise and see the value. This is called Construct Reliability. All the values should be more than 0.6. I hope this answers your Query.
@@adilzia Thank You it really cleared up some of my doubts. Can you please tell me what should I do to find the validity. I have computed the reliability by keeping your suggestion. The result was great!!! It was 0.973.
Good pm. What is the specific statistic tool use to analyze criterion and construct validity?
I have made a video on your topic. Kindly check that.
I thought likert scale is ordinal? How is it used as nominal here?
Sir, In case Sig Value of an item is
any value less than 0.05 is significant.
@@adilzia wow!! JazakAllah sir. May Allah reward you in abundance. Your videos are very helpful.
For cross cultural validation of existing scale of 40 items . What is the minimum sample size ?
thumb rule 10 X the number of items.
Sir, my sample size is 55, df 55-2=53 . What will be critical value??
The corresponding critical correlation value r_cr
for a significance level of α=0.95, for a two-tailed test is 0.009
I don't understand where did he get all the numbers
Which numbers are you talking about?
May I know if my sample size is 307, which means dF is 305, which row do I refer to on the Pearson table? Do I follow the row 300 or row 350?
I am not able to understand ur question. Just follow the critical value for a sample of 307 is 0.004. the null hypothesis is rejected if r > r_c.
I have sample size of 205, so my DF is 203. However, when I checked the correlation table, there's no 203. It's just 150 and 300. What shall I do? Thank you! :)
critical value for a sample of 205 is 0.004 for alpha 0.95.
Thanks a lot Dr. Zia, your reply was a timely help for me. Keep posting content. Support from India ❤ Cheers🥳
what is the critical value of DF 127 at 0.05?
i have replied in your last question.
Thank a lot to you 😊 sir . 🌸🌼🤲
How do we make the critical values table? Is it always needed?
do you have a video that shows construct validity?
kindly see the validity video in the playlist.
What source could support this process sir?
U can use google scholar to search the reference. Type the keyword and u will get the papers related to that concept.
Hello Adil, thank you for a great video! I was wondering if you could advice how to present the results of this validity test in a table according to the APA format?
For this, you can check various published articles available online. check my google scholar account for that.
How did you come up with the critical value table?, can it be generated with spas?
The critical value table is already developed and is available to use. It can be seen in every statistic book. Hope this answered your question.
If your degree of freedom DF is 203 how do measure it in the Pearson table
Hello. I saw your video. You have 10 items named with Q1-Q10. If we have categorized the variable on it's nature, is it allowed to do Q1,Q2,Q3,R1,R2,R3,S1,S2,S3 in one pearson corellation test just to make sure it is valid??
Yes you can do that. It all depends on how u have framed your objectives.
@@adilzia so Q1,Q2,Q3,R1,R2,R3,S1,S2,S3 in pearson correlation to validating instrument and all valid, then continuing to EFA for all variable at once. Is it okay?
@@Elsonnewgate that’s fine.
@@adilzia thankyou so much
My sample size is 330 and this table ends on 300. What should i do?
How did you come upon this procedure…any references???
Sir what will I do if out of 25 questions only 7 are significant. Should I delete the other questions?
Dear Vannn, u can see the language of the questionnaire. Can go for content validity again. 18 insignificant items of the questionnaire raises issues related to content validity.
@@adilzia thank you sir!
Thank you so much, Dr. Adil. The video is super helpful. I have one question though. My sample size is 277, so my DF is 275. However, when I checked the correlation table, there's no 275. What should I do? There are 21 items in my questionnaire. Thank you.
The sample size is n = 277, so then the number of degrees of freedom is df = n-2 = 277 - 2 = 275
The corresponding critical correlation value r_cr
for a significance level of \alpha = 0.05α=0.05, for a two-tailed test is: r_c = 0.118
Observe that in this case, the null hypothesis is rejected if |r| & gt; r_c = 0.118
hope this solves your problem.
@@adilzia Thank you very much, Dr. Adil.
Hi, Where did you find the critical values table I have over 100 participants totalling to 102, I need that very table to calculate my scores, How do i find a table just like yours?
The critical value for a sample of 102 is 0.006