Case control data analysis using SPSS | Odd ratio | Logistic Regression

[@taybahtube5384]

Thank you sir. It helped me

[@Kajidataonline]

Thanks for your comment. Really appreciate it. Happy learning

[@abdifatahfarahhamud6426]

we must give this gentle man a solute, with thanks giving, what a good explanation. he gave us a lesson like he Earns much money... Ah

[@Kajidataonline]

Thanks for your comment. Feel free to share the video, like and subscribe to our channel. Thanks again

[@abilbamark8364]

much appreciated .Good work.well understood sir

[@nurelinaroseli2294]

Good sharing and idea

[@Kajidataonline]

Thanks

[@s_ofi]

Thank you very much for the video. I have a question. I don´t know why we make crosstabs with all variables if we introduce all of them in regression analysis. I believed that I should introduced only the significant variables

[@Kajidataonline]

Thats very correct, only significant variable in differences test should only be used. But for the purpose of demo i share that, thanks for highlight

[@sirfanshamna1187]

Sir l am doing a case control study ,but l have to analyze 3 independent variables with a dependent variable/outcome so how to do it . Do l need to do the regression or do l need to do the multivariate analysis (MANOVA OR MANCOVA)

[@Kajidataonline]

Thanks for your comment. The case control study usually focus on binary outcome (sick/ not sick), (ill/ not ill). The use binary logistic regression was the correct one supply with the distribution test by chi square

[@adindadwiaurellia850]

hallo pak, untuk teknik analisis data case control apakah hanya pakai analisis bivariate dan univariate (chi square dan odds ratio)?

[@Kajidataonline]

Terima kasih utk komen tuan. Chi square dan odd ratio adalah ujian inferensi untuk mengenal pasti perkaitan/ perhubungan antara status sakit dan juga status makan. Iannya sesuai dilakukan bagi kajian cross-sectional dan case control

[@adindadwiaurellia850]

@@Kajidataonline apakah ada uji prasyarat seperti uji normalitas dan linearitas juga pak?

[@Kajidataonline]

Untuk case control tiada ujian specific normalisasi, sbb ujian normalisasi utk data numerical, namun, ada beberapa syarat dalam penggunaan chi square seperti expected count less than 5, not more than 20%. Benda nie dpt dilihat dibawah jadual chi square yg dihasilkan oleh chi square table di spss

[@adindadwiaurellia850]

@@Kajidataonline thankyou so much sir, god bless you aamiin!

[@Kajidataonline]

Ameen. Thanks for your comment.

[@annanoor8214]

hi sir, just wondering, why we need to do regression when can get OR trough simpler test like chi-square risk analysis? when to use chi-square risk analysis, and when to use regression?

[@Kajidataonline]

Thanks for your question. This is very good one.. Will proceed with video explaination later. Thanks mem. Anyway, as a quick explaination, the use of simple chi square and selection of risk element in crosstable analysis (analyze, descriptive and crosstab) are useful to generate the OR, however, there are sometimes you may need the adjustment of the variables. Example, you have the variables that is possible to disturb/ confounding.. therefore the use of logistic will be benefited. In addition, we use also the logistic to simulate the model of probability of occurance when combining the binary outcome with the numerical data straight away.. this enable us to make the crude and adjusted or much more accurate in prediction of OR. Beside the use of chi square risk element if need to conduct the additional layer/ confounding element, must undergo stratification process and add up with additional statistics of controls like mentel hazel n so on. Feel free to suggest what content.. will make about it

[@anasfm4545]

hello sir, i want to confirm something. so it is okay if the case:control sample ratio is not 1:1. for example if i have sample consist of 73 control and 87 case its okay then? i dont need to exclude some of the case in order to make it 73 case and 73 control? thank you in advance sir

[@Kajidataonline]

Thanks for your comment. The best ration for case control ideally 1:1, and any excess of case per control can be up to 1:2, if you have 73 and 87, the ratio seems suitable to proceed with case control variables. Please do conduct differences test to ensure the factors can be consider normal to each other. Normal means no statistical different in term of group proportion

[@anasfm4545]

@@Kajidataonline thank you for the reply sir, may i know what tests can i conduct to ensure my datas are normal?

[@Kajidataonline]

The normality test. But normality test are suitable for the numeric scale data. And dedicated for the dependent variable of interest.. Run, analyze > descriptive > explore. Have fun in data analysis and please consider to like, share and subscribe our channel and videos. Tq

[@abdelrahmanelsadat3644]

Thank you sir for your simple explanation... Allah bless you. I have two questions please: > Why did you do the binary logestic regression on two steps the first one without (A8, A11, A12) and the second one was the same but you added the (A8, A11, A12)?! and what are the correct values for other variables (like education, locality.. Ex).. The values of 1st step or the values of 2nd step to put them in the result table of the research? Or I have to make 2 tables one for crude OR and the other for Adjusted OR? . > The 2nd question is what is the difference between crude and adjusted odd ratio?! And when should I use each one?!

[@Kajidataonline]

Thanks for your question. The answer will be together. The used of different approach in step 1 and step 2 was to adjust the logistic regression values based on mechanism of crude odd ratio and adjusted odd ratio. The adjusted was used to ensure all the odd/ risk factor address already look into consideration of the bias/ influencial factor in our demographic elements. You can visit our previous simplified version of logistic regression. For titanic issues. If got any problem/ need further clarification let me know. Enjoy our learning, kindly visit and subscribe our youtube channel and telegram: t.me/kajidataonline_free Titatic example: kzbin.info/www/bejne/qWrNkKR3mtdkpKc

[@adrianadrin9033]

Helo sir, can i use OR value in chi square analysis for my case-control study data presentation?

[@Kajidataonline]

Yes sir. You can use that from the chi square risk analysis. Please do ensure the position of your data is correct, sick 1, not sick 2. Then should be good to go

[@adrianadrin9033]

@@Kajidataonline Thanks for the awesome explanation

[@saraelsadigibrahem4726]

thank you very much for your effort God pleased you.

[@Kajidataonline]

Thanks for your comment. Appreciate it.. hope you can share to your friend and family so that, all of them get the idea how to do it. Tq

[@drnareshchauhan]

great sir

[@Kajidataonline]

Thanks dr.

[@annanoor8214]

hi sir, do you have video on how to check data normality in case control study. We have to check data normality first before proceed with chi-square test right? do we combine both case and control for normality check?

[@Kajidataonline]

Dear Anna Noor, thanks for your question. The case-control study is one of the important epidemiological tools in analyzing the risk indexes between exposure and outcome. For your information, the outcome measures in the Case-control study usually binary (categorical) and therefore, the use of normality as numerical data may not need. The assumption of case-control study usually follows the basis of the chi-square (nonparametric approach). Anyway, the case-control study is not an analysis step but a STUDY design in epidemiology. Feel free to ask questions for any questions. Please also join our online group at TELEGRAM for robust/ immediate update on our news: t.me/kajidataonline_free

[@mohammadmohammadi4028]

@@Kajidataonline Salam alaikum. Thanks for your helpful video. Alongside this question, if we have numerical data like BMI in a case-control study how we can find out to use t-test or Mann Whitney U test? I mean how normality testing should be performed?

[@Kajidataonline]

Dear Mohammad. Thanks for your question. The normality can be run using analyze>descriptive>explore function. Then include your dependent variable of interest in the testing menu. The use of mann whitney or ttest is influence by your tabulation, normal or not.

[@mohammadmohammadi4028]

@@Kajidataonline thanks for your response and your explanation. I've done this as you said. For control the K-S is .000 but for case this number is .200 ( greater than 0.05). Which p value I should pick up in a Case control study for a variable?

[@Kajidataonline]

Dear sir. The p value of kolmogrov which more than 0.05 indicate that your variables was tabulated in normal tabulation/ normally distributed. Usually the normality will be conducted for overall dataset. Anyway, the use of normality for your BMI i believe to measure differences among case and control right? Try to understand the overall situation of the data.. then you can try to identify which factors different each other to be proceed with another test like logistics or chi square.

[@mierazahri8391]

Assalammualaikum.. Sir.. Boleh tnjuk cara utk analysis multiple choices answer

[@Kajidataonline]

Ok.. boleh puan..

[@biniyamtaye5287]

hi sir can i get how ur entry how u differentiate case and control

[@Kajidataonline]

Dear sir, thanks for your comment. Usually in case control study design, the case will be our focus of study and the control will be anyone who had been selected which having similar characteristic with the case and similar exposure towards certain hazard/ study interest. This is very important to measure the odd of the risk generated by the contigency table in chi square. Please do joint us in telegram for rapid question and answer. t.me/kajidataonline_free

[@biniyamtaye5287]

@@Kajidataonline suppose my out come is place of delivery for mother when i analysis bi variate first i code 0 case(health facility ) and code 1 control (home delivery ) my aim is to see determinate of institutional delivery service last 2 years but on classification table (predicted and observed table)what can i get suppose for me since i code case o under it i see value o o but on my control i see number of case and control is that correct

[@Kajidataonline]

Yeah.. its seems to be correctly classify sir for your dv (outcome), the case should be coded as 1 and control should be coded as 0.

[@JurgenMura]

subscribed. thank you sir

[@Kajidataonline]

Thanks for your subscribe.. appreciate if you could like and share this channel to your friends. Tq

[@anjumshaikh3270]

Sir, Can you please upload the data link as well to practice? And can you also make a video on SPSS all analysis methods,when to use with links to practice data for us to learn.This would be of much help. I get puzzeled when i see the several SPSS tests. Thank you, Reply from you would be appreciated.

[@Kajidataonline]

Thanks for your question/ inquiry, sorry for delay response, after searching my database, I already found it. Please find in the link given in description. Happy analyzing data.

[@anjumshaikh3270]

@@Kajidataonline Thank you sir

[@Kajidataonline]

Thanks for your inquiry, please consider, like, share and subscribe to our youtube channel. Any analysis ideas, you can straight ask me inside this platform. Happy learning

[@anjumshaikh3270]

@@Kajidataonline Sir, Can I get your email address? This would be beneficial for me to ask my queries. Thank you Sir.

[@Kajidataonline]

Sure can. You can whatsapp me at: spsshelper.wasap.my Email: kajidataonline [at] gmail.com Website: kajidataonline.com