Sir thank you. Very clear po sir. Di ko po maintyendihan sa subject teacher ko. pero sayo ang galing mo magturo. Wow salamat sir. Godbless dahil di kayo madamot magturo.
Yun po yung nagather na data sa research. Sa kaso po nung example sa video, pinag-take ng researcher yung mga participant ng 50-item Math test tapos kung ano po yung naging score ng participants doon sa Math test (kunwari), yun po yung nagsilbing data ng research. Doon po nanggaling yung data for T.L(x) which is yung scores sa 50-item Math test nung mga nag-undergo ng traditional learning at data for O.L(y) which yung scores sa 50-item Math test nung mga nag-undergo ng online learning. Kumbaga, given na po yung data sa example problem na nasa video, talagang hahanapin na lang kung may significant difference ba sa dalawang sets ng data na ito. Sana nasagot ko po ang inyong katanungan 🙂
Ask lang po hehe Ang akala ko po kapag ka ≤,≥ and = sign is automatically null hypothesis, pero bakit po irereject ang ≤ ? And diba dapat po ang rejected is yung >?
Hello! In null hypothesis, we assume that our groups of data are just the same. For example, we have group X and group Y. The null hypothesis always assumes that X=Y or there is no difference in the results between groups X and Y. In layman's term, pareho lang po ng resulta si X and Y. That's the null hypothesis. Now about doon sa decision rule, bakit kapag greater than the critical value yung nacompute na value ay inaaccept natin yung null hypothesis (or inaaccept natin yung fact na walang ganong kalaking pinagkaiba yung resulta sa group X at group Y)? To answer that, think of this: halimbawa, perfectly parehas lang ng ranks lahat ng data sa parehas na grupo (or in short, wala siyang difference talaga). Kung ganon, ang bawat isang data sa example natin sa video ay magkakaroon ng rank 6.5, right? Saan galing yung 6.5 na rank? Yun po yung average ng ranks 1, 2, 3, 4.... 12 (since 12 yung data sa example). Ngayon kapag pinag-add lahat ng ranks per group, halimbawa kay group X, ang lalabas kay group X na RANK SUM ay 39 (because of the fact na merong anim na participants sa group X at bawat isa sa kanila ay rank 6.5, so 6.5 x 6= 39). Now, since inassume nga natin kunwari na pare parehas ng rank lahat ng participant natin, kesyo nasa group X or Y siya, kung ano man yung resulta kay X, ganon din yung resulta kay Y, so 39 din yung RANK SUM ni Y. Now, try to plug in that "39" to the formula for Mann-Whitney. Ang lalabas na computed U ay 18, right? And 18 is the maximum possible computed U value out of 12 data that we have IF THE DATA IS PERFECTLY IDENTICAL OR WALANG SIGNIFICANT DIFFERENCE, and as we know, 18 is greater than 5 (which is the critical value for our example). That's the reason why we ACCEPT OUR NULL HYPOTHESIS IF THE COMPUTED U IS GREATER THAN CRITICAL U VALUE. (Continuation below)
Now about doon sa isa niyo pa pong question, why do we reject the null hypothesis kapag ang nacompute na U value is less than or equal to the critical value? In my previous reply, you already knew na kapag pala walang pinagkaiba yung dalawang grupo ng data at parehas lang sila ng rank sum or talagang perfectly identical yung two groups of data natin, dapat greater than the critical value yung makuha nating computed U value. Ngayon, wala namang perfect na data, right? Habang nagkakaroon ng difference yung dalawang grupo ng data natin sa example, mas lumalayo siya sa 18. Pababa ng pababa yung computed U habang nadadagdagan yung difference ng rank sum ng dalawang grupo ng data natin sa example. Now, there will be a point na kung saan parang sasabihin ng statistician na "Teka nga lang, hanggang saan yung ibababa nitong computed U value para mareject ko ang null hypothesis?" or in short "Ano ba ang limit ko para masabi ko na irereject ko na ang null hypothesis?" At yun yung function ng critical value natin na 5. Mula 18, hanggang before 5 lang ang dapat na ibaba ng computed U value mo para maaccept mo ang null hypothesis o para masabi mo na "walang significant difference yung dalawang grupo ng data"; otherwise, kung nagfall na siya sa 5 and below, kailangan mo ng ireject ang null hypothesis. Bakit? It's because sa example natin na may 12 participants, never kang makakakuha ng computed U value na less than or equal to 5 if wala or konti lang ang difference ng two groups of data mo. Para patunayan ito, let's have a new assumption. Halimbawa, yung Group X natin ay may ranks 1, 2, 3, 4, 5 and 6 (RANK SUM X= 21) tapos yung Group Y natin ay may ranks 7, 8, 9, 10, 11 and 12 (RANK SUM Y= 57). Sa example natin na ito, obviously, PERFECTLY UNIDENTICAL at sobrang laki ng difference ng rank sum ng dalawang grupo ng data para masabi natin na wala silang significant difference. Now, try to plug in these assumed rank sum data in the Mann-Whitney formula and you will be astonished to know na ang lalabas na computed U value ay zero (0) and zero is less than 5. Anong napansin natin na pattern dito? Habang lumalayo tayo sa PERFECTLY IDENTICAL data or habang nadadagdagan yung difference ng dalawang data, mas lumiliit ang computed U value natin hanggang sa mameet natin yung limit natin na critical value, and kung as in sobra sobra talaga ang laki ng difference between the two groups of data, then possible na makakuha tayo ng computed U value na talagang less than the critical value. This is the reason why WE REJECT THE NULL HYPOTHESIS IF THE COMPUTED U VALUE IS LESS THAN OR EQUAL TO THE CRITICAL U VALUE (or nirereject natin yung statement na "parehas lang ang data ng dalawang grupo"). I hope na nasagot ko ang iyong katanungan at pasensya na kung super haba ng reply ko.
Hello sir! question lang po about sa alpha. wala po kasing instruction na nakalagay sa given situation na need kong applyan ng wilcoxon mann-whitney U test. Ang magiging alpha ko po ba ay 0.05? Pahabol din po. Ito po bang process nasa video ay Wilcoxon mann whitney U test? Thank you po
Hello Karen! Most of the time, 0.05 ang alpha kapag walang sinabi ang prof or teacher niyo na alpha level, and yes, about your second question, this is the Wilcoxon-Mann-Whitney U test. 🙂
from the formula in numerator n(n+1) where n is the sample size for the group Y (which is 6). Therefore, it will become 6(6+1) which is equal to 6(7). Now, 6 x 7 is 42. Doon nakuha yung 42
Hello, sorry for my late reply kasi hindi po ito nagnotify sa aking notifcation. About your question po, nanggaling po yung mga number na yan sa data na na-gather sa actual research. Kumbaga sa problem po na iyan, ako lamang po ang nagbigay niyan.
Formation of conclusion na po with rrl-supported reasons kung bakit naging ganon ang result, as well as elaboration of the null hypothesis (e.g., "since there is no sig diff between the data, then it can be assumed that traditional learning is as effective as e-learning, etc. etc.) then yun na po. In this case, since hindi po natin nireject ang null hypothesis, wala po tayong gagawing post-hoc test. I hope nasagot ko po ang inyong katanungan. Thank you 🙂
Galing po siya sa "number of samples in a group multiplied to the number of samples in a group plus one" or n x (n + 1). For example, kung sa isang group, merong 5 samples, then 5 x (5 + 1) = 30. Sa given naman sa video, since merong anim na samples sa X, then 6 x (6 + 1)= 42. Ganon din sa Y group since meron ding anim na samples sa Y. Doon po nanggaling yung 42, following the formula
Hello po, my question po ako :) sakop po ba ng mann-whitney ang wilcoxon rank-sum test at wilcoxon signed-rank test? akala ko other terms lang sila pero medyo different approach po pala. btw, thank you so much po, galing nyo po. nalito lang ako sa isa kong napanood na gumamit ng z test hehehehe
Hello po. Sorry po for late reply. Magkaiba po ang Mann-Whitney and Wilcoxon ma'am since si Mann-Whitney po is for two groups of independent sample whereas si Wilcoxon po ay for two groups of dependent samples. And tama po, may iba po talaga na kung saan after po nila makuha yung Mann-Whitney or Wilcoxon computed value, pina-plug in po nila yung computed value sa z-test. That is still correct po. Pero kung ginamit po ang computed value ng Mann-Whitney or Wilcoxon sa z-test, yung critical value table for z-test na po ang icoconsult. Astonishingly po, whether we use the Mann-Whitney computed value or plug in that value to z-test, same conclusion lang po ang maoobtain. Maraming salamat po ma'am 💕
Hello! This is awesome. The simplest presentation yet most clear in explanation. Question though. Your equation/calculation is different from another tutorial. the video link ua-cam.com/video/jkpRGUkzFn4/v-deo.html&ab_channel=tutor2u and example of their equation: U1 = (n1)(n2) + (n1(n1+1))/2 - ∑R1 Does multiplying the n matter?
![Mann-Whitney U Test [Simply explained]](/img/n.gif)
very clear... great explanations...very helpful...looking forward to be as clear as you are.
It's my first time leaving a comment. Thank you for this great explanation, it's better than all the others (including my professors)!
TYVM. Well-explained lesson! Will recommend your videos to my students :)
Salamat po sir. 🙂
Best explanation I was looking for❤
Sir thank you. Very clear po sir. Di ko po maintyendihan sa subject teacher ko. pero sayo ang galing mo magturo. Wow salamat sir. Godbless dahil di kayo madamot magturo.
Well explained! Very clear discussion, bravo!
VERY CLEAR EXPLANATION.. THANK YOU VERY MUCH.
Thank you so much! You really explained well, it is a big help for me.
Very insightful ❤
Sir,you explained it so well...
thank you so much !! you explained it so well
Thank you po
question po, pano po kinuha yung R3 + R4?
Hello ? Can I ask ?
PAno po nakuha yung nga sagot po sa T.L(x) and E.L(y) ?
Yun po yung nagather na data sa research. Sa kaso po nung example sa video, pinag-take ng researcher yung mga participant ng 50-item Math test tapos kung ano po yung naging score ng participants doon sa Math test (kunwari), yun po yung nagsilbing data ng research. Doon po nanggaling yung data for T.L(x) which is yung scores sa 50-item Math test nung mga nag-undergo ng traditional learning at data for O.L(y) which yung scores sa 50-item Math test nung mga nag-undergo ng online learning. Kumbaga, given na po yung data sa example problem na nasa video, talagang hahanapin na lang kung may significant difference ba sa dalawang sets ng data na ito. Sana nasagot ko po ang inyong katanungan 🙂
paano naging 42yun?
Ask lang po hehe
Ang akala ko po kapag ka ≤,≥ and = sign is automatically null hypothesis, pero bakit po irereject ang ≤ ? And diba dapat po ang rejected is yung >?
Hello! In null hypothesis, we assume that our groups of data are just the same. For example, we have group X and group Y. The null hypothesis always assumes that X=Y or there is no difference in the results between groups X and Y. In layman's term, pareho lang po ng resulta si X and Y. That's the null hypothesis.
Now about doon sa decision rule, bakit kapag greater than the critical value yung nacompute na value ay inaaccept natin yung null hypothesis (or inaaccept natin yung fact na walang ganong kalaking pinagkaiba yung resulta sa group X at group Y)? To answer that, think of this: halimbawa, perfectly parehas lang ng ranks lahat ng data sa parehas na grupo (or in short, wala siyang difference talaga). Kung ganon, ang bawat isang data sa example natin sa video ay magkakaroon ng rank 6.5, right? Saan galing yung 6.5 na rank? Yun po yung average ng ranks 1, 2, 3, 4.... 12 (since 12 yung data sa example). Ngayon kapag pinag-add lahat ng ranks per group, halimbawa kay group X, ang lalabas kay group X na RANK SUM ay 39 (because of the fact na merong anim na participants sa group X at bawat isa sa kanila ay rank 6.5, so 6.5 x 6= 39). Now, since inassume nga natin kunwari na pare parehas ng rank lahat ng participant natin, kesyo nasa group X or Y siya, kung ano man yung resulta kay X, ganon din yung resulta kay Y, so 39 din yung RANK SUM ni Y. Now, try to plug in that "39" to the formula for Mann-Whitney. Ang lalabas na computed U ay 18, right? And 18 is the maximum possible computed U value out of 12 data that we have IF THE DATA IS PERFECTLY IDENTICAL OR WALANG SIGNIFICANT DIFFERENCE, and as we know, 18 is greater than 5 (which is the critical value for our example). That's the reason why we ACCEPT OUR NULL HYPOTHESIS IF THE COMPUTED U IS GREATER THAN CRITICAL U VALUE. (Continuation below)
Now about doon sa isa niyo pa pong question, why do we reject the null hypothesis kapag ang nacompute na U value is less than or equal to the critical value?
In my previous reply, you already knew na kapag pala walang pinagkaiba yung dalawang grupo ng data at parehas lang sila ng rank sum or talagang perfectly identical yung two groups of data natin, dapat greater than the critical value yung makuha nating computed U value. Ngayon, wala namang perfect na data, right? Habang nagkakaroon ng difference yung dalawang grupo ng data natin sa example, mas lumalayo siya sa 18. Pababa ng pababa yung computed U habang nadadagdagan yung difference ng rank sum ng dalawang grupo ng data natin sa example. Now, there will be a point na kung saan parang sasabihin ng statistician na "Teka nga lang, hanggang saan yung ibababa nitong computed U value para mareject ko ang null hypothesis?" or in short "Ano ba ang limit ko para masabi ko na irereject ko na ang null hypothesis?" At yun yung function ng critical value natin na 5. Mula 18, hanggang before 5 lang ang dapat na ibaba ng computed U value mo para maaccept mo ang null hypothesis o para masabi mo na "walang significant difference yung dalawang grupo ng data"; otherwise, kung nagfall na siya sa 5 and below, kailangan mo ng ireject ang null hypothesis. Bakit? It's because sa example natin na may 12 participants, never kang makakakuha ng computed U value na less than or equal to 5 if wala or konti lang ang difference ng two groups of data mo. Para patunayan ito, let's have a new assumption. Halimbawa, yung Group X natin ay may ranks 1, 2, 3, 4, 5 and 6 (RANK SUM X= 21) tapos yung Group Y natin ay may ranks 7, 8, 9, 10, 11 and 12 (RANK SUM Y= 57). Sa example natin na ito, obviously, PERFECTLY UNIDENTICAL at sobrang laki ng difference ng rank sum ng dalawang grupo ng data para masabi natin na wala silang significant difference. Now, try to plug in these assumed rank sum data in the Mann-Whitney formula and you will be astonished to know na ang lalabas na computed U value ay zero (0) and zero is less than 5.
Anong napansin natin na pattern dito? Habang lumalayo tayo sa PERFECTLY IDENTICAL data or habang nadadagdagan yung difference ng dalawang data, mas lumiliit ang computed U value natin hanggang sa mameet natin yung limit natin na critical value, and kung as in sobra sobra talaga ang laki ng difference between the two groups of data, then possible na makakuha tayo ng computed U value na talagang less than the critical value. This is the reason why WE REJECT THE NULL HYPOTHESIS IF THE COMPUTED U VALUE IS LESS THAN OR EQUAL TO THE CRITICAL U VALUE (or nirereject natin yung statement na "parehas lang ang data ng dalawang grupo").
I hope na nasagot ko ang iyong katanungan at pasensya na kung super haba ng reply ko.
Hello sir! question lang po about sa alpha. wala po kasing instruction na nakalagay sa given situation na need kong applyan ng wilcoxon mann-whitney U test. Ang magiging alpha ko po ba ay 0.05?
Pahabol din po. Ito po bang process nasa video ay Wilcoxon mann whitney U test?
Thank you po
Hello Karen! Most of the time, 0.05 ang alpha kapag walang sinabi ang prof or teacher niyo na alpha level, and yes, about your second question, this is the Wilcoxon-Mann-Whitney U test. 🙂
thank you po! :)@@vegangelo_29
pwede po gamiyin tuh pag study ay about the feasibility of fertilizer
Slr. Yes po, pwedeng pwede siya gamitin. Basta make sure na meron kayong 2 sets of data from independent samples that are not normally distributed 🙂
Kapag po Reject Null Hypothesis, ang conclusion po ba automatic not equal? Thanks po need lang sa assignment
I'm so sorry for the late reply. Yes po, kapag nareject ang null, automatically not equal yung either mean/median nung dalawa o higit pang groups. 🙂
Sir pano po nkuha ang 42/2 sa conputation
from the formula in numerator n(n+1) where n is the sample size for the group Y (which is 6). Therefore, it will become 6(6+1) which is equal to 6(7). Now, 6 x 7 is 42. Doon nakuha yung 42
Like yung mga 37,28,34,40,42,39?
Hello, sorry for my late reply kasi hindi po ito nagnotify sa aking notifcation. About your question po, nanggaling po yung mga number na yan sa data na na-gather sa actual research. Kumbaga sa problem po na iyan, ako lamang po ang nagbigay niyan.
After getting po the result of the accepted hypothesis, what will be the next step? Like for the findings po using the gathered data?
Formation of conclusion na po with rrl-supported reasons kung bakit naging ganon ang result, as well as elaboration of the null hypothesis (e.g., "since there is no sig diff between the data, then it can be assumed that traditional learning is as effective as e-learning, etc. etc.) then yun na po. In this case, since hindi po natin nireject ang null hypothesis, wala po tayong gagawing post-hoc test. I hope nasagot ko po ang inyong katanungan. Thank you 🙂
san galing po ang 42?
Galing po siya sa "number of samples in a group multiplied to the number of samples in a group plus one" or n x (n + 1). For example, kung sa isang group, merong 5 samples, then 5 x (5 + 1) = 30. Sa given naman sa video, since merong anim na samples sa X, then 6 x (6 + 1)= 42. Ganon din sa Y group since meron ding anim na samples sa Y. Doon po nanggaling yung 42, following the formula
Bkit 42?
Ok n iadd muna ang 6+1
Hello po, my question po ako :) sakop po ba ng mann-whitney ang wilcoxon rank-sum test at wilcoxon signed-rank test? akala ko other terms lang sila pero medyo different approach po pala. btw, thank you so much po, galing nyo po. nalito lang ako sa isa kong napanood na gumamit ng z test hehehehe
Hello po. Sorry po for late reply. Magkaiba po ang Mann-Whitney and Wilcoxon ma'am since si Mann-Whitney po is for two groups of independent sample whereas si Wilcoxon po ay for two groups of dependent samples. And tama po, may iba po talaga na kung saan after po nila makuha yung Mann-Whitney or Wilcoxon computed value, pina-plug in po nila yung computed value sa z-test. That is still correct po. Pero kung ginamit po ang computed value ng Mann-Whitney or Wilcoxon sa z-test, yung critical value table for z-test na po ang icoconsult. Astonishingly po, whether we use the Mann-Whitney computed value or plug in that value to z-test, same conclusion lang po ang maoobtain. Maraming salamat po ma'am 💕
Sana po masagot po ninyo
Hello! This is awesome. The simplest presentation yet most clear in explanation. Question though. Your equation/calculation is different from another tutorial.
the video link ua-cam.com/video/jkpRGUkzFn4/v-deo.html&ab_channel=tutor2u
and example of their equation: U1 = (n1)(n2) + (n1(n1+1))/2 - ∑R1
Does multiplying the n matter?