Essentially correct, but actually misleading because in this example each stratum (yes, that's the singular) is the same size. In typical cases, strata are different sizes, so when selecting the same *proportion* of each stratum we get different numbers of individuals selected. The result is that the strata in the sample are in the same proportions as those in the population we are sampling from. Also omits to say why we do this - which is to reduce sampling error.
***** no, because as the video say, having decided how many we need from each stratum, we use probability sampling (e.g. random numbers) to select them from all those in the stratum. So in the simple case, everyone in the population has an equal probability of being chosen. You can sample disproportionately to make sure you have big enough numbers from very small strata, but then you need to weight the results appropriately in analysis.
***** Sampling error is the difference between a properly drawn random sample and the actual population characteristic. It is the error that comes about by mere chance (as opposed to bias, that comes about for other reasons). Thus if you draw a sample of 4 people from a large population that is exactly 50% men and 50% women then some of the time (actually 25% of the time) you will get a sample of 1 man and 3 women. Other things being equal, the larger the sample, the smaller the sampling error.
Thanks alot for the comment Graham! If you had not commented I would have tought that he actually meant that we need to use a fixed number, and not a proportion for the stratas
Once you put the population into groups (strata), you can list them however you like. You could give each person a number, or list them alphabetically, etc. Then use SRS to randomly choose the desired number from each group.
GOD BLESS YOU FOR THIS😭🙌 Lord knows I was suffering today, my tutor made absolutely no sense in 20 minutes, & you made perfect sense in less that than 6! 🙈
+Letty D You're welcome and feel free to talk about God as much as you like. My faith in a real God that loves me and cares for me has brought me through many rough times in my life. Not to mention I believe that God is the one that created the math that we are studying. Man may have discovered it, but God created it. Many will attack us for our faith, but the Bible tells that we will be rewarded in Heaven for our persecution.
love you brother..we are unique in our own ways..we are all humans with diversities..i wish everybody in this world become like you and me..world would be so much beautiful
lol i checked this video out for my stats class an years ago... and now since i forgot about it, this came in my research class again. and you this video came to my rescue again lol THANKSS
In stratified sampling there are individual groups in a population which are known as "stratas"., which are grouped by different characteristics. Then depending on the number of things/people in each strata, you proportionately sample them to make up the total sample. You then use srs to choose which of person or thing is going to be chosen in each strata.
In this case, as the strata are the same size it is hard to see what's happening. Proportionate stratified sampling means taking the same percentage sample from each stratum (in this case 40% or 12/30). These will be different numbers if the strata are different sizes. Disproportionate sampling means taking different percentage samples from each stratum (e.g. to ensure minimum numbers in the small strata) and then weighting the data in the analysis to account for under or over representation.
thanks steve for mmaking this very easy its very easy way of explaining something and i am really very thankfull to you because I can understand all types of SRS because of you
@flakita007ec You're using a combination of sampling techniques.First you are grouping your 3000 physicians into strata.Then once your groups have been made you are using a systematic sampling technique to choose your sample from each group.For example, you've grouped the physicians (pediatrician, surgeon, OBGYN) These are your strata, but then you selected every 10th pediatrician on your list, every 10th surgeon, and every 10th OBGYN. By selecting every 10th you are using systematic sampling.
Thank you for this! Our proposal defense is next week and I couldn't understand my research adviser when he tried explaining this to me :((( Idk why but this video made a lot more sense.
You skipped the point of considering the size of each stratum. Since each stratum (or group, in simple terms) is of same size, i.e. 10, so we get to randomly choose 4 (one-third of 12, one-third, because we have 3 strata) persons from each stratum. If the stratum sizes were different, we would have chosen more persons from the larger stratum and fewer from the smaller ones. It is done to ensure that our sampled set is representative of the whole population.
@gurjeev chadha: That's because in doing unbiased statistical sampling, one should use "sampling without replacement", which means that once you choose a number, you do not use it again for the sake of giving equal opportunity for other numbers in the population (N) to participate in the sampling process. For example, if you need to do a survey to find out the average age of employees who work in company "A". Assuming there are 100 employees in that company and you started randomly choosing the employees from a randomly generated list and in this case let's assume we need only 20 sample size (n). When you start selecting random employees, let's say you selected employee number 17 on the list, to remove bias in selection process, you cross out employee number 17 from the list because you already chose him/her. If you don't do this, then there is a potential risk of duplicating the results and therefore getting inaccurate results. I hope this helps! :)
Steve Mays, thank you very much. I would love to hear from you on how to analyze likert scale level of agreement, I am an academician by profession and would love to simplify for my Uni class.
thanks for your answer. and i find your video really helpful but it's probably only me who couldn't see this part but i would've found it more helpful to me if you showed bit more detail on how you would've done the srs after. thanks!
It may have been good to have different numbers of each strata. That way you could talk about maintaining proportions when sampling. Great video though.
If you separate by colors and then again by gender, that would be considered a multi-step sample. It's allowed, but it would be like creating strata twice. This could also be considered BLOCKING if you were to use this technique in an experiment.
It's ok to choose a different number from each stratum as long as they are close. It's not ideal and I would recommend chosing a different sample size, so you can choose the same amount from each group.
I have a question. For example, we got 3 strata. and we have these values: 3, 2.5, 4.5. adding these numbers is equal to 10. so, we got 10 respondents. but there's no such thing as a "0.5" person. So, we usually round up to 3, 3, 5. adding them is equal to 11. or maybe we can also go with 3, 2, 5, or 3, 3, 4. Now, how do we distribute the remaining allocation of respondents?
If I have three non-mutually exclusive sampling strata and wish to interview a roughly equal number from each strata, what type of sampling is this called? For example, my groups are runners, jumpers, and divers. However, some jumpers are also divers and one guy is all three. Think of a three-way Venn diagram where each "pure" circle has 5 interviewees and there is additional overlap between the groups. Is this some sort of modified stratified quota sampling?
How are you selecting no. 2 from blue strata at first then 5th 7th and 10th.? Can you choose no. 1 3 5 & 7 from blue? What is the order of selection here?
ohh thanks for the quick reply and answer. so i just list within each strata right? so there are 3 groups with 10 ppl in it each. so like let's say for the blue guys group. 1.john 2. mark 3. spencer 4. rory 5. jacob 6~10. and i just use random generated number chart and pick the first 4 according to the number that first appears from the number chart? so let's say 04, 02, 01 , 07 appeared in that order.
and i agree with the guy telling u to replace with my teacher. bc my professor wouldn't explain through email. he's like, come ask during tutorial or office hours. THE HELL.
thank you very much, that was really helpful. i have a question about what if one of the groups was more than the others? lets say the red group has 12 ladies, do i still choose only 4?
I just want to say that I genuinely laughed at 4:00. Thank you for that. 🤣
Essentially correct, but actually misleading because in this example each stratum (yes, that's the singular) is the same size. In typical cases, strata are different sizes, so when selecting the same *proportion* of each stratum we get different numbers of individuals selected. The result is that the strata in the sample are in the same proportions as those in the population we are sampling from.
Also omits to say why we do this - which is to reduce sampling error.
*****
no, because as the video say, having decided how many we need from each stratum, we use probability sampling (e.g. random numbers) to select them from all those in the stratum. So in the simple case, everyone in the population has an equal probability of being chosen. You can sample disproportionately to make sure you have big enough numbers from very small strata, but then you need to weight the results appropriately in analysis.
Yeah, I was wondering why he had 10 of each...
***** Sampling error is the difference between a properly drawn random sample and the actual population characteristic. It is the error that comes about by mere chance (as opposed to bias, that comes about for other reasons). Thus if you draw a sample of 4 people from a large population that is exactly 50% men and 50% women then some of the time (actually 25% of the time) you will get a sample of 1 man and 3 women. Other things being equal, the larger the sample, the smaller the sampling error.
Thanks alot for the comment Graham! If you had not commented I would have tought that he actually meant that we need to use a fixed number, and not a proportion for the stratas
Thanks from 4 years later haha!
Once you put the population into groups (strata), you can list them however you like. You could give each person a number, or list them alphabetically, etc. Then use SRS to randomly choose the desired number from each group.
The best informative/hilarious vid on startified sampling
I just can't 4:00 😂😂😂😂😂😂😂😂😂😂😂😂😂😂😂😂
4:01 made me laugh ;) thank you very helpful videos
GOD BLESS YOU FOR THIS😭🙌 Lord knows I was suffering today, my tutor made absolutely no sense in 20 minutes, & you made perfect sense in less that than 6! 🙈
+Letty D You're welcome and feel free to talk about God as much as you like. My faith in a real God that loves me and cares for me has brought me through many rough times in my life. Not to mention I believe that God is the one that created the math that we are studying. Man may have discovered it, but God created it. Many will attack us for our faith, but the Bible tells that we will be rewarded in Heaven for our persecution.
Thank you for your help!!! This was great! Also, coming from a black guy, it is ok to say black guys. haha
love you brother..we are unique in our own ways..we are all humans with diversities..i wish everybody in this world become like you and me..world would be so much beautiful
From 11 years and still helpful for my graduation project thanks a lot
lol i checked this video out for my stats class an years ago... and now since i forgot about it, this came in my research class again. and you this video came to my rescue again lol THANKSS
In stratified sampling there are individual groups in a population which are known as "stratas"., which are grouped by different characteristics. Then depending on the number of things/people in each strata, you proportionately sample them to make up the total sample. You then use srs to choose which of person or thing is going to be chosen in each strata.
Black characters hahaha. Dude I don't think anyone would be offended if you said black men.
Those anti-fascist cunts all over the world spread this shit.People are even scared to talk!!!!!
ahaha expected this to be top when I heard it
I thought the exact same thing. Congratulations SJWs of the world. This is the society you have created through your over-sensitivity.
@Kurt T Ikr lmao anti-SJWs are so cringy.
@@apoellasapoelazio5377 So you're pro-fascist?
Wish all teachers were this clear. Studying for a test tomorrow in a class with a frustrating teacher who won't explain things!
In this case, as the strata are the same size it is hard to see what's happening. Proportionate stratified sampling means taking the same percentage sample from each stratum (in this case 40% or 12/30). These will be different numbers if the strata are different sizes. Disproportionate sampling means taking different percentage samples from each stratum (e.g. to ensure minimum numbers in the small strata) and then weighting the data in the analysis to account for under or over representation.
Great... So simple explanation. Love it👍
ahaha, he didn't want to say black people. Something is happening subconsciously...
😂😂true
I must thank you! I used this in my exam...and i perfectly understood all of them! :D
thanks steve for mmaking this very easy
its very easy way of explaining something and i am really very thankfull to you because I can understand all types of SRS because of you
Glad I could help.
Internet is good for learning, I am using for learning the English, thank very much you.
Its great to see that, even in 2021, statistics does not offend
"black guys strata" lol
Compton
Perfect explanation thank you! Just one tiny thing the individual group is a stratum, the plural of it is strata 😁
Thanks a lot...you have made it easier to understand. Though i was already very clear about the concept. :) ;)
thank you so much for this , you made it a lot easier xD
this illustration is very helpful. thank you
that was really helpful thanks a lot!!! Steve
@flakita007ec
You're using a combination of sampling techniques.First you are grouping your 3000 physicians into strata.Then once your groups have been made you are using a systematic sampling technique to choose your sample from each group.For example, you've grouped the physicians (pediatrician, surgeon, OBGYN) These are your strata, but then you selected every 10th pediatrician on your list, every 10th surgeon, and every 10th OBGYN. By selecting every 10th you are using systematic sampling.
thank you so much for helping me for my report ❤❤❤
Thank you for this! Our proposal defense is next week and I couldn't understand my research adviser when he tried explaining this to me :((( Idk why but this video made a lot more sense.
You SIR are AMAZING thanks for this. its so helpful
Each one of these groups is called STRATUM, they all are called STRATA...
Vladimir M thank you
Thank you this is helpful.
This is spoonfeeding thank you
The Blue Man Group is offended.
You skipped the point of considering the size of each stratum. Since each stratum (or group, in simple terms) is of same size, i.e. 10, so we get to randomly choose 4 (one-third of 12, one-third, because we have 3 strata) persons from each stratum. If the stratum sizes were different, we would have chosen more persons from the larger stratum and fewer from the smaller ones. It is done to ensure that our sampled set is representative of the whole population.
The black guys😂😂
@gurjeev chadha: That's because in doing unbiased statistical sampling, one should use "sampling without replacement", which means that once you choose a number, you do not use it again for the sake of giving equal opportunity for other numbers in the population (N) to participate in the sampling process. For example, if you need to do a survey to find out the average age of employees who work in company "A". Assuming there are 100 employees in that company and you started randomly choosing the employees from a randomly generated list and in this case let's assume we need only 20 sample size (n). When you start selecting random employees, let's say you selected employee number 17 on the list, to remove bias in selection process, you cross out employee number 17 from the list because you already chose him/her. If you don't do this, then there is a potential risk of duplicating the results and therefore getting inaccurate results. I hope this helps! :)
Steve Mays, thank you very much. I would love to hear from you on how to analyze likert scale level of agreement, I am an academician by profession and would love to simplify for my Uni class.
Thanks for easy explanation!
So good and helpful!! Thank you!!
thanks bro cluster and stratified seemed the same to me until these videos
This has been incredibly helpful thank you! Could you do Purposive sampling as well?
Thank you so much, this was very helpful
Oh our lecturer mentioned proportionate vs disproportionate stratified sampling. U've done proportionate..wish you could hv shown disproportionate. :)
Nothing wrong with saying black people XD Ahh that was funny
Great video though! Really helped, I subbed
really helpfull, thankyouu
thanks this solved so many issues
thanks for your answer. and i find your video really helpful but it's probably only me who couldn't see this part but i would've found it more helpful to me if you showed bit more detail on how you would've done the srs after. thanks!
Awesome, thanks for sharing!
thx for all your videos! This is coming in very handy for my Data Management class! short videos are always better than confusing lectures.
Thank you for video 👌
It may have been good to have different numbers of each strata. That way you could talk about maintaining proportions when sampling. Great video though.
nice tutorial! thanks a lot for the video
Maybe I can do an example with disproportionate stratified sample when I get a chance. Thanks for the suggestion.
You sound like ross from friends.
I dunno why
Good Job & thnx I needed that.
lmao i knew this video was gonna be funny 🤣
If you separate by colors and then again by gender, that would be considered a multi-step sample. It's allowed, but it would be like creating strata twice.
This could also be considered BLOCKING if you were to use this technique in an experiment.
Thank you ❤
excellent.... thanx
What if each strata has different size, do we still pick the same number of sample for each strata? Your answers will be very helpful. Thank you
no you make it proportional to the number in each strata
very good explanation
great explanation
Anyone else notice the black man try to sneak in behind the red ladies?
kinda racist
ha ha :D.
It's ok to choose a different number from each stratum as long as they are close. It's not ideal and I would recommend chosing a different sample size, so you can choose the same amount from each group.
Bro loved the presentation
I'm reading Dane right now and this was a great way to summarize Ch. 5 Sampling.
very helpful video - thanks
Video made my day :)
Can you please come to Chico and re-teach my entire high school curriculum?
I have a question. For example, we got 3 strata. and we have these values: 3, 2.5, 4.5. adding these numbers is equal to 10. so, we got 10 respondents. but there's no such thing as a "0.5" person. So, we usually round up to 3, 3, 5. adding them is equal to 11. or maybe we can also go with 3, 2, 5, or 3, 3, 4. Now, how do we distribute the remaining allocation of respondents?
I use Smart Notebook software. It comes with Smart Recorder, which is what I use to capture the video.
This is correct. If I interchanged them in the video, I'm sorry.
how would you take a sample of 9 from 30 observations where there are only 18 males and 12 females?
If I have three non-mutually exclusive sampling strata and wish to interview a roughly equal number from each strata, what type of sampling is this called?
For example, my groups are runners, jumpers, and divers. However, some jumpers are also divers and one guy is all three. Think of a three-way Venn diagram where each "pure" circle has 5 interviewees and there is additional overlap between the groups. Is this some sort of modified stratified quota sampling?
nice job. thnks
what methods, Kent and Coker (1992) used in vegetation discription
Thank you extremely helpful
So we have Smurfs, Basketball people and Ed Sheeran as three strata
i came to learn but hearing this guy desperately try not to seem like a racist was hilarious
the sound quality was not the best for me personally...but the video was very informative though!! thanks! liked it!!
1:22 i lol'd
Very helpful!
What if each strata population doesnt have an equal frequency?
top quality😁😁🙌🙌
Thank you
This is porportianant stratified random sampling. It would be great if you could make disproportianant one too.
How are you selecting no. 2 from blue strata at first then 5th 7th and 10th.? Can you choose no. 1 3 5 & 7 from blue? What is the order of selection here?
Jug your beer for every time "group" is said. ;)
should the "strata" be equal or it's not necessary ?????
What if groups had not an equal number of individuals?
ohh thanks for the quick reply and answer. so i just list within each strata right? so there are 3 groups with 10 ppl in it each. so like let's say for the blue guys group. 1.john 2. mark 3. spencer 4. rory 5. jacob 6~10. and i just use random generated number chart and pick the first 4 according to the number that first appears from the number chart? so let's say 04, 02, 01 , 07 appeared in that order.
and i agree with the guy telling u to replace with my teacher. bc my professor wouldn't explain through email. he's like, come ask during tutorial or office hours. THE HELL.
thank you very much, that was really helpful.
i have a question about what if one of the groups was more than the others? lets say the red group has 12 ladies, do i still choose only 4?
Thank you!
Laughed when he recounts at 3:15 hahahaha (Im sorry) but thanksss!
+Queenie Joy Resmundo love u
Tnx for this video
Is there any math to this like a formula or a way to figure this out on a calculated? I want to know how to do that
What are advantages and disadvantages?
thx! i can understand it in a minute
Does that mean for a Probability Random technique, we need to have a list of the population compared to Non Probability Sampling technique?