Sampling

define each type of sampling & describe d its main strength 

    1. purposive sampling- (nonprobability) trying 2 find ppl most relevant 4 d 2pic;particular groups pf ppl who can tell me ? i want 2 know so i go look 4 dem. conveys idea dat a sample of a pop will adequ8ly rep an entire pop;using prior knowledge2 choose participants;gets ol possible cases dat fit particular areas,using various methods &chooses sample based on who researcher thinks would b appropri8 4 study.b.snowball sampling- (nonprobability) U use it wen ppl R hard 2 reach & it gives U a way in thus,using d ad: works as a chain reaction;it starts with a few participants & hs dem lead researchers 2 new cases;existing study subjects recruit additional subjects frm their contact groups.dis causes d sample group 2 keep growing similar 2 how a snowball grows.1ce d original subject hs been studied,researcher asks subject 2 find ppl w similar traits 2 expand d sampling group.it allows access 2 a group of ppl dat would usually b diff 2 get in 2uch w. c.random cluster sampling- cluster- (probability) generally used wen U wanna cover a large geographic area (still rando sample places u go & within dat sample household,ppl,wh8ver U’re doing) ad: less expensive;tradeoff btwn accuracy&cost;thus,it cuts down on cost of preparing a sampling frame &reducing travel &other administr8ve costs.it’s process of combining &elimin8on in dat it takes a large sample frame thru clusters 2 in-turn making 1-stage random samples of cluster 2cre8 a sample frame win each selected clusters in which r included in d sample,lastly in a 2-stage,drawing a rando sample of those cases;thus,subset of elements win selected clusters r randomly selected in sample.d.str8fied random sampling- (probability) wh8ver proportion U want it 2 b;we wanna oversample dis pop by dis much & dat op w dat much – cause if u wanna draw concluions U have 2 have a certain amount of ppl frm d partic group.U strat rando sampling so UR literally putting in2 pockets of types of ppl & within each pocket U rando sample (ex: 600 ppl btwn 20-25 yrs old-make sure u wanna have a certain #,U draw wh8ver proportion/r8o U’re drawing fom each 1 & do dat within each of d pockets.


2.? is a sampling frame? sampling frame is nt always 4mal,it is a method used 4 constructing a sample by making d sample as well as choosing ppl frm it.an example would b a list of names: direc2ry,census,membership list,etc.it’s rel8d 2 a research popul8on in dat a researcher’s popul8on specifies d unit being sampled,d geographical loc8on,& d temporal boundaries of popul8ons.thus,a researcher begins with an idea of d popul8on & precisely defines it in2 a target popul8on which is d specific popul8on he/she wants 2 study.due 2 popul8on being an abstract concept,researcher needs 2 oper8onalize by developing a specific list dat closely approxim8s ol d elements in d popul8on 2 research.it is relatable 2 external validity in dat d purpose of probability sampling is 2 maximize.thus,d biggest strength is 2 b able 2 generalize d 2 popul8ons U have & d degree 2 which u can generalize it 2 other popul8ons,which is external validity.



3.explain d logic of d central limit theorem.cltt–  in d probability theory,it st8s dat ne given conditions,d mean of a sufficiently large # of independent random variables,each with finite mean & variance,will b approxim8ly normally distributed.it hs a # of variants & in common 4m,random variables must b identically distributed.a major function consists of oder of experience with d help of concepts as well as selects relevant aspects & data amonly d elements multitude of “facts” dat confront d researcher.? it is & ? it does: ol d probability sample R d more samples (each dot is a sample) & d more we draw,it starts 2 make a big peak (p.229) out of 5,000 marbles (100 of each) –take ol of samples,& how many red marbles (we usually don’t know but in dis we do—pulling out 100 marbles,42 put there;put dem back,draw again – it hs been d1 enough mathe we knoe if we pull a sample out it’s likely it’ll b in dat range—very few of then can b far away-dat’s ? sampling error is;but usually we’re going 2 b in d middle/pretty close) r in each sample (only 1ce dey gor 42 wen dey pulled out 50) in d samples,most of d time dey got 50.cltt tells us,(mathem8cal theory) d # of diff rando samples in d distribution.we c ? is most predictable/most common sample) dis is ? allows us 2 generalize with samples;make a st8ment dat says if we truly sample randomly then dat sample will represent d pop pretty well.allows us 2 generalize 2 generalize samples in d popul8on as long as we use probability sampling.pull samples rando out of a pop & confiderntly say d # of sample will rep d broader popul8on (based on clt should rep d rest of d group cause gonna b in d middle somewhere)—dat theory allows us 2 generalize frm samples 2 popul8ons.important part-dey R rando selected if nt d cltt hs no applic8on.d more samples used leads 2 continuance 2 draw more samples leading 2 getting closer & closer 2 know d popul8on.

4) define each of d following 3 concepts adequ8 sample size 4 a probability sample.

wen using random sampling,U nid 2 know those 3 2 figure out how big yur sample’s going 2 b.confidence interval- determines how big U want yur sample 2 b in which d researcher determines (dict8 d range) d average of mistake dat d researcher will require 2 2ler8;set by d researcher.thus,d range around d # U might estim8 dat U will fil confident d range of popul8on will b in parameter which is determined by d finding.it is nt d range,rather d amount of assurance (popul8on parameter falls in2 dis) 4 example,d sample size would include how close dey R 2 d hoops;& d closer 1 is,d closer dey R 2 make it.) thus,confidence interval is wen d decrease is equivalent 2 d increase of sample size/d smaller d confidence interval,d larger d sample size.  due 2 d sample accuracy being based on probability & distributional assumption,confidence levels specify d levels of risk dat d research is willing 2 take;dis is associ8d with d confidence level,which,again,is d range of values in which d value of d popul8on along with d probability R included in d results.thus,d precise estim8s must include hi confidence intervals 2 b equivalent with hi confidence levels & a necessary increased sample size.dis is also important 4 surveys 4 it estim8s reported.a range of estim8d values dat is d best guess as 2 d tru popul8on’s value.confidence intervals R usually calcul8d 4 d sample mean.st8stically,dis means dat if 100 random samples were drawn frm a popul8on & confidence intervals were calcul8d 4 d mean of each of d samples,95 of d confidence intervals would contain d popul8on’s mean.4 example,a 95% confidence interval 4 iq of 95 2 105,indic8s with 95% certainty dat d actual average iq in d popul8on lies btwn 95 & 105.a random sample of a popul8on,which ensures dat each member of d popul8on hs a chance of being selected 4 d sample.confidence level tells us how far/big UR going 2 go away frm dat (ci goes outside frm dat) in making a claim dat d pop parameter falls in there somewhere.in d findings,how big of a range RU gonna allow in making st8ment dat d pop parameter falls in d range (how big d range/target) bigger target,easier 2 hit.more likely 2 get closer 2 d actual pop parameter with d ___with how close u can fall & if U want… dis worksheet is tellin how many U nid 2 have;its saying,if u set a large ci,U don’t nid as many ppl but if u have a small ci,U nid more so U can hit d target (pop) hoop=confidence level.

probability level- (dict8 d level of precision).d average abnormality score dat expresses d percentage of a variable’s values dat fall within a set of interval wen d variable is normally distributed.d lower d probability level is equivalent 2 d lower nid of precision.d most common is 1.56-1.58,it measures how certain U want yur sample’s ci will over d popul8on interval & 1ce dat’s set,how sure i want 2 b dat d popul8on lies in there.ex: 1 making a basketball hoop with their arms & telling some1 2 make it in d hoop- d bigger 1 makes their arms,d higher d probability level will b,whereas,d smaller d hoop,d lower d probability level.(or precision levels…d higher d precision level=d higher d likelihood,vice versa).a random sample of a popul8on,which ensures dat each member of d popul8on hs a chance of being selected 4 d sample.  d most common is 1.56-1.58 measuring how certain U want yur sample’s confidence interval will cover d popul8on interval.thus,1ce U have d set,how sure U want 2 b d popul8on lies there.U can check d probability level if U want 2 b 95% sure d popul8on falls in2 d confidence interval (parameter popul8on).d lower d pl,d lower d nid 2 b precise.shooter=pl.variance- a commonly used measure of dispersion 4 variables.d variance is calcul8d by squaring d standard devi8on.d variance is based on d square of d difference btwn d values 4 each observ8on & d mean value.d degree of similarity on d key variable in a sample popul8on;so d variance leads 2 a level of similarity which then is equivalent 2 d sample size being less.it jst exists,researcher doesn’t control it.d best 1 can do is equivalent 2 d best estim8.thus,a lot of variability would b d variance remaining d same.it measures how far a set of numbers is spread out;in other words a parameter in which explains whether or nt d actual probability distribution of numbers frm d observed popul8on &/or d distribution of a theoretical probability of a sample of numbers;in other words nt a complete observ8on of d numbers.small variance indic8s d data points R near d expected value whereas hi variance indic8s d data points R very spread out.variance(s) R important in probability sampling coz dey R often applied within d estim8on of d variance of consistent distribution frm a sample of dat distribution.a random sample of a popul8on,which ensures dat each member of d popul8on hs a chance of being selected 4 d sample.it is determined by d researcher which is d1 by their best estim8on 2 figure out how many ppl.example,? is d % sample of ppl who smoke.pop looking @- 60% of those who smoke 2 yur variance would b .06,.07 & U nid 2 determine dat thru estim8on.if d ppl R more alike,U don’t nid as many coz U can assume.d more variance U have,d more likely it is going 2 b dey’re going 2 b different on ol things.more variance nid more ppl — less variance-nid less ppl coz there is more hemogence in behavior.d more diff u expect pop 2 b on d main variable,then more ppl u nid in yur study 2 b able 2 identify dat.–larger ci,U can go w less ppl.large ci=can go w less ppl — hi prob level-nid more ppl,more variance=more ppl,less variance,less ppl.wen U use random sampling,U nid 2 know these 2 how big yur sample is going 2 b.