https://padlet.com/cushla_thomson/sampling en-us 2018-05-30 22:56:43 UTC 2018-05-30 23:11:14 UTC hello@padlet.com https://padlet.com/cushla_thomson/sampling/wish/264648859 We must use bootstrapping as we are just looking at a singular sample from the population.  Bootstrapping allows us to sample with replacement 1000 times therefore it will help us discover the true population median/give us a brief idea of where it lies displayed with bootstrap confidence intervals.  
If we took another sample from our population we would could see a different median.]]> 2018-05-30 22:57:38 UTC https://padlet.com/cushla_thomson/sampling/wish/264648859 https://padlet.com/cushla_thomson/sampling/wish/264648878 Bootstrapping is definitely required as when we want to make a claim on the whole population, we can't rely on just our sample, and as we will never see the population median we have to find an alternative method]]> 2018-05-30 22:57:52 UTC https://padlet.com/cushla_thomson/sampling/wish/264648878 https://padlet.com/cushla_thomson/sampling/wish/264648893 If we took a different sample there would be different data and we would get different values. Unfortunately we are unable to take different samples due to limitations and restriction, and that is why we bootstrap. Bootstrapping allows us to make a claim on the whole population through the sample that we are given. 1000 times!]]> 2018-05-30 22:58:05 UTC https://padlet.com/cushla_thomson/sampling/wish/264648893 https://padlet.com/cushla_thomson/sampling/wish/264648896 We cannot just rely on our sample in order to make accurate inference about the population, as a sample does not give us a full picture of the population. With a different sample, there would be different data points which would result in different values and a potentially different result. This is called sampling variability, and is a fact that we have to deal with. In order to minimize the effect that sampling variability ]]> 2018-05-30 22:58:07 UTC https://padlet.com/cushla_thomson/sampling/wish/264648896 https://padlet.com/cushla_thomson/sampling/wish/264648900 2018-05-30 22:58:09 UTC https://padlet.com/cushla_thomson/sampling/wish/264648900 https://padlet.com/cushla_thomson/sampling/wish/264648909
Boot strapping is nessacery after looking at 2 samples. the analysis because we never see the real population median. Boot strapping lets us see what the median will be with 1000 tries. This means when we take another sample, (the dataset is not the same as sample) we will get a different median because the data set is not the same. boot strapping is useful as in real life this process will be too expensive and too long to re sample the existing values.]]> 2018-05-30 22:58:13 UTC https://padlet.com/cushla_thomson/sampling/wish/264648909 https://padlet.com/cushla_thomson/sampling/wish/264648939 - due to there being only one sample we will never see the population median due to price and other factors, as we cannot take different samples the best way to get the closest to the population median we will use bootstrapping which is sampling with RESAMPLING ]]> 2018-05-30 22:58:27 UTC https://padlet.com/cushla_thomson/sampling/wish/264648939 https://padlet.com/cushla_thomson/sampling/wish/264648943 The reason we use bootstrapping in our report is to find a range of values in which we are pretty sure the true population median lies between. We cannot just use the median from the sample, as there is sampling variability. This means that if we were to take another sample, we would get a different median, as the dataset will be slightly different. In real life, we can also usually only take one sample, due to costs and accessibility, so bootstrapping allows us to simulate the action of getting multiple sample medians with just one sample, provided the initial sample is representative of the population.]]> 2018-05-30 22:58:29 UTC https://padlet.com/cushla_thomson/sampling/wish/264648943 https://padlet.com/cushla_thomson/sampling/wish/264648953 There is a population median, which we never see because we are just looking at a sample of the population. If we took a different sample, we would get different values. We can't do lots of samples in real life, so we need to use the sample median as the basis for a range of values where we can be pretty sure that the population median lies. To do this, we use a bootsrapping confidence interval. We do bootstrapping by resampling from the original samples with replacement. We calculate the median of the two new sample groups 1000 times, and then calculate the difference between these medians 1000 times. The central 95% of these values are selected as the confidence interval. We remove the the lower 2.5% and the higher 2.5%. When we do bootstrapping, even though we're doing resampling with replacement, we get a very similar picture to the real picture of repeated samples that we don't ever see. We use NZGrapher to create a bootstrapping confidence interval. We then interpret the confidencd interval, and make a call about whether we can make a claim that one median is bigger than the other.]]> 2018-05-30 22:58:35 UTC https://padlet.com/cushla_thomson/sampling/wish/264648953 https://padlet.com/cushla_thomson/sampling/wish/264649097 Should I repeat the sampling process? ........, I will get different values, and therefore my graphs and results would be slightly different. Because of this, we are unsure of whether or not my current sample will represent or be close enough to the true median of the population. In order to be confident that we can make a call, and answer our question, we use a bootstrapping confidence interval. The bootstrapping process involves ]]> 2018-05-30 22:59:51 UTC https://padlet.com/cushla_thomson/sampling/wish/264649097 https://padlet.com/cushla_thomson/sampling/wish/264649129 By looking at the sample we see features that reflect what we see in the population. However, because we aim the see the true population and we never see it, we use bootstrapping to solve this issue. Also consider that if we were to take a different sample we would see different results, this is sampling variability. Bootstrapping consists of resampling from the original sample with replacement.]]> 2018-05-30 23:00:12 UTC https://padlet.com/cushla_thomson/sampling/wish/264649129