Quantitative data analysis with examples

In this video, we're going to jump into the often  confusing world of quantitative data analysis.   We're going to explore what quantitative data  analysis is, some of the most popular analysis   methods and how to choose the right methods for  your research. We'll also cover some useful tips,   as well as common pitfalls to avoid when  you're undertaking quantitative analysis.   So grab a cup of coffee, grab a cup of tea,  whatever works for you and let's jump into it! Hey! Welcome to Grad Coach TV - where we demystify  and simplify the oftentimes intimidating world of   academic research my name is Emma and today  we're going to unwrap the topic of quantitative   data analysis if you're new here be sure to hit  that subscribe button for more videos covering   all things research-related also if you're  looking for hands-on help with your research   check out our one-on-one coaching services where  we help you through your dissertation thesis   or research project step by step it's basically  like having a professor in your pocket whenever   you need it so if that sounds interesting to you  you can learn more and book a free consultation   with a friendly coach at www all right  with that out of the way let's jump into it quantitative data analysis is one of those things  that often strikes fear into students it's totally   understandable quantitative analysis is a complex  topic full of daunting lingo like medians modes   correlations and regression suddenly we're all  wishing we'd paid a little more attention in math   class now the good news is that while quantitative  data analysis is a mammoth topic gaining a working   understanding of the basics isn't that hard even  for those of us who avoid numbers and math at all   costs in this video we'll break quantitative  analysis down into simple bite-sized chunks   so you can get comfy with the core concepts  and approach your research with confidence   so let's start with the most basic question what  exactly is quantitative data analysis despite   being quite a mouthful quantitative data analysis  simply means analyzing data that's numbers based   or data that can be easily converted into  numbers without losing any meaning for example   category based variables like gender ethnicity  or native language can all be converted into   numbers without losing meaning for example  english could equal one french could equal two   and so on this contrasts against qualitative data  analysis where the focus is on words phrases and   expressions that can't be reduced to numbers  if you're interested in learning about   qualitative analysis we've got a video covering  that as well i'll include a link below so the   next logical question is what is quantitative  analysis used for well quantitative analysis is   generally used for three purposes first it's used  to measure differences between groups for example   average height differences between different  groups of people second it's used to assess   relationships between variables for example  the relationship between weather temperature   and voter turnout and third it's used to test  hypotheses in a scientifically rigorous way   for example a hypothesis about  the impact of a certain vaccine   again this contrasts with qualitative analysis  which can be used to analyze people's perceptions   and feelings about an event or situation in other  words things that can't be reduced to numbers   so how does quantitative analysis work you ask  well since quantitative data analysis is all   about analyzing numbers it's no surprise that it  involves statistics statistical analysis methods   form the engine that powers quant analysis these  methods can vary from pretty basic calculations   for example averages and medians to more  sophisticated analyses for example correlations   and regressions sounds like a bunch of gibberish  don't worry we will explain all of that in this   video importantly you don't need to be a  statistician or a math whiz to pull off a   good quantitative analysis we'll break down  all the technical mumbo jumbo in this video   so let's start by taking a look at the  two main branches of quantitative analysis as i mentioned quantitative analysis is powered  by statistical analysis methods there are two main   branches of statistical methods that are used  descriptive statistics and inferential statistics   in your research you might only use descriptive  statistics or you might use a mix of both   depending on what you're trying to figure out in  other words depending on your research questions   aims and objectives i'll explain how to  choose your methods later in this video   so what are descriptive and inferential statistics  well before i can explain that we need to take a   quick detour to explain some lingo to understand  the difference between these two branches   of statistics you need to understand two  important words these words are population   and sample first step population in statistics  the population is the entire group of people or   animals or organizations or whatever that  you're interested in researching for example   if you were interested in researching tesla  owners in the us then the population would be   all tesla owners in the united states however  it's extremely unlikely that you're gonna be   able to interview or survey every single tesla  owner in the u.

S realistically you'll only get   access to a few hundred or maybe a few  thousand owners using an online survey   this smaller group of accessible people whose  data you actually collect is called your sample   so to recap the population is the entire group of  people you're interested in and the sample is the   subset of that population that you can actually  get access to in other words the population is   the full chocolate cake whereas the sample is just  a slice of that cake can you see what i've got on   my mind anyhow why is this sample population thing  important well descriptive statistics focuses on   describing the sample while inferential statistics  aim to make predictions about the population   based on the findings within the sample in other  words we use one group of statistical methods   descriptive statistics to investigate the slice  of cake and another group of methods inferential   statistics to draw conclusions about the entire  cake and there i go with the cake analogy again   but to be fair i always have chocolate on my  mind so with that out of the way let's take a   closer look at each of these branches in more  detail starting with descriptive statistics descriptive statistics serve a simple but  critically important role in your research   to describe your data set hence the name in other  words they help you understand the details of   your sample unlike inferential statistics which  we'll get to later descriptive statistics don't   aim to make inferences or predictions about the  entire population they're purely interested in   the details of your specific sample when you're  writing up your analysis descriptive statistics   are the first set of stats you'll cover before  moving on to inferential statistics but depending   on your research objectives and research questions  they may be the only type of statistics that you   use we'll explore that a little later so what kind  of statistics are usually covered in this section   well some common statistical tests used in this  branch include the following the mean this is   simply the mathematical average of a range of  numbers nothing too complicated here next is the   median this is the midpoint in a range of numbers  when the numbers are all arranged in order if the   data set makes up an odd number then the median  is the number right in the middle of the set   if the data set makes up an even number then  the median is the midpoint between the two   middle numbers next up is the mode this is simply  the most commonly repeated number in the data set   then we have standard deviation this metric  indicates how dispersed a range of numbers is   in other words how close all the numbers are  to the mean the average in cases where most   of the numbers are quite close to the average  the standard deviation will be relatively low   conversely in cases where the numbers are  scattered all over the place the standard   deviation will be relatively high lastly we have  skewness as the name suggests skewness indicates   how symmetrical a range of numbers is in other  words do they tend to cluster into a smooth bell   curve shape in the middle of the graph this  is called a normal or parametric distribution   or do they lean to the left or right this is  called a non-normal or non-parametric distribution   okay are you feeling a bit confused  let's look at a practical example   on the left hand side is the data set this data  set details the body weight in kilograms of a   sample of 10 people on the right hand side we  have the descriptive statistics for this data set   let's take a look at each of them first we can  see that the mean weight is 72.

4 kilograms in   other words the average weight across the sample  is 72.4 kilograms pretty straightforward next   we can see that the median is very similar to the  mean the average this suggests that this data set   has a reasonably symmetrical distribution in other  words a relatively smooth center distribution of   weights clustered towards the center moving on to  the mode well there is no mode in this data set   this is because each number presents itself only  once and so there cannot be a most common number   if hypothetically there were two people who were  both 65 kilograms then the mode would be 65.   next up is the standard deviation 10.6 indicates  that there's quite a wide spread of numbers we   can see this quite easily by just looking at the  numbers which range from 55 to 90. this is quite a   stretch from the mean of 72.4 so we would expect  the standard deviation to be well above zero   and lastly let's look at the skewness a result  of negative 0.

2 tells us that the data is very   slightly negatively skewed in other words it has  a very slight lean this makes sense since the   mean and the median are only slightly different as  you can see these descriptive statistics give us   some useful insight into the data set of course  this is a very small data set only 10 records   so we can't read into these statistics too much  but hopefully this example helps you understand   how these statistics play out in reality also keep  in mind that this is not a list of all possible   descriptive statistics just the most common  ones so at this point you might be wondering   but why do these matter well while these  descriptive statistics are all fairly basic   they're important for a few reasons firstly  they help you get both a macro and micro level   view of your data they help you understand  both the big picture and the finer details   secondly they help you spot potential errors in  the data for example if an average is way higher   than you'd expect or responses to a question  are highly varied this can act as a warning   sign that you need to double check the data and  lastly these descriptive statistics help inform   which inferential statistical methods you can use  as those methods depend on the shape of the data   we'll explore this a little bit more later  on simply put descriptive statistics are   really important even though the  statistical methods used are pretty basic   all too often at grad coach we see students  rushing past the descriptives in their eagerness   to get to the more exciting inferential methods  and then landing up with some very flawed results   don't be a sucker give your descriptive  statistics all the love and attention they deserve all right now that we've looked at  descriptive stats let's move on to   the second branch of quantitative  analysis inferential statistics as i mentioned while descriptive statistics are  all about the details of your specific data set   your sample inferential statistics aim to make  inferences about the population in other words   you'll use inferential statistics to make  predictions about what you'd expect to find   in the full population what kind of predictions  you ask well generally speaking there are two   common types of predictions that research  try to make using inferential stats firstly   predictions about differences between groups  for example height differences between children   grouped by their favorite sport and secondly  relationships between variables for example   the relationship between body weight and the  number of hours a week a person does yoga   in other words inferential statistics when  done correctly allow you to connect the   dots and make predictions about what you'd  expect to see in the real world population   based on what you observe in your sample data  for this reason inferential statistics are used   for hypothesis testing in other words to test  hypotheses that predict changes or differences   of course when you're working with inferential  statistics the composition of your sample is   really important in other words if your  sample doesn't accurately represent the   population you're researching then your findings  won't necessarily be very useful for example   if your population of interest is a mix of 50  male and 50 female but your sample is 80 male   you can't make inferences about the population  based on your sample since it's not representative   this area of statistics is called sampling but we  won't go down that rabbit hole here it's a deep   one we'll save that for another video so what kind  of statistics are usually covered in this section   well there are many many different statistical  analysis methods within the inferential branch   and it would be impossible for us to discuss  them all here so we'll just take a look at   some of the most common inferential statistical  methods so that you have a solid starting point first up are t-tests t-tests compare the  means the averages of two groups of data   to assess whether they are different to a  statistically significant extent in other words   to see whether they have significantly different  means standard deviations and skewness for example   you might want to compare the mean blood pressure  between two groups of people one that has taken a   new medication and one that hasn't to assess  whether they are significantly different   simply looking at the two means  is not enough to draw a conclusion   you need to assess whether the differences  are statistically significant and that's   what t-tests allow you to do right next up is  anova anova stands for analysis of variance   this test is similar to a t-test in that it  compares the means of various groups but anova   allows you to analyze multiple groups not just  two so it's basically a t-test but on steroids next we have correlation analysis this type of  analysis assesses the relationship between two   variables in other words if one variable increases  does the other variable also increase decrease   or stay the same for example if the average  temperature goes up do average ice cream sales   increase too we'd expect some sort of relationship  between these two variables intuitively   but correlation analysis allows us to  measure that relationship scientifically lastly we have regression analysis regression  analysis is similar to correlation in that it   assesses the relationship between variables but  it goes a step further to understand the cause   and effect between variables not just whether they  move together in other words does the one variable   actually cause the other one to move or do they  just happen to move together naturally thanks to   another force just because two variables correlate  doesn't necessarily mean that one causes the other   to make this all a little more tangible let's  take a look at an example of correlation in   action here's a scatter plot demonstrating the  correlation or the relationship between weight and   height intuitively we'd expect there to be some  sort of relationship between these two variables   which is what we see in this scatter plot in other  words the results tend to cluster together in a   diagonal line from bottom left to top right the  more tightly the results cluster together to form   a line in any direction the more correlated they  are and therefore the stronger the relationship   between the variables as i mentioned these are  just a handful of inferential methods there   are many many more importantly each statistical  method has its own assumptions and limitations   for example some methods only work with normally  distributed or parametric data while other methods   are designed specifically for data that are  not normally distributed and that's exactly why   descriptive statistics are so important they're  the first step to knowing which inferential   methods you can and can't use of course this  all begs the question how do i choose the right   quantitative analysis methods for my research  well that's exactly what we'll look at next now that we've looked at some of the most common  statistical methods used within quantitative   analysis let's look at how you go about choosing  the right tool for the job to choose the right   statistical methods for your research you need  to think about two important factors one the   type of quantitative data you have specifically  level of measurement and the shape of the data   and two your research questions and hypotheses  let's take a closer look at each of these the   first thing you need to consider is the type of  data you've collected or the data you will collect   by data types i'm referring to the four levels  of measurement namely nominal ordinal interval   and ratio if you're not familiar with this lingo  you should hit the pause button real quick and   go check out our post over on the grad coach blog  that explains each of these levels of measurement   i'll include the link below okay so why does this  matter well because different statistical methods   require different types of data this is one of the  assumptions i mentioned earlier every method has   its assumptions regarding the type of data for  example some methods work with categorical data   like yes or no type questions while others work  with numerical data like age weight or income if   you try to use a statistical method that doesn't  support the data type you have your results will   be largely meaningless so make sure you have a  clear understanding of what types of data you've   collected or will collect once you have this you  can then check which statistical methods support   your data types i'll include a link below the  video that explains which methods support which   data types now if you haven't collected your data  yet you can of course reverse engineer the process   and look at which statistical methods would  give you the most useful insights and then   design your data collection strategy around this  to ensure that you collect the correct data types another important factor to  consider is the shape of your data   specifically does it have a normal distribution  in other words is it a bell-shaped curve   centered in the middle or is it  very skewed to the left or right   again different statistical methods work for  different shapes of data some are designed   for symmetrical data while others are designed  for skewed data this is another reminder of why   descriptive statistics are so important since  they tell you all about the shape of your data the next thing you need to consider is your  specific research questions as well as your   hypotheses if you have some the nature of your  research questions and research hypotheses   will heavily influence which statistical  methods you should use if you're just   interested in understanding the attributes of your  sample as opposed to the entire population then   descriptive statistics might be all you need for  example if you just want to assess the means or   averages and the medians or center points of  variables in a group of people descriptives   will do the trick on the other hand if you aim  to understand differences between groups or   relationships between variables and to  infer or predict outcomes in the population   then you'll likely need both descriptive  statistics and inferential statistics so   it's really important to get very  clear about your research aims   and research questions as well as your hypotheses  before you start looking at which statistical   methods to use never shoehorn a specific method  into your research just because you like it   or have experience with it your choice of methods  must align with all the factors we've covered here all right now that we've looked  at what quantitative analysis is   the two main branches of statistics and how  to choose the right methods for your research   let's recap and bring it all together we've covered a lot in this video  well done on making it this far   let's recap on the key points we've looked at  first we asked the question what is quantitative   data analysis as we discussed quantitative  analysis is all about analyzing number based data   which can include both categorical and numerical  data these data are analyzed using statistical   methods the two main branches of statistics are  descriptive statistics and inferential statistics   descriptives describe your sample the slice of  the cake while inferentials make predictions   about what you'll find in the population the full  cake based on what you've observed in the sample   as we saw common descriptive statistical  metrics include the mean the median the mode   standard deviation and skewness on the inferential  side we looked at t tests anovas correlation   analysis and regression analysis all of which can  help you make predictions about the population   lastly we asked the important question how  do i choose the right statistical methods   as we discussed to choose the right  statistical methods you need to consider   the type of data you're working as well  as your research questions and hypotheses   remember in this video we've only looked at a  handful of the most common quantitative methods   there are many many more so be sure to check out  the grad coach blog as well as the other links   below this video to get a fuller picture of what  all's on offer in terms of statistical methods   also if you'd like us to cover any of the methods  in more detail be sure to leave a comment below alright that wraps it up for today if you  enjoyed the video hit that like button and   leave a comment if you have any questions also  be sure to subscribe to the grad coach channel   for more research related content lastly if you  need a helping hand with your research check out   our private coaching service where we work with  you on a one-on-one basis chapter by chapter to   help you craft a winning dissertation thesis or  research project if that sounds interesting to   you book a free consultation with a friendly  coach at www www.

Bradcoach.com as always i'll   include a link below that's all for this episode  of grad coach tv until next time good luck you.